Luminous Landscape Forum

Equipment & Techniques => Cameras, Lenses and Shooting gear => Topic started by: ErikKaffehr on October 02, 2009, 01:09:34 am

Title: AA-filtering CCD and CMOS
Post by: ErikKaffehr on October 02, 2009, 01:09:34 am
Hi,

I have scanned trough some articles regarding MTF and "fill factors". It seems that there is a lot of research going on about MTF of sensors at the subpixel level. One issue that came to my mind is that CCD sensor cells approach 100% fill factor (according to some of the articles) while CMOS sensors have a much smaller "fill factor" and the actual sensel area is quite irregular, usually "L-shaped".

As CCD-s seem to have larger fill factor, it seems that they would be in less need of AA-filtering (or OLP-filtering to use another word). This may explain, in part, while MFDBs and Leica can do without AA-filter.

Partly, because it's obvious that both MFDBs and Leica M9 can have "color moiré" in certain cases.

Just "Google" for "fill factor and MTF" and you will find a lot of articles....

Best regards
Erik


Title: AA-filtering CCD and CMOS
Post by: BernardLanguillier on October 02, 2009, 01:19:44 am
Quote from: ErikKaffehr
I have scanned trough some articles regarding MTF and "fill factors". It seems that there is a lot of research going on about MTF of sensors at the subpixel level. One issue that came to my mind is that CCD sensor cells approach 100% fill factor (according to some of the articles) while CMOS sensors have a much smaller "fill factor" and the actual sensel area is quite irregular, usually "L-shaped".

Beyond these totally relevant technical factors, the biggest difference between CCD and CMOS is probably the fact that Japanese companies all use CMOS with AA filters while non Japanese companies all use CCD without AA filters.

It would be interesting to think about the reasons why these different technological options were selected by different engineering cultures, but I am more interested today in the downstream part of the process, the communication between brands and reviewers.

It is obvious that Western reviewers like Michael have much closer contacts with Western brands using CCDs without AA filter than they have with Japanese brands using CMOS with AA filters.

It seems fair to say that the promoters of CMOS with AA filters have not invested as much effort in communicating to Western reviewers about their motivations. We are basically only hearing one side of the story.

Cheers,
Bernard
Title: AA-filtering CCD and CMOS
Post by: Slough on October 02, 2009, 03:23:17 am
Another area that gets little attention is gamut, which surely is due to the choice of colour filters. I suspect the Japanese manufacturers attend more to things that look good in specs and tets such as high ISO performance, and pixel count i.e. the things that dpreview pixie peepers focus on.
Title: AA-filtering CCD and CMOS
Post by: BernardLanguillier on October 02, 2009, 04:28:22 am
Quote from: Slough
Another area that gets little attention is gamut, which surely is due to the choice of colour filters. I suspect the Japanese manufacturers attend more to things that look good in specs and tets such as high ISO performance, and pixel count i.e. the things that dpreview pixie peepers focus on.

Well... they are not alone. I remember being laughed at big time when I claimed last year on this very forum that true RGB devices like Betterlight backs delivered significantly better colors...

Cheers,
Bernard
Title: AA-filtering CCD and CMOS
Post by: georgl on October 02, 2009, 04:56:34 am
The fill-rate (amount of light-sensitive area) is IQ-wise one of the most important technical aspects of sensors. The higher the fill-factor, more photons hit the (larger) light-sensitive surface and more photons can hit it before saturation sets in (important for DR). Of course the analogue sensor signal then has to be amplified and converted in digital information - but what's lost in the sensor, cannot be restored.

Full-frame CCDs need special manufacturing sites (they're expensive), they're slow, they consume much energy and they're stupid (they're just a large array of photodiodes). But they have one big advantage: fill-rate, it's still not 100% but close.
So they are able to gather more information, create a high-quality and still pretty much unprocessed (no filtering etc.) signal. That's why they are used for prefessional digital cameras.

CMOS-based systems have come a long way, but their fill-rate is still quite low but they're capable of very sophisticated image-processing to reduce noise (and they have to, especially certain kinds of pattern noise due to the specific amplification process of CMOS), that's why high-iso-files from these cameras appear quite clean but also tend to be more artificial.
When comparing noise, we have to compare processed (because CMOS-files are always filtered internally) RAW-files regarding noise AND detail. Regarding CMOS and CCD we then have to care about the same pixel-pitch, usage of microlenses and the age of the architecture. The common 6.8µm-CCDs are from 2004! We will have such a rare opportunity to compare "CMOS vs. CCD" again with the appearence of the S2 - it has microlenses (the other 6µm-CCD-systems don't have those) and is comparable to the 20+MP-DSLRs and when the engineers didn't mess up the processing (conversion...) the noise/detail will look different but won't be worse than with CMOS-based systems.

CMOS is the future, the new EVIL-systems need their capability to implement live-view, even for professional camera systems. But the approach will be different, when not using AA-filters, reduce internal processing to minimum and similar HQ-color-filters as in current CCDs, their "look" won't be much different than current CMOS - if there will be a difference at all!
CMOS or CCD-sensors itself don't have a "look", they're just electronic devices to convert light into electric information! The size of the sensels, the microlenses, the architecture etc. don't even affect sharpness/MTF because these things are all affecting individual pixels (just like sharpness is no longer an issue with TFT-displays instead of CRTs). Only the AA-filter affects "sharpness" sensor-wise (given the same pixel-pitch/size)

AA-filters are needed to reduce alaising artifacts "out of the sensor" and have nothing to do with the choice of CCD vs. CMOS. That's what the Japanese DSLRs are designed for: professional press-photography in the upper end and amateurs on the lower end. They both need their images fast without post-processing, even if they have to compromise IQ. MFDB-files are processed carefully in post.
CCDs can be used without microlenses, making them more sensitive to oblique light rays and avoid aberrations - important for technical cameras. But you loose about one stop of sensitivity, because microlenses try to compensate the low fill-rate and focus light only on light-sensitive areas of the sensor. That's why they are more important with CMOS-designs, their "loss in sensitivity" due to the lack of microlenses would be much bigger.
Title: AA-filtering CCD and CMOS
Post by: Daniel Browning on October 02, 2009, 02:31:44 pm
Quote from: ErikKaffehr
As CCD-s seem to have larger fill factor, it seems that they would be in less need of AA-filtering (or OLP-filtering to use another word). This may explain, in part, while MFDBs and Leica can do without AA-filter.

Yes, the optical fill factor does relate to how bad the aliasing will be without an OLPF; however, CCD has not had an advantage in optical fill factor for many, many years thanks to the invention of microlenses. Digicams got up to 100% optical fill factor ~4 years ago, but it was only two years ago that 36x24mm DLSRs finally hit that last few percent to reach 100.

So no, this has not been a reason for the difference in AA-filtering for a long time.
Title: AA-filtering CCD and CMOS
Post by: aaykay on October 02, 2009, 05:55:21 pm
Quote from: BernardLanguillier
Beyond these totally relevant technical factors, the biggest difference between CCD and CMOS is probably the fact that Japanese companies all use CMOS with AA filters while non Japanese companies all use CCD without AA filters.

And then the fact that there are Japanese sourced CCDs with AA filters....   .....proving that the usage of AA filters have nothing to do with CCD or CMOS but a deliberate design choice by the camera (not sensor) manufacturers.

Title: AA-filtering CCD and CMOS
Post by: BernardLanguillier on October 02, 2009, 06:36:38 pm
Quote from: aaykay
And then the fact that there are Japanese sourced CCDs with AA filters....   .....proving that the usage of AA filters have nothing to do with CCD or CMOS but a deliberate design choice by the camera (not sensor) manufacturers.

Yes, that is indeed correct if you factor in compact cameras. Aren't all Japanese reflex CMOS nowadays?
 
Regards,
Bernard
Title: AA-filtering CCD and CMOS
Post by: ErikKaffehr on October 02, 2009, 06:40:49 pm
Beranrd,

There are some CCD based designs around.

I suggest that you check out this: http://www.imx.nl/photo/leica/camera/page157/page157.html (http://www.imx.nl/photo/leica/camera/page157/page157.html)

Best regards
Erik

Quote from: BernardLanguillier
Yes, that is indeed correct if you factor in compact cameras. Aren't all Japanese reflex CMOS nowadays?
 
Regards,
Bernard
Title: AA-filtering CCD and CMOS
Post by: BernardLanguillier on October 02, 2009, 06:58:21 pm
Quote from: ErikKaffehr
Beranrd,

There are some CCD based designs around.

I suggest that you check out this: http://www.imx.nl/photo/leica/camera/page157/page157.html (http://www.imx.nl/photo/leica/camera/page157/page157.html)

I was speaking about Japanese DSLRs Erik.

Cheers,
Bernard
Title: AA-filtering CCD and CMOS
Post by: ErikKaffehr on October 03, 2009, 12:46:06 am
Sorry Bernard!

The two comments were not related. Some of Sony's low end DSLRs are CCD-based, there may be some others.

The Erwin Puts article is interesting because he compares the optical performance of the Nikon D3X with the M9. In short:

- D3X comes out on top
- M9 is very good
- But M9 has also Moiré

That helicopter shoot of yours is very sharp, BTW, you have a recepie for sharpening?

Best regards
Erik

Quote from: BernardLanguillier
I was speaking about Japanese DSLRs Erik.

Cheers,
Bernard
Title: AA-filtering CCD and CMOS
Post by: TheSuede on October 03, 2009, 09:22:51 am
Hi all... :-)

I'd actually say that working without an AA-filter only makes your pictures more detailed (and sharper) in SOME cases. As Fraunhofer (and Kodak, and Dalsa, and several independant studies I've seen) have found, more complex targets with smaller detail or colour-driven detail more often than not get a lower dE ("perfect" original to raw-interpolated picture, pixel-per-pixel value deviation) when you shoot (or simulate) a sensor with an AA-filter and sharpen it right. This is quite logical if you know something about sampling theory and the Bayer CFA principle. Any one pixel HAS to have some relevant and statistically conformant relationship to the surrounding pixels if the interpolation of the two missing colours per pixel is to have any chance at all to do the right estimation. Most studies use the standard Kodak picture targets, software simulated as CFA filtered raws.

The absolutely best results in "mixed" real world picture targets seems to be a controlled 8% light spread in the X- and Y-axis (none in the 45º axises), this gives the raw-interpolation engine maximum detail to work with, with reasonably good relevance to the reality in front of the lens. What is interpreted as "detail" and "sharpness" in an AA-filter-less sensor package is actually mostly interpolation noise and high-frequency detail miscalculations. It has very little to do with real detail and sharpness - except in cases as simple as the Imatest slanted edge measurement - which is a very simple "black square on white background, how sharp is the outline"-test. No detail, no colour.

Easy target, no fine detail or fine colour detail - AA-less is better. ALL other cases - Controlled AA-filter with the right strength is better. Unless you're truly comfortable with the fact that the picture isn't conformant with reality. In Fraunhofer's study, the impression of "more detail" in the AA-filter less sensor could be totally negated by adding a very fine pitch noise to the sharpened AA-filtered version - then the filtered version was perceived as "more detailed", and STILL had better conformance with reality - pixel for pixel as measured by standard deviation from the "perfect" original.

I'm also very surprised that such a seasoned and knowledgeable person as Erwin gives the Leica 250% oversharpening and forgets to scale the resolutions evenly in the comparison graphs... :-) Note that the D3x resolutions are measured out to "2500", and the M9 uses the same graph width, but is stretched to make "1800" the maximum value out right in the graph.... Do a mental correction for the scale differences before you compare the graphs - then the story looks even more "definitely not in favour" of the M9... :-)


Quite a rant for a first post, but I was actually quite "not impressed" with Erwin's latest study - he is usually a VERY dependable and neutral person. Even if his entire salary comes from the Leica PR-department... :-)
Title: AA-filtering CCD and CMOS
Post by: aaykay on October 03, 2009, 09:49:49 am
Quote from: BernardLanguillier
I was speaking about Japanese DSLRs Erik.

Cheers,
Bernard

And I think that is precisely what is being talked about too.  DSLRs.  The  current Sony A300/350/330/380 along with some of the lower-end Nikon models (D40/D40X/D60), the previous D200 and all of the Pentax models prior to the recent ones employing the Samsung CMOS chip, used CCDs with AA filters.  

In all of their semi-pro and above models, Sony has completely transitioned into CMOS and only uses CCD in their lowest-end consumer models.  Same story with Nikon, where in the past only the D2X came with CMOS but now all of their semi-pro and Pro models have moved away from CCD into CMOS.  Pentax too have transitioned from CCD into CMOS but do retain some CCDs in their lowest-end consumer models.
Title: AA-filtering CCD and CMOS
Post by: ErikKaffehr on October 03, 2009, 09:59:58 am
Hi,

It may be helpful if you could give links to studies you refer to. Even without references I'd say that your comments put the AA-filtering issue in perspective. Regarding Erwin Puts's articles I have not noticed the 250% oversharpening. That said I often feel that data presented by Erwin Puts is not always easy to grasp or compare but I get the impression that he puts a lot of efforts into his tests.

To me it seems odd that DSLR makers would spend resources on quite expensive filters, something that would reduce image quality. It simply doesn't make any sense. Therefore I assume that they are employed for good reason.

Best regards
Erik

Quote from: TheSuede
Hi all... :-)

I'd actually say that working without an AA-filter only makes your pictures more detailed (and sharper) in SOME cases. As Fraunhofer (and Kodak, and Dalsa, and several independant studies I've seen) have found, more complex targets with smaller detail or colour-driven detail more often than not get a lower dE ("perfect" original to raw-interpolated picture, pixel-per-pixel value deviation) when you shoot (or simulate) a sensor with an AA-filter and sharpen it right. This is quite logical if you know something about sampling theory and the Bayer CFA principle. Any one pixel HAS to have some relevant and statistically conformant relationship to the surrounding pixels if the interpolation of the two missing colours per pixel is to have any chance at all to do the right estimation. Most studies use the standard Kodak picture targets, software simulated as CFA filtered raws.

The absolutely best results in "mixed" real world picture targets seems to be a controlled 8% light spread in the X- and Y-axis (none in the 45º axises), this gives the raw-interpolation engine maximum detail to work with, with reasonably good relevance to the reality in front of the lens. What is interpreted as "detail" and "sharpness" in an AA-filter-less sensor package is actually mostly interpolation noise and high-frequency detail miscalculations. It has very little to do with real detail and sharpness - except in cases as simple as the Imatest slanted edge measurement - which is a very simple "black square on white background, how sharp is the outline"-test. No detail, no colour.

Easy target, no fine detail or fine colour detail - AA-less is better. ALL other cases - Controlled AA-filter with the right strength is better. Unless you're truly comfortable with the fact that the picture isn't conformant with reality. In Fraunhofer's study, the impression of "more detail" in the AA-filter less sensor could be totally negated by adding a very fine pitch noise to the sharpened AA-filtered version - then the filtered version was perceived as "more detailed", and STILL had better conformance with reality - pixel for pixel as measured by standard deviation from the "perfect" original.

I'm also very surprised that such a seasoned and knowledgeable person as Erwin gives the Leica 250% oversharpening and forgets to scale the resolutions evenly in the comparison graphs... :-) Note that the D3x resolutions are measured out to "2500", and the M9 uses the same graph width, but is stretched to make "1800" the maximum value out right in the graph.... Do a mental correction for the scale differences before you compare the graphs - then the story looks even more "definitely not in favour" of the M9... :-)


Quite a rant for a first post, but I was actually quite "not impressed" with Erwin's latest study - he is usually a VERY dependable and neutral person. Even if his entire salary comes from the Leica PR-department... :-)
Title: AA-filtering CCD and CMOS
Post by: ErikKaffehr on October 03, 2009, 10:04:14 am
Yes,

Just would point out that normally a double layer of OLP are employed, one vertical and one horizontal. Some of the Pentax models seem to have only one OLP (either vertical or horizontal).

Best regards
Erik


Quote from: aaykay
And I think that is precisely what is being talked about too.  DSLRs.  The  current Sony A300/350/330/380 along with some of the lower-end Nikon models (D40/D40X/D60), the previous D200 and all of the Pentax models prior to the recent ones employing the Samsung CMOS chip, used CCDs with AA filters.  

In all of their semi-pro and above models, Sony has completely transitioned into CMOS and only uses CCD in their lowest-end consumer models.  Same story with Nikon, where in the past only the D2X came with CMOS but now all of their semi-pro and Pro models have moved away from CCD into CMOS.  Pentax too have transitioned from CCD into CMOS but do retain some CCDs in their lowest-end consumer models.
Title: AA-filtering CCD and CMOS
Post by: georgl on October 05, 2009, 01:04:37 pm
@ErikKaffehr
AA-filters are needed if you have to avoid moire at any cost. Press-photographers usually don't have the time to post-process the images to remove moire - they rather live the loss of fine detail/contrast.

Mr. Puts has made some interesting articles about lenses and optical design, this was also his last direct connection to Leica (a brochure in 2005?). But his articles about digital photography are quite often misleading and sometimes even wrong.

@Daniel Browning
The "look" of moire has definitely changed with the various sensor-generations (with increasing fill-rate) but alaising itself appears nevertheless and never seemed to influenced the choice of AA-filter vs. unfiltered!?

@TheSuede
The aspect of bayer-interpolation is quite interesting, but I'm not familiar with these studies.

But the basics of alaising and AA-filters remain unchanged, moire doesn't destroy detail and AA-filters don't preserve it.
Title: AA-filtering CCD and CMOS
Post by: BJL on October 05, 2009, 02:01:56 pm
Quote from: BernardLanguillier
Aren't all Japanese reflex CMOS nowadays?
No. Not the brand new Nikon D3000 for example, or this year's Sony A220, A330 and A380, or the Pentax K2000 (K-m) and K200D. The Sony 10MP CCD is the main one left is use though; five of the above six use it. I suspect that Sony is squeezing the last value out of its "paid for" DSLR CCD production lines.

Anyway, over the years there have been a great many Japanese DSLR's with CCD sensors, all with AA filters, so I see no connection between AA filter and CCD/CMOS.

It might have more to do with the high end customers (MF, Leica) leaning towards optimizing sharpness and such, and being willing to do some occasional PP on moiré to get the best IQ, whereas the mainstream is more likely to sacrifice the last few percent of sharpness for more consistently usable "out of the box" results.
Title: AA-filtering CCD and CMOS
Post by: BernardLanguillier on October 05, 2009, 07:09:30 pm
Quote from: ErikKaffehr
It may be helpful if you could give links to studies you refer to. Even without references I'd say that your comments put the AA-filtering issue in perspective.

Which is exactly what I was refering to above.

The opinion of the back/M8/M9 owners has mostly been formed based on information coming from companies doing AA filter less devices. Once you start to listen to both sides of the story, it becomes a lot less clear cut... and as we are seeing with this contribution, there are many people around who believe that the lack of AA filter has little value if any in actual detail rendering.

Cheers,
Bernard
Title: AA-filtering CCD and CMOS
Post by: BernardLanguillier on October 05, 2009, 07:24:01 pm
Quote from: georgl
@TheSuede
The aspect of bayer-interpolation is quite interesting, but I'm not familiar with these studies.

But the basics of alaising and AA-filters remain unchanged, moire doesn't destroy detail and AA-filters don't preserve it.

What is being said here is different. According to this contribution, AA filter less devices are intrinsically unable to capture the reality of a scene.

We have gotten used to the smooth fonts we see on our screens but sometimes fail to realize that the disapearance of stair cases on our display is the result of a voluntary blurring of their edges... another example where anti-aliasing is required to make a shape look the way it actually is beyond the limited spatial resolution of the digital information.

All in all, the over-simplications that are often proposed by one party to sell a technology are typically poor basis for judgement. The only option is to listen to ask the question to both parties and listen to both over-simplifications.

Another way is to look at 100% crops and make up one's mind based on the actual results delivered by the technology. Every time I do so, I end up preferring the way the correctly sharpened AA filter image looks.

Cheers,
Bernard
Title: AA-filtering CCD and CMOS
Post by: telyt on October 07, 2009, 07:45:27 pm
Quote from: BernardLanguillier
The opinion of the back/M8/M9 owners has mostly been formed based on information coming from companies doing AA filter less devices.

I'm not sure how you determined this.  The M8/M9/DMR owners I know of who prefer the AA filter-less approach formed their opinions by comparing output from the AA filter-less camera to their other, AA-filtered, cameras.
Title: AA-filtering CCD and CMOS
Post by: rljones on October 08, 2009, 07:58:09 pm
MaxMax (http://www.maxmax.com/hot_rod_visible.htm) offers a AA-filter removal service for several cameras. They've posted results from a Canon 5D.

They claim up to a 30% improvement in resolution. (However, any sensor cleaning features of the camera are lost due to the conversion process.)

It was my impression that the greater the sensor resolution, then moire is less of an issue (or, is this a function more of smaller pixel size, rather than simply more of them per unit area). if this is true, then removing the AA-filter from a 5D2 or a D3x might be useful.
Title: AA-filtering CCD and CMOS
Post by: jing q on October 09, 2009, 04:01:16 am
Quote from: rljones
MaxMax (http://www.maxmax.com/hot_rod_visible.htm) offers a AA-filter removal service for several cameras. They've posted results from a Canon 5D.

They claim up to a 30% improvement in resolution. (However, any sensor cleaning features of the camera are lost due to the conversion process.)

It was my impression that the greater the sensor resolution, then moire is less of an issue (or, is this a function more of smaller pixel size, rather than simply more of them per unit area). if this is true, then removing the AA-filter from a 5D2 or a D3x might be useful.

I removed the AA filter on my 5d MkII using Maxmax.
just to let you know they only remove one of two AA filters in the camera since one of the filters is connected directly to the sensor.

Anyway you will see a difference in tiny details especially in an architectural shot, I did one rcently where the amount of tiny detail was beyond what I expected, and yes, there was moire (which is really not a big problem here)

People mention sharpening as having a similar effect but what we're looking at here is more distinct definition of edges of objects in an image.
Sharpening helps abit but seems abit more muddy and flat.
Title: AA-filtering CCD and CMOS
Post by: BernardLanguillier on October 09, 2009, 09:42:30 am
Quote from: jing q
I removed the AA filter on my 5d MkII using Maxmax.
just to let you know they only remove one of two AA filters in the camera since one of the filters is connected directly to the sensor.

Anyway you will see a difference in tiny details especially in an architectural shot, I did one rcently where the amount of tiny detail was beyond what I expected, and yes, there was moire (which is really not a big problem here)

People mention sharpening as having a similar effect but what we're looking at here is more distinct definition of edges of objects in an image.
Sharpening helps abit but seems abit more muddy and flat.

Would it be possible for you to post before/after crops?

Thank you.

Cheers,
Bernard
Title: AA-filtering CCD and CMOS
Post by: jing q on October 09, 2009, 11:28:45 am
Quote from: BernardLanguillier
Would it be possible for you to post before/after crops?

Thank you.

Cheers,
Bernard

I'll post a shot from the modified camera, together with some moire. unfortunately don't have a comparison shot to show you, but it may give some insight into the effect of having an AA filter removed

Here are some 100% crops.

Sharpening Amount 0, Radius 1.0, Detail 0
[attachment=17065:wide00.JPG]


Sharpening Amount 25, Radius 1.0, Detail 0
[attachment=17066:wide25.JPG]


Sharpening Amount 50, Radius 1.0, Detail 0
[attachment=17067:wide50.JPG]

Sharpening Amount 50, Radius 1.0, Detail 15
[attachment=17068:wide5015.JPG]

Title: AA-filtering CCD and CMOS
Post by: ejmartin on October 09, 2009, 12:32:10 pm
Quote from: rljones
MaxMax (http://www.maxmax.com/hot_rod_visible.htm) offers a AA-filter removal service for several cameras. They've posted results from a Canon 5D.

They claim up to a 30% improvement in resolution. (However, any sensor cleaning features of the camera are lost due to the conversion process.)

It was my impression that the greater the sensor resolution, then moire is less of an issue (or, is this a function more of smaller pixel size, rather than simply more of them per unit area). if this is true, then removing the AA-filter from a 5D2 or a D3x might be useful.

Translation of marketing-speak: 30% increase in aliasing.

The AA filter strength is tuned to the pixel size, one only needs to blur the image over the size of a Bayer RGGB quartet in order to damp the aliasing.  As the pixels get smaller, the needed blur gets smaller.  In addition, the effects of lens aberrations and diffraction are fixed, and can do the job of blurring if the pixels are small enough.  So as pixels get smaller, the AA filter can be even weaker since the optics is doing part of the blurring already.  My understanding is that most current P&S cameras have no AA filter for this reason -- diffraction already provides sufficient blur.  Current high-res DSLR's are not yet in the territory that allows removal of the AA filter without substantial aliasing effects.
Title: AA-filtering CCD and CMOS
Post by: Ray on October 09, 2009, 12:36:59 pm
Quote from: jing q
I'll post a shot from the modified camera, together with some moire. unfortunately don't have a comparison shot to show you, but it may give some insight into the effect of having an AA filter removed


No! Comparisons are all that count when pixel-peeping. No comparisons, no conclusions can be drawn.
Title: AA-filtering CCD and CMOS
Post by: jing q on October 09, 2009, 12:37:02 pm
this is the file where the crop is from. I can't remember the settings for this. But it'll give you an idea of how small the crop is relative to the whole 21mp picture.

things I noticed after modifying my 5D mkII:

edges are more defined (and CA is more visible also)
abit more subtle detail overall

[attachment=17069:wide1_smalla.jpg]
Title: AA-filtering CCD and CMOS
Post by: jing q on October 09, 2009, 12:42:22 pm
Quote from: Ray
No! Comparisons are all that count when pixel-peeping. No comparisons, no conclusions can be drawn.

Ray...
I don't owe you anything. I'm posting this stuff as a general favour. And i'm sharing my experience. If you don't like it, pay someone to do the testing for you, or pay to get it done yourself.
Title: AA-filtering CCD and CMOS
Post by: rljones on October 09, 2009, 07:08:05 pm
thanks jing for posting the images.
Title: AA-filtering CCD and CMOS
Post by: BernardLanguillier on October 10, 2009, 06:15:18 am
Quote from: jing q
I'll post a shot from the modified camera, together with some moire. unfortunately don't have a comparison shot to show you, but it may give some insight into the effect of having an AA filter removed

Here are some 100% crops.

Thanks a lot for the crops.

I am honestly not too impressed by these crops, I feel that correct sharpening applied to d3x files results in both sharper and more pleasing images. This could be depending on the lens though, 90% or my images are shot with the Zeiss 100mm f2.0 that is really sharp.

Cheers,
Bernard
Title: AA-filtering CCD and CMOS
Post by: ErikKaffehr on October 10, 2009, 09:57:46 am
Hi,

It may be my view that all pictures need som more agressive sharpening. In Bruce Fraser's book on sharpening there was an example that a 6MP camera sans AA-filter would be sharpened amount=240 and radius=0.6, while a 6MP camera with AA filter would need around amount=500. For > 11MP cameras he suggested radius=0.4.

You could try to sharpen at 0.4 and say 300%

Best regards
Erik

Quote from: jing q
I'll post a shot from the modified camera, together with some moire. unfortunately don't have a comparison shot to show you, but it may give some insight into the effect of having an AA filter removed

Here are some 100% crops.

Sharpening Amount 0, Radius 1.0, Detail 0
[attachment=17065:wide00.JPG]


Sharpening Amount 25, Radius 1.0, Detail 0
[attachment=17066:wide25.JPG]


Sharpening Amount 50, Radius 1.0, Detail 0
[attachment=17067:wide50.JPG]

Sharpening Amount 50, Radius 1.0, Detail 15
[attachment=17068:wide5015.JPG]
Title: AA-filtering CCD and CMOS
Post by: ErikKaffehr on October 10, 2009, 09:59:55 am
Hi,

I don't think they are correctly sharpened.

Best regards
Erik

Quote from: BernardLanguillier
Thanks a lot for the crops.

I am honestly not too impressed by these crops, I feel that correct sharpening applied to d3x files results in both sharper and more pleasing images. This could be depending on the lens though, 90% or my images are shot with the Zeiss 100mm f2.0 that is really sharp.

Cheers,
Bernard
Title: AA-filtering CCD and CMOS
Post by: bjanes on October 10, 2009, 11:18:40 am
Quote from: ErikKaffehr
It may be my view that all pictures need som more agressive sharpening. In Bruce Fraser's book on sharpening there was an example that a 6MP camera sans AA-filter would be sharpened amount=240 and radius=0.6, while a 6MP camera with AA filter would need around amount=500. For > 11MP cameras he suggested radius=0.4.

You could try to sharpen at 0.4 and say 300%

Best regards
Erik

Yes, in the original sharpening book, Bruce did capture sharpening in two steps. The first step involved global application (with blend if sliders to protect the shadows and highlights) of the unsharp filter. The radius was determined by the megapixel count of the camera and the amount by the strength of the blur filter. The next step was to sharpen for image content (i.e low frequency or high frequency images) using an edge mask. The final step was output sharpening for screen or print output.

A very interesting Zeiss article (http://www.zeiss.com/C12567A8003B8B6F/EmbedTitelIntern/CLN_30_MTF_en/$File/CLN_MTF_Kurven_EN.pdf) on MTF showed point spread functions (PSPs) for a 12 MP full frame digital camera (likely the Nikon D3). The white square represents the pixel size. Coma and other aberrations are shown. Image 7 represents an ideal PSF and image 8 shows the effect of a blur filter. The effect of the blur filter appears rather nasty, but one should remember that the effective pixel size of a Bayer array camera approaches twice the pixel size, since demosaicing involves interpolation from a 2x2 pixel array.

[attachment=17107:ZeissPSFs.png]

Sharpening is critically important for cameras with a blur filter. Rather than using the unsharp mask for source sharpening, some photographers use a deconvolution filter such as Focus Magic. Deconvolution is theoretically attractive since it actually removes blur rather than merely improving edge contrast as with the unsharp mask. However, deconvolution requires that one derive a PSP describing how the blur was introduced. In astronomy when one is dealing with point sources, this is relatively simple. However, for normal photography deriving the PSP is more difficult. The Photoshop Smart Sharpen filter is another deconvolution method and it might also be tried. However, one should still sharpen for image content and output. It is not clear if these steps were done with the posted images. BTW, setting the black point improves the images considerably.
Title: AA-filtering CCD and CMOS
Post by: jing q on October 10, 2009, 11:33:44 am
Quote from: BernardLanguillier
Thanks a lot for the crops.

I am honestly not too impressed by these crops, I feel that correct sharpening applied to d3x files results in both sharper and more pleasing images. This could be depending on the lens though, 90% or my images are shot with the Zeiss 100mm f2.0 that is really sharp.

Cheers,
Bernard

heh this wasn't shot with the sharpest lens. it's a 24-70mm.but I was just posting this as an example.
you wouldn't be able to get better photos with your 100mm because this was shot with a wide angle.

it's not a comparison either so it's not a lab test,it's just a personal opinion, from someone who uses dslrs and medium format backs and scanned film of small to large format, who prints up to 68" for exhibitions.
what i'm saying is that this level of detail is not what I would have expected in the past from dslrs if there was an AA filter present

And what I am saying is that there is value for people (a niche perhaps) not to have AA filters on cameras, and that the doom and gloom about moire is overrated.
It's no big deal removing the AA filter, and I like the quality of files coming out from my camera after having the AA filter removed.
quoting you,
'there are many people around who believe that the lack of AA filter has little value if any in actual detail rendering."

and i would like to respond that there are people around who believe that the lack of AA filter HAS value.
and they aren't just leica users who need to justify their costs.
Title: AA-filtering CCD and CMOS
Post by: ErikKaffehr on October 10, 2009, 01:10:42 pm
Hi,

I'm aware of the Zeiss article. Just a couple of points:

1) The SPF in the Zeiss article seems to contain lightbeams in 45 degree angles, I'd expect 90 degrees. That said I have no expertise in this area.
2) Focus Fixer cares about camera used, maybee they have some kind of PSF for AA-filter

Looking at the stuff from the practical side, Focus Magic doesn't work on Intel Macs if not using Rosetta. I have tried Focus Magic, Focus Fixer and Smart Sharpen in PS and could not really say that one is preferable over the other.

Best regards
Erik




Quote from: bjanes
Yes, in the original sharpening book, Bruce did capture sharpening in two steps. The first step involved global application (with blend if sliders to protect the shadows and highlights) of the unsharp filter. The radius was determined by the megapixel count of the camera and the amount by the strength of the blur filter. The next step was to sharpen for image content (i.e low frequency or high frequency images) using an edge mask. The final step was output sharpening for screen or print output.

A very interesting Zeiss article (http://www.zeiss.com/C12567A8003B8B6F/EmbedTitelIntern/CLN_30_MTF_en/$File/CLN_MTF_Kurven_EN.pdf) on MTF showed point spread functions (PSPs) for a 12 MP full frame digital camera (likely the Nikon D3). The white square represents the pixel size. Coma and other aberrations are shown. Image 7 represents an ideal PSF and image 8 shows the effect of a blur filter. The effect of the blur filter appears rather nasty, but one should remember that the effective pixel size of a Bayer array camera approaches twice the pixel size, since demosaicing involves interpolation from a 2x2 pixel array.

[attachment=17107:ZeissPSFs.png]

Sharpening is critically important for cameras with a blur filter. Rather than using the unsharp mask for source sharpening, some photographers use a deconvolution filter such as Focus Magic. Deconvolution is theoretically attractive since it actually removes blur rather than merely improving edge contrast as with the unsharp mask. However, deconvolution requires that one derive a PSP describing how the blur was introduced. In astronomy when one is dealing with point sources, this is relatively simple. However, for normal photography deriving the PSP is more difficult. The Photoshop Smart Sharpen filter is another deconvolution method and it might also be tried. However, one should still sharpen for image content and output. It is not clear if these steps were done with the posted images. BTW, setting the black point improves the images considerably.
Title: AA-filtering CCD and CMOS
Post by: aaykay on October 10, 2009, 04:01:47 pm
Quote from: BJL
No. Not the brand new Nikon D3000 for example, or this year's Sony A220, A330 and A380, or the Pentax K2000 (K-m) and K200D. The Sony 10MP CCD is the main one left is use though; five of the above six use it. I suspect that Sony is squeezing the last value out of its "paid for" DSLR CCD production lines.

The "paid for CCD production lines" of Sony Semiconductor Kyushu, also make ultra-high-end CCD sensors for their video/cine-alta line.  The Sony F35 Cine-alta, where the bare-bones camera costs $250,000 - not including the cost of a single accessory or lens - comes with a 36x24mm 35mm sized CCD sensor.   They have several other models in the Cine-Alta range that also use CCDs exclusively.  Thus Sony will not only "squeeze the last value" but will continue to promote that line for a variety of other applications that may or may not be  DSLR related.  

There have also been rumors about an upcoming 35mm Studio camera with a CCD-sensor from the Sony stable, with around 40MP.  Whether true or not, I have no confirmation but being Sony, I would not rule out anything.  It is well known that for low-ISO applications where shooting speed is not an overriding consideration, CCDs are the preferred sensor type since they don't have areas on the sensor taken up by the secondary circuitry, like CMOS sensors and thus have a lot more of their area dedicated to gathering light, which in turn translates into better quality images (at least at low ISOs and well lit studio conditions).




Title: AA-filtering CCD and CMOS
Post by: Daniel Browning on October 10, 2009, 10:33:51 pm
Great post, aaykay.

Quote from: aaykay
The Sony F35 Cine-alta, where the bare-bones camera costs $250,000 - not including the cost of a single accessory or lens - comes with a 36x24mm 35mm sized CCD sensor.

Minor correction: it's actually 23.6x13.3mm (active area). In cinema, the film runs through the camera "hot dog way" (vertically). Still photographers run film through the camera "hamburger way" (horizontally). They both call it "35mm", but one is much bigger than the other.

The F35 sensor (same as the Panavision Genesis) is interesting for a variety of reasons, and not just its astronomical cost. It has 12 megapixels, in an RGB array (not Bayer), with alternating rows having an ND filter. All 6 pixels are then combined in a proprietary process to generate one single output pixel with high dynamic range (thanks to the ND). So it only puts out 2 MP, which is the maximum that could be utilized by the highly expensive Sony tape system anyway. While I would have expected the design to make it possible for very little aliasing (chroma or luma), the unfortunate reality is that the F35 has a much higher degree of aliasing than its contemporaries.
Title: AA-filtering CCD and CMOS
Post by: aaykay on October 11, 2009, 11:20:04 am
Quote from: Daniel Browning
Great post, aaykay.



Minor correction: it's actually 23.6x13.3mm (active area). In cinema, the film runs through the camera "hot dog way" (vertically). Still photographers run film through the camera "hamburger way" (horizontally). They both call it "35mm", but one is much bigger than the other.

The F35 sensor (same as the Panavision Genesis) is interesting for a variety of reasons, and not just its astronomical cost. It has 12 megapixels, in an RGB array (not Bayer), with alternating rows having an ND filter. All 6 pixels are then combined in a proprietary process to generate one single output pixel with high dynamic range (thanks to the ND). So it only puts out 2 MP, which is the maximum that could be utilized by the highly expensive Sony tape system anyway. While I would have expected the design to make it possible for very little aliasing (chroma or luma), the unfortunate reality is that the F35 has a much higher degree of aliasing than its contemporaries.


Cool stuff.  I did not know some of these finer points and some of the Film vs Still differences.  Thanks for clarifying.  So is such an RGB sensor design adaptable for (future) dSLRs, than employing Bayer designs ?  Is this similar to the Foveon RGB design ?
Title: AA-filtering CCD and CMOS
Post by: Daniel Browning on October 11, 2009, 12:05:18 pm
Quote from: aaykay
So is such an RGB sensor design adaptable for (future) dSLRs, than employing Bayer designs?

It's possible, but not a good idea. It only made sense for Sony because their tape system cannot record over 2 MP. Our flash cards can record almost any resolution because we have much lower frame rates in still photography, so it does not make sense for us to throw away so much resolution.

Quote from: aaykay
Is this similar to the Foveon RGB design ?

No, Foveon takes 3 samples at a single location,  whereas the F35 takes 6 samples at 6 separate locations, then combines them into 1.
Title: AA-filtering CCD and CMOS
Post by: thierrylegros396 on October 12, 2009, 03:27:56 am
A better idea is to use double read CCD at different polarization to obtain 2 files for HDR.

Almost instantaneously so that you can shoot HDR without tripod !!!

Don't remember where I found that, but they are projects of some manufacturers.

Have a Nice Day.

Thierry
Title: AA-filtering CCD and CMOS
Post by: BernardLanguillier on October 12, 2009, 08:24:03 am
Quote from: jing q
and i would like to respond that there are people around who believe that the lack of AA filter HAS value.
and they aren't just leica users who need to justify their costs.

I want to believe in that, and wouldn't hesitate to have the AA filter of my d3x removed, but still haven't seen any samples proving the point.

Cheers,
Bernard
Title: AA-filtering CCD and CMOS
Post by: ErikKaffehr on October 12, 2009, 12:58:41 pm
Hi Bernard,

The question is really if removing the the AA-filter actually increases resolution or just reduces the need of sharpening. So the question is if we can just increase sharpening to compemsate for AA-filtering. It seems that if remove the AA-filter we will have same artifacts, colorful or not, but it seems no one cares about the artifacts as long as they are monochrome.

That helicopter picture of yours was impressive, how was it processed?

Best regards
Erik



Quote from: BernardLanguillier
I want to believe in that, and wouldn't hesitate to have the AA filter of my d3x removed, but still haven't seen any samples proving the point.

Cheers,
Bernard
Title: AA-filtering CCD and CMOS
Post by: aaykay on October 12, 2009, 02:24:59 pm
Quote from: BernardLanguillier
I want to believe in that, and wouldn't hesitate to have the AA filter of my d3x removed, but still haven't seen any samples proving the point.

Cheers,
Bernard

I believe there was a company mentioned eariler, that removes AA filters for a charge, right ?  Would they not have before and after samples ?  What if we write to them or call them up ?
Title: AA-filtering CCD and CMOS
Post by: ErikKaffehr on October 12, 2009, 02:39:39 pm
Hi,

They have. Problem is that they are JPEG. I downloaded a couple of their samples and increased sharpening on the AA-filtered image and could not see a significant difference.

Here is my posting on the issue: http://luminous-landscape.com/forum/index....st&p=144346 (http://luminous-landscape.com/forum/index.php?s=&showtopic=20081&view=findpost&p=144346)

and here is one from Christopher: http://luminous-landscape.com/forum/index....st&p=145485 (http://luminous-landscape.com/forum/index.php?s=&showtopic=20081&view=findpost&p=145485)

Both have samples

Best regards
Erik

Quote from: aaykay
I believe there was a company mentioned eariler, that removes AA filters for a charge, right ?  Would they not have before and after samples ?  What if we write to them or call them up ?
Title: AA-filtering CCD and CMOS
Post by: Plekto on October 12, 2009, 07:06:06 pm
No AA filter in raw is always noticeable.   But you will get moires and jaggies unless you are using a Fuji or Foveon sensor or something without a Bayer pattern (or that technically does bracketing and blending in-camera)

Note - bracketing and blending half or even a quarter of a stop apart will make artifacts disappear.  It's a cheap and dirty method to get clean results if you are shooting something that isn't moving.  In this case, the AA filter isn't even needed.  But... finding a camera without one is essentially impossible these days as they cater to consumers instead of pros these days.
Title: AA-filtering CCD and CMOS
Post by: BernardLanguillier on October 12, 2009, 07:06:42 pm
Quote from: ErikKaffehr
That helicopter picture of yours was impressive, how was it processed?

Erik,

It was converted in C1 Pro with capture sharpening set at something like 400, 0.6 if I recall (my typical capture sharpening), works great on the d3x files at low ISO since there simply no noise from highlights to shadows.

I may have added a bit of sharpening in PS, not too sure what I did, but anyway this sample it representative of what I get with the d3x when using good lenses like the Zeiss 100mm f2.0 or the 300 f2.8 VR.

I keep thinking that it looks better when viewed at 100% than all the other AA filter less samples I have seen, excluding perhaps some of the P65+ samples.

Cheers,
Bernard
Title: AA-filtering CCD and CMOS
Post by: Daniel Browning on October 12, 2009, 07:38:49 pm
George,

You made a post last week, but I didn't notice it until today.

Quote from: georgl
@Daniel Browning
The "look" of moire has definitely changed with the various sensor-generations (with increasing fill-rate) but alaising itself appears nevertheless and never seemed to influenced the choice of AA-filter vs. unfiltered!?

I don't understand if you're rephrasing my statement in the form of a question or making your own statement in the form of a comment, but perhaps it will help if I just add a clarification. My position is that aliasing characteristics are dependent on the optical fill factor (with microlenses considered), not the electronic fill factor (of the bare sensor with no microlenses). I think that microlenses have given CMOS optical fill factors comparable to CCD for at least several years; therefore it has not influenced the choice to use OLPF or not.

Quote from: georgl
AA-filters are needed if you have to avoid moire at any cost. Press-photographers usually don't have the time to post-process the images to remove moire - they rather live the loss of fine detail/contrast.

First of all, there is no post-processing that can truly remove moire. The best software (C1P IMHO) can only smear the moire into the surrounding detail.

Second, moire is only the worst and most offensive aliasing artifact. But there are many other aliasing artifacts that I find displeasing and unnatural, including jaggies, stair-stepping, sparkling, "snap to grid", wavy lines, bands, fringing, popping, strobing, noise, and false detail. These, too, are impossible to remove through any automated or semi-automated software process.

In real life, when you pour two liters of water into a one liter container, water spills out and makes a mess. But camera design is different: when you pour two liters of water into a one liter container, the water folds back on itself and corrupts the entire container. The amount of water is the level detail (spatial frequency), and the volume of the container is the number of megapixels in the camera. Aliasing is the corruption. Anti-aliasing filters reduce detail down to a level that can fit within the pixel resolution.

But of course, that's just me personally. Two people can look at the exact same image and each see something different. Take an aliased image for example. Where one sees overly harsh and sudden transitions from black to white in just 2 pixels, another sees microcontrast. Where one gets the impression of fakeness, another gets the feeling of sharpness. Where one is jarred by the conformation of small details into a slightly different location than they exist in nature, another is awestruck by the high acuity.

Same thing with anti-aliased images. Where one sees slow, smooth, and careful transitions from black to white, over 3 or more pixels, another sees mushy detail. One gets the impression of natural, life-like renditions, another gets the feeling of haze and low contrast.

So you may percieve an unfiltered image as having high microcontrast, sharpness, and acuity; while I see the same one as harsh and unnatural. The OLPF'd images you perceive as mushy, haze, lowcon images are, to me, smooth and natural. So it would be difficult for us to come to agreement on how big of an issue aliasing is.

I think part of the reason why some manufacturers (e.g. MFDB) exclude OLPF is cost. A good OLPF is lab-grown, high-grade, ground, and polished Lithium Niobate crystal. The cost scales exponentially with area because even the tiniest defect will show up on the image, thanks to being so close to the sensor. Even though MFDB is only three times more area than 35mm, the cost can be an order of magnitude (or more) higher. More importantly, Canon/Nikon ship millions of units a year, compared to less than 6,000/year for all MFDB combined, so economies of scale is a huge factor.

If camera manufacturers were shrewd enough, they would stop wasting so much money and effort to fight it. It's not like they get appreciated for it, they're more often lambasted (IMHO). I think the market of people that like aliasing is large enough. But since I happen to belong to the smaller part that prefers anti-aliased images, I'm glad they haven't (yet) given up on the expensive AA filters, and hope they never do. (At least until we start hitting diffraction cutoff frequencies at faster f-numbers.)

Quote from: georgl
But the basics of alaising and AA-filters remain unchanged, moire doesn't destroy detail and AA-filters don't preserve it.

I kindly disagree. For my taste, the AA-filtered image is better, because although the detail is low contrast, at least it's real. The detail in an unfiltered image is very high contrast, but it's false detail. But I understand that others have different personal preferences.
Title: AA-filtering CCD and CMOS
Post by: Plekto on October 12, 2009, 10:56:11 pm
This last comment is important.  The reality is that high contrast(actual rather than artificial as you noted) and good definition is often of far more importance than absolute sharpness.
Title: AA-filtering CCD and CMOS
Post by: jing q on October 12, 2009, 11:31:06 pm
Quote from: BernardLanguillier
I want to believe in that, and wouldn't hesitate to have the AA filter of my d3x removed, but still haven't seen any samples proving the point.

Cheers,
Bernard

I think if you were truly interested in that you would have seeked out some AA filterless images and tested it out yourself.
I'm not the only person who removed their AA filter however it's interesting to see a lot of people who don't use AA filterless cameras coming out of the woodworks to slam the idea of removing the AA filter everytime this topic is mentioned.

MR himself seems to like his images without AA filters, maybe these AA unfiltered images are not quite as terrible an idea as everyone makes them out to be?

just sayin' you know.
Title: AA-filtering CCD and CMOS
Post by: jing q on October 13, 2009, 12:44:24 am
Quote from: aaykay
I believe there was a company mentioned eariler, that removes AA filters for a charge, right ?  Would they not have before and after samples ?  What if we write to them or call them up ?

the company is Maxmax.
I think it is a good idea to call them up to post new samples, or maybe they should lend a few cameras to a few people to try out and give their opinion.
I guess marketing isn't their strong point.
Title: AA-filtering CCD and CMOS
Post by: BernardLanguillier on October 13, 2009, 04:25:56 am
Quote from: jing q
I think if you were truly interested in that you would have seeked out some AA filterless images and tested it out yourself.
I'm not the only person who removed their AA filter however it's interesting to see a lot of people who don't use AA filterless cameras coming out of the woodworks to slam the idea of removing the AA filter everytime this topic is mentioned.

I did look, and never liked what I saw, but I respect the opinion of those tho think/see different.

I have obviously not seen images from all possible cameras with/without AA filter, and I believe that the gain will vary from make to make, which is why I was interested in your samples.

Cheers,
Bernard



Title: AA-filtering CCD and CMOS
Post by: telyt on October 13, 2009, 07:58:13 am
Show me the artifacts:

(http://www.wildlightphoto.com/birds/accipitridae/feha02.jpg)

(http://www.wildlightphoto.com/mammals/lagomorphs/btha04.jpg)

Both photos: Leica R8 with DMR digital back (10 MP, no AA filter), 280mm f/4 APO-Telyt-R

I don't see any artifacts in 16" x 24" prints.
Title: AA-filtering CCD and CMOS
Post by: ejmartin on October 13, 2009, 08:18:16 am
Quote from: telyt
Show me the artifacts:

http://www.wildlightphoto.com/birds/accipitridae/feha02.jpg (http://www.wildlightphoto.com/birds/accipitridae/feha02.jpg)

http://www.wildlightphoto.com/mammals/lagomorphs/btha04.jpg (http://www.wildlightphoto.com/mammals/lagomorphs/btha04.jpg)

Both photos: Leica R8 with DMR digital back (10 MP, no AA filter), 280mmm f/4 APO-Telyt-R

Hard to show any artifacts when they have been eliminated by downsampling.

Quote
I don't see any artifacts in 16" x 24" prints.

Do you know what you're looking for?
Title: AA-filtering CCD and CMOS
Post by: telyt on October 13, 2009, 08:41:54 am
Quote from: ejmartin
Hard to show any artifacts when they have been eliminated by downsampling.

Buy a print and see for yourself.  Contact Appel Gallery in Sacramento.

Quote from: ejmartin
Do you know what you're looking for?

Yes I do.  So does the gallery owner.
Title: AA-filtering CCD and CMOS
Post by: michael on October 13, 2009, 09:07:15 am
I can't believe the amount of misinformation and disinformation that this topic has engendered.

Has the fact that every medium format back since Noah has not had an AA filter escaped anyone's notice? Do those that think that all or even most images taken without an AA have aliasing and artifacting think that the countless pros and highly critical photographers who use MF backs would tolerate this if it were true? Come on!

The reality is that aliasing is only visible in "some" images, and unless you shoot fabrics for a living, it's a very small number.

The advantage of not having an AA is clearly visible to anyone that has a good eye. In some cases it simply jumps out at you, image after image. In most cases its visible even in small prints and on screen as a type of increased clarity, as if a veil were lifted.

And as for sharpening compensating for having an AA, what are you thinking? Sharpening is about edge sharpness; AKA accutance. Removing an AA filter is about increasing resolution. Yes, adding USM increases apparent sharpness, but not real resolution. Two different things, though related in the real world.

Bumble bees can fly, though aeronautical engineers used to say that it was impossible. They just didn't bother asking the bees. And yes, removing the AA filter on a properly designed system does increase both real and apparent resolution. Talk to the engineers that design these systems, and also just trust your own eyes.

And as for removing moire in software – don't be too sure it can't be done. I'm working with one company now that is likely to have something exciting in this area in the months ahead.  

Michael
Title: AA-filtering CCD and CMOS
Post by: telyt on October 13, 2009, 09:24:51 am
Quote from: Daniel Browning
For my taste, the AA-filtered image is better, because although the detail is low contrast, at least it's real.

Nonsense.  Almost nothing in a photograph is real.  All that matters is that it's believable.

Common artifacts we take for granted:

film grain
digital noise
2-dimesions
limited DOF
a moment in time
limited field of view
lens aberations

AA blurring is an artifact too.  Pick the one(s) you're willing to live with
Title: AA-filtering CCD and CMOS
Post by: thierrylegros396 on October 13, 2009, 09:40:43 am
Quote from: michael
And as for sharpening compensating for having an AA, what are you thinking? Sharpening is about edge sharpness; AKA accutance. Removing an AA filter is about increasing resolution. Yes, adding USM increases apparent sharpness, but not real resolution. Two different things, though related in the real world.

just trust your own eyes.

Michael

For those who have knowledge in signal transmission, sharpening can be compared to equalization used in high speed data line to recover the signal !

It's not perfect, but it works.

So yes, the result is the only real important thing !

But in some very rare instance, aliasing cannot be eliminated.

In real world however, bandwidth is limited enough by the lens.

Have a Nice Day !

Thierry
Title: AA-filtering CCD and CMOS
Post by: ejmartin on October 13, 2009, 10:52:28 am
Quote from: michael
I can't believe the amount of misinformation and disinformation that this topic has engendered.

I think what you're seeing is two groups of people talking past one another.  One group understands the technical aspects of signal processing, and what aliasing is in a precise quantitative sense; another group has a rough idea of what aliasing is in terms of a set of experiences and examples of converted RAW images.  

It's a bit like the use of the term "dynamic range".  There is a precise engineering definition of DR (max signal divided by noise without signal), and a working photographer's definition of DR (max range of signal with 'acceptable' S/N ratio).  The difference between the two (the working photographer's version is typically several stops lower) causes endless heated 'discussions'.

I don't want to derail this thread into a foodfight on DR, I merely brought it up as another instance where different usages of the same terminology lead to each side misunderstanding the other's point of view.

Quote
Has the fact that every medium format back since Noah has not had an AA filter escaped anyone's notice? Do those that think that all or even most images taken without an AA have aliasing and artifacting think that the countless pros and highly critical photographers who use MF backs would tolerate this if it were true? Come on!

The reality is that aliasing is only visible in "some" images, and unless you shoot fabrics for a living, it's a very small number.

The advantage of not having an AA is clearly visible to anyone that has a good eye. In some cases it simply jumps out at you, image after image. In most cases its visible even in small prints and on screen as a type of increased clarity, as if a veil were lifted.

Actually, all that "pop" and "sizzle" of sharp edges in converted images from an AA-less camera is largely aliasing.  Aliasing is not just moire.

What the "signal processing camp" is saying is that, in terms of fidelity to the scene being imaged, a properly processed image from a sensor with an AA filter yields a more faithful digitization of the scene being captured than the same scene captured with a sensor without an AA filter.

Now, it may be that accurate image capture is not the goal; many people like all the pop and sizzle that aliasing provides, just as many people prefer oversaturated images.  What the "end user" camp is saying is that they like the effect of aliasing, so long as it doesn't manifest itself in forms objectionable to them such as moire.

Quote
And as for sharpening compensating for having an AA, what are you thinking? Sharpening is about edge sharpness; AKA accutance. Removing an AA filter is about increasing resolution. Yes, adding USM increases apparent sharpness, but not real resolution. Two different things, though related in the real world.

The DPReview report on the DP2 provides a good example of the increased "resolution" provided by the lack of an AA filter:

http://www.dpreview.com/reviews/sigmadp2/page22.asp (http://www.dpreview.com/reviews/sigmadp2/page22.asp)

Count the number of lines at one end of the DP2 resolution pattern and compare it to the other end, for any orientation.  The pattern looks nice and sharp, and if you were just to look at the pattern from say 22 to 24 on the chart you would say that the camera is resolving all the way out to the limit of the chart.  But actually it is just aliasing a more finely spaced set of lines into a more coarsely spaced set of lines; it is not actually resolving at that level.  Aliasing definitely helps the accutance though  
Title: AA-filtering CCD and CMOS
Post by: BJL on October 13, 2009, 11:18:26 am
Michael, I agree with most of what you say about the pros and cons of AA filters (in a nutshell I would say that the higher the pixel count, the better the argument for omitting the AA filter)

But I have to object to yet another statement of the following modern myth, which is an insult to many of my scientific friends and colleagues who work in the area of aerodynamics:
Quote from: michael
Bumble bees can fly, though aeronautical engineers used to say that it was impossible.
Actually what the aeronautical engineer said is completely true and very useful information, which sadly has been corrupted into an anti-scientific myth, often trotted out when people wish to dismiss scientific conclusions that they do not wish to accept.

What was discovered was that a bee would not be able to fly, if its wings were rigid, instead of flexing during flight as they in fact do. The flexing of a bee's wings during flight adds significantly to their propulsional effectiveness. This is good science; both true (experimentally confirmed) and useful: for example it is relevant to the design of helicopter rotors, making use of flexibility.


I have to wonder about the many people who believe this myth of the incompetence of aeronautical engineers ... and yet happily fly in aircraft that those "incompetents" help to design!
Title: AA-filtering CCD and CMOS
Post by: thierrylegros396 on October 13, 2009, 01:12:57 pm
Same problem occurs with digital audio !

Do you prefer sharp edge high frequencies from your tweeters or softer ones ?!

But the problem is more important when you have clipping !

The sound can become really crappy.

Link to The Nyquist Theorem tells us that we can successfully sample and play back frequency components up to one-half the sampling frequency. (http://www.earlevel.com/Digital%20Audio/Aliasing.html)

Have a Nice Day.

Thierry
Title: AA-filtering CCD and CMOS
Post by: georgl on October 13, 2009, 02:33:11 pm
The fact is that without real scientific measurements under controlled conditions we cannot say for sure what's actually "better" or "worse" - we not only compare filtered vs. unfiltered images, we compare different processing, different sensors, different lenses and slightly different shooting conditions (like the precise alignment of the sensor grid and photographic detail). I was surprised to see so unsharp unfiltered samples from the 5D which really show barely any more detail than the filtered version - it seems that the lens/focusing limited the IQ already.

The Mamiya ZD was available with and without OLPF - it would be the best "measurement" system we could get hands on.

What we see in these non-scientific "comparisons" are superior (besides absolute image size) MFDB-results. It's not just the AA-filter it's various aspects. But these MFDBs are used for the most demanding photographic situations for a good reason. In the real world, the artifacts which are cannot be handled by post-processing in comparison to the artifacts still left in AA-filtered images are usually negligible.  But every AA-filter reduces contrast below Nyqist and therefore information - that's a scientific fact. All of our bayer-based imaging systems are far from being perfect, most of the image is interpolated! We're talking about poor compromises on both sides anyway.

@Daniel Browning
I'm sorry that I chose my words imprecisely - of course moire or alaising cannot be removed, just as sharpening cannot retain any information. But it can be supressed and it seems that those processed images usually still contain more "real information" than AA-filtered images.
As I mentioned, the Mamiya ZD was available with an OLPF and all MFDBs contain highly specialized technologies (microlenses, cover glasses, sensor technology) - AA-filters aren't different. They chose not to use AA-filters and Pros (and most of them know AA-filtered systems) seem to be happy about that.
Title: AA-filtering CCD and CMOS
Post by: NikoJorj on October 13, 2009, 02:59:03 pm
Quote from: thierrylegros396
Same problem occurs with digital audio !
Sir yes Sir, but with a caveat (at least to my taste) : I'd think human perception is (much) more sensitive to high frequencies in sound, than to those in images.
Eg, the MP3 compression (that among others discards those harmonics we're not supposed to hear at, or am I wrong?) is more evident than the JPEG one.
Title: AA-filtering CCD and CMOS
Post by: Plekto on October 13, 2009, 05:12:48 pm
Quote
http://www.dpreview.com/reviews/sigmadp2/page22.asp (http://www.dpreview.com/reviews/sigmadp2/page22.asp)

Count the number of lines at one end of the DP2 resolution pattern and compare it to the other end, for any orientation.

Here is what they do with each image:

    *  Load RAW file into Adobe Camera RAW (Auto mode disabled)
    * Set Sharpness to 0 (all other settings default)
    * Open file to Photoshop
    * Apply a Unsharp mask: 80%, Radius 1.0, Threshold 0
    * Save as a TIFF (for cropping) and as a JPEG quality 11 for download

And their comment about that test:
***
Because Adobe Camera Raw doesn't currently render the same level of detail as resolved by the camera's JPEG engine, we've taken the unusual step of using Phase One's Capture One. Again we've minimized sharpening and applied the same unsharp mask in Photoshop.
***

So basically they mangle it by shooting in JPEG format, tweaking and massaging it, and then converting it *again* to another JPEG.  The test is worthless to compare what we're looking for here.
Title: AA-filtering CCD and CMOS
Post by: BernardLanguillier on October 13, 2009, 06:15:32 pm
Quote from: georgl
The fact is that without real scientific measurements under controlled conditions we cannot say for sure what's actually "better" or "worse" - we not only compare filtered vs. unfiltered images, we compare different processing, different sensors, different lenses and slightly different shooting conditions (like the precise alignment of the sensor grid and photographic detail). I was surprised to see so unsharp unfiltered samples from the 5D which really show barely any more detail than the filtered version - it seems that the lens/focusing limited the IQ already.

Another possibility being that, since DSLR sensors are designed and optimized with the presence of the AA filter in mind, removing it simply turns a highly tuned device into a non optimized one?

Cheers,
Bernard
Title: AA-filtering CCD and CMOS
Post by: bjanes on October 13, 2009, 06:24:32 pm
Quote from: ejmartin
The DPReview report on the DP2 provides a good example of the increased "resolution" provided by the lack of an AA filter:

http://www.dpreview.com/reviews/sigmadp2/page22.asp (http://www.dpreview.com/reviews/sigmadp2/page22.asp)

Count the number of lines at one end of the DP2 resolution pattern and compare it to the other end, for any orientation.  The pattern looks nice and sharp, and if you were just to look at the pattern from say 22 to 24 on the chart you would say that the camera is resolving all the way out to the limit of the chart.  But actually it is just aliasing a more finely spaced set of lines into a more coarsely spaced set of lines; it is not actually resolving at that level.  Aliasing definitely helps the accutance though  

An excellent observation by Emil. The resolution of the DP2 is 1760 pixels/picture height. The Nyquist frequency is 880 lp/ph, whereas DPReview is reporting an "absolute" resolution of 1400 to 1500 lp/ph, which is preposterous!

880 lp/ph would be at 8 on their scale.

The conclusion is that aliasing works wonders on a resolution chart.
Title: AA-filtering CCD and CMOS
Post by: BernardLanguillier on October 13, 2009, 06:28:04 pm
Quote from: michael
Bumble bees can fly, though aeronautical engineers used to say that it was impossible. They just didn't bother asking the bees. And yes, removing the AA filter on a properly designed system does increase both real and apparent resolution. Talk to the engineers that design these systems, and also just trust your own eyes.

Michael,

When was the last time you spoke with a sensor R&D chief engineer at Canon, Nikon or Sony? Wouldn't be interesting to hear about their motivation for equiping their sensors with AA filters?

If you are interested, I might be able to arrange an interview with one of these people. Cannot commit anything but I would try.

Please let me know.

Cheers,
Bernard
Title: AA-filtering CCD and CMOS
Post by: Daniel Browning on October 13, 2009, 06:46:50 pm
As I mentioned above, two people can look at the exact same image and get a totally different perception because they have very different tastes. It applies to everything in photography, like saturation. One will look at a certain red tone in a Velvia image and see it as a fake, unnatural-looking color. The other will see it as the ideal pleasing tone. One will look at a certain detail of the image and see it as a fake, unnatural-looking aliasing artifact. The other will see it as the ideal microcontrast. So the very things that are pleasing to one person may be the very thing that another dislikes.

Image A:

(http://thebrownings.name/photo/2009-10-aliasing/2009-01-30-3481-rt-400-point.png)

Image B:

(http://thebrownings.name/photo/2009-10-aliasing/2009-01-30-3481-rt-400-lanczos.png)

To see the differences, it may help to open each image in a different browser tab for direct A/B comparison.

Many would prefer the first image, describing it as sharp, crunchy, high microcontrast, with lots of fine detail, such as stubble. Others would see it as fake-looking, with harsh transitions, jagged edges, and lots of false detail, such as stubble that should be too small to see and jagged edges on the ear.

Some would prefer the second image, describing it as smooth, natural, with the appropriate amount of detail for its size. Others would see it as mushy, hazy, low contrast, and lacking in fine detail.

Which image do you prefer?
Title: AA-filtering CCD and CMOS
Post by: Daniel Browning on October 13, 2009, 07:56:57 pm
Doug,

Quote from: telyt
Show me the artifacts:

Both photos: Leica R8 with DMR digital back (10 MP, no AA filter), 280mm f/4 APO-Telyt-R

I don't see any artifacts in 16" x 24" prints.

I think there is a difference in how you and I perceive fine details. If that's the case, then even if you did provide 100% crops, it would not do any good for me to point out what I consider to be "artifacts", because to you there are not artifacts at all, but the ideal fine detail.

To confirm whether or not we do have a difference in taste, would you consider comparing Image A and B above? Do you see lots of aliasing artifacts? Or just a high-detail vs. low-detail image?

I would also be curious to hear how Michael, George, Nicolas, and Jing Q perceive the difference between the two.

Quote from: michael
Has the fact that every medium format back since Noah has not had an AA filter escaped anyone's notice? Do those that think that all or even most images taken without an AA have aliasing and artifacting think that the countless pros and highly critical photographers who use MF backs would tolerate this if it were true? Come on!

The reality is that aliasing is only visible in "some" images, and unless you shoot fabrics for a living, it's a very small number.

Michael,

You are correct that moiré does not affect most images taken without an AA filter. It only occurs with regularly-repeating fine patterns, like fabric, which is rare for most types of photography. However, moiré is not the only aliasing artifact. My theory is that the details you like in unfiltered images (high microcontrast) are the very things that I consider aliasing artifacts. If you see good microcontrast in image A, above, then the theory is proven. On the other hand, if you see the image as plagued by terrible aliasing, then the theory is not proven.

Quote from: michael
The advantage of not having an AA is clearly visible to anyone that has a good eye. In some cases it simply jumps out at you, image after image. In most cases its visible even in small prints and on screen as a type of increased clarity, as if a veil were lifted.

I agree completely. AA filters reduce contrast severely. It's a very steep price to pay, but I'll gladly pay it to avoid aliasing, because I consider that to be an even worse problem.

Quote from: michael
And as for removing moire in software – don't be too sure it can't be done. I'm working with one company now that is likely to have something exciting in this area in the months ahead.

Thanks for the heads up, I'll look forward to it.

Quote from: telyt
AA blurring is an artifact too.  Pick the one(s) you're willing to live with

Doug, that's a good way to put it. Aliasing and AA blurring are both artifacts in a sense. To me, blurring is far more acceptable than aliasing.

Quote from: georgl
But every AA-filter reduces contrast below Nyqist and therefore information - that's a scientific fact.

George, I agree with you on that. Fortunately, there is a very straightforward method to get that contrast back: reduce pixel size by 30%. If you want 12 MP with no loss of contrast from the AA filter, you have two choices:


That is one of the things I noticed about the 5D2, even in just an 8x10 print. Although the AA filter was still the same relative to Nyquist, it was much weaker in the absolute sense. That weaker filtration resulted in much higher contrast even in print sizes where the pixel resolution itself did not contribute (8x10).

That's why I prefer filtered photographs.
Title: AA-filtering CCD and CMOS
Post by: ejmartin on October 13, 2009, 07:59:54 pm
Quote from: Plekto
Here is what they do with each image:

    *  Load RAW file into Adobe Camera RAW (Auto mode disabled)
    * Set Sharpness to 0 (all other settings default)
    * Open file to Photoshop
    * Apply a Unsharp mask: 80%, Radius 1.0, Threshold 0
    * Save as a TIFF (for cropping) and as a JPEG quality 11 for download

And their comment about that test:
***
Because Adobe Camera Raw doesn't currently render the same level of detail as resolved by the camera's JPEG engine, we've taken the unusual step of using Phase One's Capture One. Again we've minimized sharpening and applied the same unsharp mask in Photoshop.
***

So basically they mangle it by shooting in JPEG format, tweaking and massaging it, and then converting it *again* to another JPEG.  The test is worthless to compare what we're looking for here.

That's not how I understand what is written.  It looks to me as though they used C1 rather than ACR to convert the RAW files, and then proceeding according to their usual protocol.  The jpeg resolution test is on the previous page.
Title: AA-filtering CCD and CMOS
Post by: ErikKaffehr on October 13, 2009, 09:15:08 pm
Hi Michael,

I have great respect for your views and writings. That said I have a couple of points to make.

The first one is that many posters on this forum are involved with imaging theory and it seems mostly that AA filtering is necessary from a theoretical viewpoint.  As an example. it's quite obvious that we get som artifacts like the spurious resolution on the Sigma DP2 in the DPReview test on systems without OLP.

As you have pointed out MF Digital backs didn't employ OLPs since Noah. The problem in my view is that it is very difficult to compare filtered and nonfiltered systems, as we normally don't have access to identical systems in two versions, one having OLP and the other not. A pair of cameras of which one is modifed by Mad Max may be a good comparison. (AFAIK MadMax only removes one of the OLP-filters as the other is bonded to the IR-filter or sensor).

That MF cameras perform better on an per pixel basis may depend on other factors, like better optics (relative to pixel size), plentyful resolution or different processing.

An intresting comparison was made by Erwin Puts, here: http://www.imx.nl/photo/leica/camera/page157/page157.html (http://www.imx.nl/photo/leica/camera/page157/page157.html)

His writing indicates that the Nikon D3X is actually resolving better than the M9 but he feels that the Leica images are more pleasant to the eye.

Best regards
Erik


Quote from: michael
I can't believe the amount of misinformation and disinformation that this topic has engendered.

Has the fact that every medium format back since Noah has not had an AA filter escaped anyone's notice? Do those that think that all or even most images taken without an AA have aliasing and artifacting think that the countless pros and highly critical photographers who use MF backs would tolerate this if it were true? Come on!

The reality is that aliasing is only visible in "some" images, and unless you shoot fabrics for a living, it's a very small number.

The advantage of not having an AA is clearly visible to anyone that has a good eye. In some cases it simply jumps out at you, image after image. In most cases its visible even in small prints and on screen as a type of increased clarity, as if a veil were lifted.

And as for sharpening compensating for having an AA, what are you thinking? Sharpening is about edge sharpness; AKA accutance. Removing an AA filter is about increasing resolution. Yes, adding USM increases apparent sharpness, but not real resolution. Two different things, though related in the real world.

Bumble bees can fly, though aeronautical engineers used to say that it was impossible. They just didn't bother asking the bees. And yes, removing the AA filter on a properly designed system does increase both real and apparent resolution. Talk to the engineers that design these systems, and also just trust your own eyes.

And as for removing moire in software – don't be too sure it can't be done. I'm working with one company now that is likely to have something exciting in this area in the months ahead.  

Michael
Title: AA-filtering CCD and CMOS
Post by: telyt on October 13, 2009, 10:14:06 pm
Quote from: Daniel Browning
I think there is a difference in how you and I perceive fine details. If that's the case, then even if you did provide 100% crops, it would not do any good for me to point out what I consider to be "artifacts", because to you there are not artifacts at all, but the ideal fine detail.

Here are a couple of 100% crops, no sharpening applied:

(http://www.wildlightphoto.com/birds/accipitridae/feha02crop.jpg)

(http://www.wildlightphoto.com/mammals/lagomorphs/btha04crop.jpg)

Keep in mind, these are real-world photos not test photos and as such were made under less-than-ideal conditions: 1/125 sec shutter speed, moving subjects, imperfect support.
Title: AA-filtering CCD and CMOS
Post by: Daniel Browning on October 14, 2009, 02:42:12 am
Quote from: telyt
Here are a couple of 100% crops, no sharpening applied:

Thanks for the response, Doug. The subject, composition, blur, and light are really great; you are very talented.

Quote from: telyt
Keep in mind, these are real-world photos not test photos and as such were made under less-than-ideal conditions: 1/125 sec shutter speed, moving subjects, imperfect support.

That's an excellent point. The more imperfect the conditions, the less likely there is to be aliasing. All it takes to reduce aliasing is a very slight blur, and there are so many things that can cause a very slight blur, including DOF, camera shake, slightly missed focus, subject motion, diffraction, slight lens imperfections, etc. Plus, the smaller and smaller the pixels are, the harder and harder it is to get aliasing.

I don't see much aliasing in the first photo, but in the second one I do. Keep in mind that this is my perception of the detail, I'm sure that others see it very differently. There are many hairs with the width of 1 pixel that occur on a perfectly horizontal line. A bit of stair-stepping from one horizontal line to the next. Some of the very short, bright hairs appear to me to sparkle slightly. The 1-pixel hairs that bend significantly do it on a jagged curve. There are a few hairs that are completely unattached from the head, such as the one in the top left. Some hairs are missing a few pixels half-way through the hair, so that the ends appear unattached to the base. The two-pixel hairs become irregular in shape and thickness when they pass over contrasting backgrounds, like a dark hair on a light background, so they appear to have little bumps. There appear to be tiny spots of green discoloration, though perhaps that is a normal part of the fur.

If any image has sharp transitions from dark to light in just two pixels (black, white), I think we would agree that it has such content. The only difference of opinion is in how we interpret that content. I interpret two-pixel transitions as unnatural. To me, a transition requires 3 pixels (black, gray, white) in order to have a natural appearance. To other viewers, a two-pixel transition is the sought-after ideal of microcontrast.
Title: AA-filtering CCD and CMOS
Post by: jing q on October 14, 2009, 03:11:41 am
wow now that is serious pixel peeping.

what's your opinion on MFDB images then?

Quote from: Daniel Browning
I don't see much aliasing in the first photo, but in the second one I do. Keep in mind that this is my perception of the detail, I'm sure that others see it very differently. There are many hairs with the width of 1 pixel that occur on a perfectly horizontal line. A bit of stair-stepping from one horizontal line to the next. Some of the very short, bright hairs appear to me to sparkle slightly. The 1-pixel hairs that bend significantly do it on a jagged curve. There are a few hairs that are completely unattached from the head, such as the one in the top left. Some hairs are missing a few pixels half-way through the hair, so that the ends appear unattached to the base. The two-pixel hairs become irregular in shape and thickness when they pass over contrasting backgrounds, like a dark hair on a light background, so they appear to have little bumps. There appear to be tiny spots of green discoloration, though perhaps that is a normal part of the fur.

If any image has sharp transitions from dark to light in just two pixels (black, white), I think we would agree that it has such content. The only difference of opinion is in how we interpret that content. I interpret two-pixel transitions as unnatural. To me, a transition requires 3 pixels (black, gray, white) in order to have a natural appearance. To other viewers, a two-pixel transition is the sought-after ideal of microcontrast.
Title: AA-filtering CCD and CMOS
Post by: Daniel Browning on October 14, 2009, 03:50:12 am
Quote from: jing q
wow now that is serious pixel peeping.

Thanks!  I appreciate your complement on my attention to detail; however, aliasing doesn't actually require pixel peeping to see, it affects the perception of the entire image as a whole. Detailed descriptions can help illustrate how individual artifacts contribute to that perception, and since Doug asked me to, I tried to describe some.

Quote from: jing q
what's your opinion on MFDB images then?

Compared to 35mm, as far as image quality is concerned, I like the improved contrast resolution of the lenses operating at lower spatial frequencies. I also appreciate the improved SNR over the most important stops of the dynamic range, leading to less noise and deeper color (tonal resolution/gradations). I appreciate the correct implementation of metadata ISO. What I dislike the most is the aliasing.
Title: AA-filtering CCD and CMOS
Post by: ErikKaffehr on October 14, 2009, 05:28:48 am
Hi!

So you can observe some aliasing in the image. I presume that this is one of the things that you can train eye/brain to see. What happens when you print the image?

I found this image on the Zeiss website demonstrating aspect of resolution/MTF/sharpening: http://www.zeiss.de/C12567A8003B8B6F/Graph...ile/Bild_10.jpg (http://www.zeiss.de/C12567A8003B8B6F/GraphikTitelIntern/CLN31MTF-KurvenBild10/$File/Bild_10.jpg)

It's an illustration to this article: http://www.smt.zeiss.com/C12567A8003B8B6F/...Kurven_2_en.pdf (http://www.smt.zeiss.com/C12567A8003B8B6F/EmbedTitelIntern/CLN_31_MTF_en/$File/CLN_MTF_Kurven_2_en.pdf)

I have printed that image in A2 and studied it along the article,  a recommended exercise.

See also below
http://83.177.178.241/ekr/index.php/photoa...-and-perception (http://83.177.178.241/ekr/index.php/photoarticles/22-a-very-god-article-about-mtf-and-perception)

Best regards
Erik

My experience is that the perceptive system can be trained to detect small differences, once you have seen them you detect quite easily.
Quote from: Daniel Browning
Thanks!  I appreciate your complement on my attention to detail; however, aliasing doesn't actually require pixel peeping to see, it affects the perception of the entire image as a whole. Detailed descriptions can help illustrate how individual artifacts contribute to that perception, and since Doug asked me to, I tried to describe some.



Compared to 35mm, as far as image quality is concerned, I like the improved contrast resolution of the lenses operating at lower spatial frequencies. I also appreciate the improved SNR over the most important stops of the dynamic range, leading to less noise and deeper color (tonal resolution/gradations). I appreciate the correct implementation of metadata ISO. What I dislike the most is the aliasing.
Title: AA-filtering CCD and CMOS
Post by: ErikKaffehr on October 14, 2009, 05:49:51 am
Hi!

If you draw a line from just below the nose to just below the eye you can see four regions:

1) At the nose no hair is actually resolved, this may depend on very little or thin hair or beeing slightly out of focus
2) A bit from the nose the hair is actually resolved
3) About half way there is a lot of artifacing. Many strains are discontinous
4) Just below the hair is not resolved but just goes mush. This is not likely a depth of field effect as the eye and is well resolved.

That said I'm quite impressed with your technique! The sharpness is excellent even without sharpening.

Best regards
Erik

Quote from: telyt
Here are a couple of 100% crops, no sharpening applied:

(http://www.wildlightphoto.com/birds/accipitridae/feha02crop.jpg)

(http://www.wildlightphoto.com/mammals/lagomorphs/btha04crop.jpg)

Keep in mind, these are real-world photos not test photos and as such were made under less-than-ideal conditions: 1/125 sec shutter speed, moving subjects, imperfect support.
Title: AA-filtering CCD and CMOS
Post by: jing q on October 14, 2009, 07:28:08 am
Quote from: Daniel Browning
Thanks!  I appreciate your complement on my attention to detail; however, aliasing doesn't actually require pixel peeping to see, it affects the perception of the entire image as a whole. Detailed descriptions can help illustrate how individual artifacts contribute to that perception, and since Doug asked me to, I tried to describe some.



Compared to 35mm, as far as image quality is concerned, I like the improved contrast resolution of the lenses operating at lower spatial frequencies. I also appreciate the improved SNR over the most important stops of the dynamic range, leading to less noise and deeper color (tonal resolution/gradations). I appreciate the correct implementation of metadata ISO. What I dislike the most is the aliasing.

After all the technical talk, I still don't see your point or why aliasing is such a big issue.
Perhaps that's where the problem lies, a lot of people who use AA filterless images actually are not bothered by the aliasing?
Granted, there are times when it is visible (esp in certain highlights)
but from my personal experience no matter how much I sharpen images I still get a slightly mushy and blurred effect with DSLRs with their AA filters.
And the mushiness/blurring is MUCH more disturbing to me than any aliasing.

I guess I like my images out of the box having a nice pop instead of having to go through rounds of processing with it.
And also although people keep repeating that you can sharpen to regain detail, from personal experience no matter how much you sharpen a mushy edge you still get stuck with softer, mushier transitions upon interpolation.

which goes back to your explanation of 3 pixel natural edge vs 2 pixel unnatural edge
I actually don't find 2 pixel edges unnatural. I actually think that's what makes lines so much more defined.

Another thing is that how visible are aliasing effects when files are actually printed?
Title: AA-filtering CCD and CMOS
Post by: telyt on October 14, 2009, 07:59:24 am
Quote from: Daniel Browning
I don't see much aliasing in the first photo, but in the second one I do.

You're only seeing aliasing with some extreme pixel-peeping, and in details that would have been mushed out by an AA filter.  In a 16" x 24" print you'd have to use a loupe and your imagination to see any of this.  So, pick your artifacts: mush or aliasing.
Title: AA-filtering CCD and CMOS
Post by: telyt on October 14, 2009, 08:11:17 am
Quote from: ErikKaffehr
1) At the nose no hair is actually resolved, this may depend on very little or thin hair or beeing slightly out of focus
2) A bit from the nose the hair is actually resolved

limited DOF.  The photo was made with a 280mm lens @ f/4 and about 3 meters distance.


Quote from: ErikKaffehr
3) About half way there is a lot of artifacing. Many strains are discontinous

can you point these out?  With extreme pixel-peeping there are some whiskers that show stair-stepping but this is not evident in a large print.  Many hairs are crossing over other hairs, is this what you're seeing?

Quote from: ErikKaffehr
4) Just below the hair is not resolved but just goes mush. This is not likely a depth of field effect as the eye and is well resolved.

I'm not sure which area you're referring to.  Mush isn't an aliasing artifact.
Title: AA-filtering CCD and CMOS
Post by: michael on October 14, 2009, 08:45:58 am
Trying to bring things back to practical reality...

I've spent the past 10 days shooting with four different cameras, a Phase One 645 with P65+, a Sony A900 with Zeiss glass, a Leica M9 and a Panasonic GF1. These were all of fall colour in Northern Ontario, almost all of it tripod or monopod mounted. (One of the most spectacular fall colour shows in recorded history I'm told by park rangers).

I have now made some 40 prints, ranging in size from 13X19" to 24X36".

Here's what I see. The GF1 shots are very nice in print sizes up to 11X17". They are easily distinguished though from any of the others. A wonderful casual use camera.

The Sony A900 with Zeiss glass comes next. Really excellent. No one would kick it out of bed.

Then there's the Leica M9. Visibly superior to the A900, even on small prints. This is a combination of resolution, clarity, accutance and modeling. Much of this may well be caused by the Leica lenses. The point being that while the A900 at low ISOs, and when used with Zeiss lenses, is as good as any DSLR on the market, the M9 using Leica glass just looks superior to my eyes. There's a palpable three dimenionality to the images that's hard to put into words, but is akin to the sweetness of a tube amp over a transistor amp. (Pure Class A over Class A/B).

The P65+ is in a class of it's own, but only when prints over about 20X24" are made. Otherwise I find it hard to differentiate between them and the M9, except that I usually can identify the M9 shots by their other "drawing" characteristics, especially with the 35mm Summilux, which seems to be in a class by itself; almost luminous, if you'll pardon the expression. The P65+ has other superior characteristics, such as dynamic range, but that's another discussion.

The question is, could I make these differentiations in a double blind test? Probably not 100%, but since I took the opportunity to do some sider-by-side shots, and can clearly see the differences (though admittedly they usually are subtile) I'm confident that likely score quite a bit better than random chance much of the time.

As for artifacting, well, try as I might, on prints and at 300% on screen, I just don't see it, as long as sharpening is carefully and properly done. If I'm confusing artifacting with resolution, then long live the revolution.

As for the question asked – if AA filters are so limiting, why do smart companies like Nikon and Canon use them? I would conversely ask, if AA filters are such a problem, why do smart companies like Phase One and Hasselblad and Leica not use them? Clearly designers and engineers make design choices based on what they'd like their equipment to accomplish. We as consumers do similarly. We each have different criteria and needs, not to mention budgets.

The discussions of theory are fine, and I enjoy reading them. But, just as in the world of audio I trust my ears over measurements, in photography I trust my eyes above all else, even if what they are seeing can be described as an illusion.

If you want more from me on this, come to my talk at B&H Photo next Thursday (http://www.bhphotovideo.com/find/eventDetails.jsp/id/530), 10am till noon (though it appears to be sold out). It's titled "The Photographic Illusion".

Michael
Title: AA-filtering CCD and CMOS
Post by: pcunite on October 14, 2009, 10:01:20 am
Quote from: michael
As for the question asked – if AA filters are so limiting, why do smart companies like Nikon and Canon use them? I would conversely ask, if AA filters are such a problem, why do smart companies like Phase One and Hasselblad and Leica not use them? Clearly designers and engineers make design choices based on what they'd like their equipment to accomplish. We as consumers do similarly. We each have different criteria and needs, not to mention budgets.

The discussions of theory are fine, and I enjoy reading them. But, just as in the world of audio I trust my ears over measurements, in photography I trust my eyes above all else, even if what they are seeing can be described as an illusion.

This is why some people prefer over-saturated and warmer images while others don't. If I had spent $30K + on a digital back I can assure you I would prefer it's results as well. I think that photographers are too emotional to ever reach any conclusion on which system is better for their needs.
Title: AA-filtering CCD and CMOS
Post by: ErikKaffehr on October 14, 2009, 10:12:49 am
Hi,

Sorry for the term mush, may be my english. I mean a region without resolved detail.

I include the region which contains artifacts in my view: [attachment=17192:aliasing.jpg]

Seen as discontinous hairs and brightness variations along strains.

According to imaging theory there should be aliasing artifacts if the lens resolves above the Nyquist limit, so it just may prove that you have a good lens and technique. I won't argue that they are harmful.

Best regards
Erik



Quote from: telyt
limited DOF.  The photo was made with a 280mm lens @ f/4 and about 3 meters distance.




can you point these out?  With extreme pixel-peeping there are some whiskers that show stair-stepping but this is not evident in a large print.  Many hairs are crossing over other hairs, is this what you're seeing?



I'm not sure which area you're referring to.  Mush isn't an aliasing artifact.
Title: AA-filtering CCD and CMOS
Post by: jing q on October 14, 2009, 10:25:23 am
Quote from: pcunite
This is why some people prefer over-saturated and warmer images while others don't. If I had spent $30K + on a digital back I can assure you I would prefer it's results as well. I think that photographers are too emotional to ever reach any conclusion on which system is better for their needs.

are you saying that michael is giving his opinion as such because he spent $30k on the camera?
actually all you need to do is get a kodak proback for dirt cheap and try it out and see the difference vs a 5D Mk II
Title: AA-filtering CCD and CMOS
Post by: BernardLanguillier on October 14, 2009, 10:50:28 am
Quote from: michael
The point being that while the A900 at low ISOs, and when used with Zeiss lenses, is as good as any DSLR on the market, the M9 using Leica glass just looks superior to my eyes. There's a palpable three dimenionality to the images that's hard to put into words, but is akin to the sweetness of a tube amp over a transistor amp. (Pure Class A over Class A/B).

Michael,

Allow me to disagree with your A900 comment. Not all DSLR are equal. They may appear close compared to the quality of a P65+, but important differences exist in detail and DR, even at equal pixel count.

I have never seen any A900 image getting close to the sharpness shown in the file below for instance. To my eyes, this is as detailed (and aliased for that matter) as any AA filter less file.

Original image reduced to 2000 pixel for posting purpose:

(http://farm3.static.flickr.com/2673/4010915467_9f34d4f8ea_o.jpg)

100% crop of the central part of the image:

(http://farm3.static.flickr.com/2500/4011680522_e6a785b6fc_b.jpg)

And for reference, the final 150 megapixel pano using this image...

(http://farm3.static.flickr.com/2544/4004073853_5bac0406d8_o.jpg)

Cheers,
Bernard
Title: AA-filtering CCD and CMOS
Post by: pcunite on October 14, 2009, 11:06:33 am
Quote from: jing q
are you saying that michael is giving his opinion as such because he spent $30k on the camera?
actually all you need to do is get a kodak proback for dirt cheap and try it out and see the difference vs a 5D Mk II

I can clearly seen the benefits of MFD at low ISO and large viewing sizes. But to say that MFD is also superior in other ways (such as 8x10 size) is tempting after you have spent $30K. Michael is well respected for hosting this site and giving us his opinions. I don't know the limit of 35mm as I have never printed beyond 16x24 (it was from a 10mp file and I did not really like it personally, the customer loved it) but I can say that MFD is not the best for every situation, micro contrast or not. It is the $30K price that causes one to see something special in a small print. That is all I am saying.
Title: AA-filtering CCD and CMOS
Post by: tho_mas on October 14, 2009, 12:02:03 pm
Quote from: BernardLanguillier
To my eyes, this is as detailed (and aliased for that matter) as any AA filter less file.
to my eyes it is not.
But what does your view or my view proof? ... unless you shoot the same motif side by side with and without AA filter sensors and (roughly) comparable lenses?
Title: AA-filtering CCD and CMOS
Post by: michael on October 14, 2009, 12:13:32 pm
Quote from: pcunite
This is why some people prefer over-saturated and warmer images while others don't. If I had spent $30K + on a digital back I can assure you I would prefer it's results as well. I think that photographers are too emotional to ever reach any conclusion on which system is better for their needs.

You are presuming that I appreciate a P65+ back (or an M9 for that matter) because of how much I spent on it. That is erroneous.

Indeed it is just the opposite. I spend the money because I can see the difference that this level of gear provides for the type of shooting that I do and the size of prints that I make, exhibit and sell.

I am subject to self delusion and rationalization as much as most people, but not when it comes to spending that kind of money. I am fortunate in that I get to field test all sorts of high end gear. No guess work, no buyers remorse (at least not often), and always based on what I can see.

Others may differ with my opinions, but I put my money where my mouth is – or at least where my eyes suggest they should be.

Michael
Title: AA-filtering CCD and CMOS
Post by: bjanes on October 14, 2009, 12:25:13 pm
Quote from: michael
As for the question asked – if AA filters are so limiting, why do smart companies like Nikon and Canon use them? I would conversely ask, if AA filters are such a problem, why do smart companies like Phase One and Hasselblad and Leica not use them? Clearly designers and engineers make design choices based on what they'd like their equipment to accomplish. We as consumers do similarly. We each have different criteria and needs, not to mention budgets.

Michael

There may be other considerations as to why MFDBs and Leicas lack a blur filter. As pointed out elsewhere, a blur filter for a MFDB would be quite expensive to produce, since the cost increases exponentially with the sensor size. Also, MDFBs are very high resolution, and alaising could be downsized out of existence when the print is made. The Leica is limited by a short back focal distance (see Erwin Puts (http://www.imx.nl/photo/leica/camera/page155/m9part2.html)) which limits what kind of filter can be placed in front of the sensor. In the M8 they were not even to implement an IR filter. Erwin states "The M9 filter thickness is now 0.8mm as compared to 0.5mm for the M8 and more than 2mm for the D3x". A 2mm filter could cause severe problems with the short back focal distance of the Leicas.

Title: AA-filtering CCD and CMOS
Post by: pcunite on October 14, 2009, 12:34:33 pm
Quote from: michael
Others may differ with my opinions, but I put my money where my mouth is – or at least where my eyes suggest they should be.

Thank you for your thoughts Michael.
Title: AA-filtering CCD and CMOS
Post by: jing q on October 14, 2009, 12:39:59 pm
Quote from: pcunite
I can clearly seen the benefits of MFD at low ISO and large viewing sizes. But to say that MFD is also superior in other ways (such as 8x10 size) is tempting after you have spent $30K. Michael is well respected for hosting this site and giving us his opinions. I don't know the limit of 35mm as I have never printed beyond 16x24 (it was from a 10mp file and I did not really like it personally, the customer loved it) but I can say that MFD is not the best for every situation, micro contrast or not. It is the $30K price that causes one to see something special in a small print. That is all I am saying.

I think he stated that he was hard pressed to tell a difference in cameras until a larger size print was made...
Title: AA-filtering CCD and CMOS
Post by: Daniel Browning on October 14, 2009, 12:43:18 pm
Quote from: jing q
After all the technical talk, I still don't see your point or why aliasing is such a big issue.
Perhaps that's where the problem lies, a lot of people who use AA filterless images actually are not bothered by the aliasing?

Yes, that is my best guess. What to me is aliasing is to another person the ideal microcontrast.

Quote from: jing q
Granted, there are times when it is visible (esp in certain highlights)
but from my personal experience no matter how much I sharpen images I still get a slightly mushy and blurred effect with DSLRs with their AA filters.
And the mushiness/blurring is MUCH more disturbing to me than any aliasing.

I can understand your point of view. For me, sharpening compensates for the OLPF very well, but it leaves me between a rock and a hard place when noise is high.

However, even if you use no sharpening whatsoever, you can completely remove the effects of the OLPF by simply buying a sensor with 1.3X more linear resolution than you need. If you need a sharp 12 MP, buy 20 MP. Then you get the rid of the mushy/blurred effect without sharpening and without aliasing. However, if aliasing is the very thing that you are trying to get, then you would not like it.

Quote from: jing q
which goes back to your explanation of 3 pixel natural edge vs 2 pixel unnatural edge
I actually don't find 2 pixel edges unnatural. I actually think that's what makes lines so much more defined.

I think that helps explain the difference in perception. When you look at Image A/B comparsion from my earlier post, do you find Image A to be more pleasing?

Quote from: jing q
Another thing is that how visible are aliasing effects when files are actually printed?

If the smallest details are visible, then so too is the aliasing. I see a ton of aliasing artifacts in prints all the time, but they're usually caused by the downsampling during post, not the camera itself. Photoshop does poorly with large resampling ratios, such as taking 21 MP down to 2 MP for a 4x6. Lightroom 2.0 was slightly better, and later in 2.x (I think) they improved the downsampling algorithm further, but it still results in far too much aliasing for my taste. Qimage, irfanview, and ImageMagick are much better.


Quote from: telyt
So, pick your artifacts: mush or aliasing.

I pick mush.

Quote from: michael
As for artifacting, well, try as I might, on prints and at 300% on screen, I just don't see it, as long as sharpening is carefully and properly done. If I'm confusing artifacting with resolution, then long live the revolution.

One man's ugly artifacts are another man's fine detail. I think it's a difference in taste/perception.

Quote from: pcunite
I can clearly seen the benefits of MFD at low ISO and large viewing sizes. But to say that MFD is also superior in other ways (such as 8x10 size)...

I kindly disagree. There are many circumstances when a MF back is superior even at 8x10 size. For example, let's say the viewer's CoC is small enough to see 1000 LP/PH in the 8x10. On 35mm, that corresponds to 42 lp/mm. But on Medium Format the same print size is only 26 lp/mm due to smaller reproduction magnification. Many wide angle 35mm lenses are not able to provide the same level of aberration correction at much higher spatial frequencies, which is clearly reflected in their MTF charts. Worse still, they must operate at wider f-numbers to get the same depth of field and diffraction.
Title: AA-filtering CCD and CMOS
Post by: BernardLanguillier on October 14, 2009, 04:24:54 pm
Quote from: tho_mas
to my eyes it is not.
But what does your view or my view proof? ... unless you shoot the same motif side by side with and without AA filter sensors and (roughly) comparable lenses?

True, but this was not my main point anyway.

As far as this M9 vs DSLR topic is concerned, there is at least one person who did the test on both a D3x and M9, and his conclusion are different from Michael's regarding the amount of detail captured (http://www.imx.nl/photo/leica/camera/page155/m9part2.html).

Citing him:

"We may infer from these graphs that the Nikon images benefit more from the post processing, but get that now familiar (and not always pleaseant) digital look, where the Leica images are more closely related to the classical film look and here the sharpening effect is less pronounced. The Leica M9 is quite close to The Nikon D3x in definition and resolution, but Nikon photographers do not need to fear that the M9 will dethrone the D3x as the reference camera for state of the art quality. Stunning as the M9 pictures are, they must be put in context and then the Nikon D3x images are just better. "

Since Michael did not test the D3x against the M9 (Michael doesn't own a D3x as far as I know), but based his comment on a generic belief that all DSLRs are basically the same (conclusion being based on a limited amount of D3x usage), I tend to trust Erwin's conclusions more on this particular point, especially knowing that he is known to be very close to Leica.

Cheers,
Bernard
Title: AA-filtering CCD and CMOS
Post by: tho_mas on October 14, 2009, 05:01:02 pm
Quote from: BernardLanguillier
As far as this M9 vs DSLR topic is concerned, there is at least one person who did the test on both a D3x and M9, and his conclusion are different from Michael's regarding the amount of detail captured
But Michael is not referring to "details captured" only as it would be boring anyhow:
Quote
a combination of resolution, clarity, accutance and modeling. Much of this may well be caused by the Leica lenses (...)
There's a palpable three dimenionality to the images that's hard to put into words
This notion is brought up quite often. And indeed one can't disprove it with pure measurement of data.
Title: AA-filtering CCD and CMOS
Post by: Daniel Browning on October 14, 2009, 05:08:20 pm
Quote from: tho_mas
But Michael is not referring to "details captured" only as it would be boring anyhow:
This notion is brought up quite often. And indeed one can't disprove it with pure measurement of data.

All the talk I've ever heard about three dimensionality can be attributed to one or more of the following:


For example, MFDB has exactly those factors. If you take them away, you take away the "3D-ness". For example, shoot telephoto, deep DOF, with a lowcon filter (or junky lens) to remove contrast, and you'll end up with a non-3D photo from a MFDB.

In the case of the Leica, the rangefinder advantage is especially strong with wide angles and thin DOF. They also provide good contrast and strong aliasing.

Foveon, too, was described by many as 3D. In that case it was just the aliasing.
Title: AA-filtering CCD and CMOS
Post by: BernardLanguillier on October 14, 2009, 05:16:49 pm
Quote from: tho_mas
But Michael is not referring to "details captured" only as it would be boring anyhow:
This notion is brought up quite often. And indeed one can't disprove it with pure measurement of data.

It does read like a Stereophile review, doesn't it?

Cheers,
Bernard
Title: AA-filtering CCD and CMOS
Post by: tho_mas on October 14, 2009, 05:27:55 pm
Quote from: BernardLanguillier
It does read like a Stereophile review, doesn't it?
well, maybe :-)
Everyone has his own experiences and value certain properties over others ... and there's nothing wrong with that.
Title: AA-filtering CCD and CMOS
Post by: bjanes on October 14, 2009, 05:29:14 pm
Quote from: michael
Trying to bring things back to practical reality...

Then there's the Leica M9. Visibly superior to the A900, even on small prints. This is a combination of resolution, clarity, accutance and modeling. Much of this may well be caused by the Leica lenses. The point being that while the A900 at low ISOs, and when used with Zeiss lenses, is as good as any DSLR on the market, the M9 using Leica glass just looks superior to my eyes. There's a palpable three dimenionality to the images that's hard to put into words, but is akin to the sweetness of a tube amp over a transistor amp. (Pure Class A over Class A/B).

Quote from: BernardLanguillier
It has been obvious for years that high performance and small volume in high tech are incompatible requirements in the mid/long run.

The way nuforce is totally trouncing high end hifi equipment costing 5 or 10 times more is yet another example in the supposedely different high end hifi world too.

Cheers,
Bernard

Quote from: BernardLanguillier
It does read like a Stereophile review, doesn't it?

Cheers,
Bernard

Good heavens, that nuforce appears to be solid state .  Could it possibly best a tube amplifier?
Title: AA-filtering CCD and CMOS
Post by: michael on October 14, 2009, 05:45:20 pm
Bernard,

I used to have a lot of respect for Erwin's lens-related work, but his recent digital camera reports are showing cracks in his methodology and understanding. Doing his tests using in-camera JPGs, for example.

I have no desire to knock him here, but I have heard from more than a couple of knowledgeable tech journalists recently who are privately musing that everything is not all as it should be with his recent work on cameras and sensors as opposed to lenses.

In other words, I disagree with Erwin's M9 results based on my own tests, and while I haven't done the D3x comparison I have done it with the A900 and the 1Ds MKIII, which others have measured as having comparable resolution to the D3x. The D3x's forte, as you know, is its high ISO capability, and not any particular lead in the resolution department. And I have no need to quibble over Nikon glass vs. Zeiss. The best of both are comparable, and probably differ more in individual samples that in type.

Michael
Title: AA-filtering CCD and CMOS
Post by: BernardLanguillier on October 14, 2009, 08:11:24 pm
Quote from: bjanes
Good heavens, that nuforce appears to be solid state .  Could it possibly best a tube amplifier?

I know... either way, my personnal preference remains to have my own private orchestra playing on demand in the kiosk, you can't beat that kind of DR and even nuforce can't factor in the influence of the birds singing in the background.

Cheers,
Bernard
Title: AA-filtering CCD and CMOS
Post by: BernardLanguillier on October 14, 2009, 08:20:13 pm
Quote from: michael
In other words, I disagree with Erwin's M9 results based on my own tests, and while I haven't done the D3x comparison I have done it with the A900 and the 1Ds MKIII, which others have measured as having comparable resolution to the D3x. The D3x's forte, as you know, is its high ISO capability, and not any particular lead in the resolution department. And I have no need to quibble over Nikon glass vs. Zeiss. The best of both are comparable, and probably differ more in individual samples that in type.

Michael,

Thanks for the feedback.

Having not yet used a M9 myself, I cannot comment further on the comparison between these 2. I would be very surprised if typical usage of the M9 resulted in critically focused files most of the time anyway, so this is very theoretical I believe.

As far as the D3x vs the other DSLRs goes, the key is that the amount of sharpening that a file can take at low ISO is directly impacted by the level of noise present in the file. This is where the main strenght of the D3x relative to the competition, larger DR and lower noise in low ISO files, plays a major role and makes it possible to extract significantly more detail with proper sharpening (this being further facilitated by the weaker AA filter the sensor is equiped with). I guess that the example I posted above speaks for itself.

This conclusion is coming from shooting tens of thousands of images with the D3x, mostly with the best Zeiss glass available in Nikon mount.

I guess that we'll have to agree to disagree on this point.

Cheers,
Bernard
Title: AA-filtering CCD and CMOS
Post by: telyt on October 14, 2009, 09:25:57 pm
Quote from: pcunite
If I had spent $30K + on a digital back I can assure you I would prefer it's results as well.

If I were about to spend that much money on a camera I'd be damned sure it provided a tangible real-world benefit before I opened my wallet.

Quote from: pcunite
I think that photographers are too emotional to ever reach any conclusion on which system is better for their needs.

On this we agree completely!  The number of times people have defended their mushy AA-filtered images is staggering!

Quote from: ErikKaffehr
Sorry for the term mush, may be my english. I mean a region without resolved detail.

It might be my technique.  After all 1/125 sec on a monopod & shoulder stock with no mirror pre-release or ST @ full aperture isn't quite optimum.

Quote from: michael
I used to have a lot of respect for Erwin's lens-related work, but his recent digital camera reports are showing cracks in his methodology and understanding. Doing his tests using in-camera JPGs, for example.

My thoughts too.  He also has compared cameras' raw files after processing all samples with the same raw developer.  I don't think he'd have done that when comparing films.  IMHO he's only started on the digital learning curve.
Title: AA-filtering CCD and CMOS
Post by: ErikKaffehr on October 14, 2009, 09:46:59 pm
Michael,

When you talk about the A900 and Zeiss lenses are you discussing the zooms or the primes? My take is that the zooms are pretty good but I would guess that both the Zeiss primes (specially the 135/1.8) and the Leica lenses may be a different league.

Going back to the Erwin Puts articles I don't think he uses in camera JPEGS his testing, is based on Capture One on Nikon and I guess Lightroom on the Leica. He certainly does have no bias against Leica. I often have difficulty understanding his writing but still am thankful for information.

Most of us don't really have the option to have access to all equipment and therefore depend on independent reviews and sample pictures. Sites like Imaging Resource have lot of hopefully high quality images, even in RAW. I have unfortunately not seen any MFDB images of the ISO 12233 test charts for instance.

The nearest thing I have seen was actually the MFDB shootout on Luminious Landscape. I have tried to reproduce it partly for an article I have in progress comparing MF Velvia and 24MP DSLR. I actually made prints from small crops and found that the Alpha 900 I have was remarkably good but even a blind man could see that the P45 was better. Stopping the P45 down to f/16 gave significantly worse results, however. Even at f/16 the P45 was better than the Alpha 900.

My guess is that you need to enlarge above A2 to really benefit from MFDBs over DSLRs with 20+ MP. Experienced viewers may see differences easier.



Quote from: michael
Bernard,

I used to have a lot of respect for Erwin's lens-related work, but his recent digital camera reports are showing cracks in his methodology and understanding. Doing his tests using in-camera JPGs, for example.

I have no desire to knock him here, but I have heard from more than a couple of knowledgeable tech journalists recently who are privately musing that everything is not all as it should be with his recent work on cameras and sensors as opposed to lenses.

In other words, I disagree with Erwin's M9 results based on my own tests, and while I haven't done the D3x comparison I have done it with the A900 and the 1Ds MKIII, which others have measured as having comparable resolution to the D3x. The D3x's forte, as you know, is its high ISO capability, and not any particular lead in the resolution department. And I have no need to quibble over Nikon glass vs. Zeiss. The best of both are comparable, and probably differ more in individual samples that in type.

Michael
Title: AA-filtering CCD and CMOS
Post by: aaykay on October 14, 2009, 10:58:32 pm
Quote from: BernardLanguillier
Since Michael did not test the D3x against the M9 (Michael doesn't own a D3x as far as I know), but based his comment on a generic belief that all DSLRs are basically the same (conclusion being based on a limited amount of D3x usage), I tend to trust Erwin's conclusions more on this particular point, especially knowing that he is known to be very close to Leica.

And Erwin does not own an A900 (as far as I know) and he did not test it head-to-head with the M9, by say putting something like the ultra-fine Zeiss 135mm f/1.8 on the A900 and the 135mm f/3.4 M Leica lens on the M9 for a one-to-one comparison.  He would probably have reached the same conclusion with the A900, that he did with the D3X - maybe even more so, since even the lofty D3X does not get to be coupled with as fine a lens as the Sony's Zeiss, to my knowledge.   The lens obviously has an outsized influence here on the final result, and my personal finding is that the Zeiss primes in the Sony/Alpha system (the 135mm even more than the 85mm) have that extra wallop that even the excellent Zeiss zooms like the 24-70 f/2.8 and the 16-35 f/2.8 does not have.

Clearly, it seems like Michael and Erwin have differing views on what they find to be "better" and you find Erwin's view more convenient and believable.    I personally am pretty neutral on this point.
Title: AA-filtering CCD and CMOS
Post by: ErikKaffehr on October 14, 2009, 11:43:53 pm
Hi,

I'm actually a bit confused by Erwin Puts test. For instance he has a picture of a couple of combs. In this case I certainly perceive the M9 image to be sharper than the 3DX image although Erwin says that the D3X is sharper.

Also, the MTF he shows for Nikon has excessive sharpening, I have done a lot of testing with Imatest, although not professionally, and I never had that level of sharpening Erwin shows.

Erwin's stuff is interesting and worth the read, but I don't really feel that the information is complete, especially it is hard to know exactly what he means.

Best regards
Erik


Quote from: aaykay
And Erwin does not own an A900 (as far as I know) and he did not test it head-to-head with the M9, by say putting something like the ultra-fine Zeiss 135mm f/1.8 on the A900 and the 135mm f/3.4 M Leica lens on the M9 for a one-to-one comparison.  He would probably have reached the same conclusion with the A900, that he did with the D3X - maybe even more so, since even the lofty D3X does not get to be coupled with as fine a lens as the Sony's Zeiss, to my knowledge.   The lens obviously has an outsized influence here on the final result, and my personal finding is that the Zeiss primes in the Sony/Alpha system (the 135mm even more than the 85mm) have that extra wallop that even the excellent Zeiss zooms like the 24-70 f/2.8 and the 16-35 f/2.8 does not have.

Clearly, it seems like Michael and Erwin have differing views on what they find to be "better" and you find Erwin's view more convenient and believable.    I personally am pretty neutral on this point.
Title: AA-filtering CCD and CMOS
Post by: BernardLanguillier on October 14, 2009, 11:50:37 pm
Quote from: aaykay
And Erwin does not own an A900 (as far as I know) and he did not test it head-to-head with the M9, by say putting something like the ultra-fine Zeiss 135mm f/1.8 on the A900 and the 135mm f/3.4 M Leica lens on the M9 for a one-to-one comparison.  He would probably have reached the same conclusion with the A900, that he did with the D3X - maybe even more so, since even the lofty D3X does not get to be coupled with as fine a lens as the Sony's Zeiss, to my knowledge.   The lens obviously has an outsized influence here on the final result, and my personal finding is that the Zeiss primes in the Sony/Alpha system (the 135mm even more than the 85mm) have that extra wallop that even the excellent Zeiss zooms like the 24-70 f/2.8 and the 16-35 f/2.8 does not have.

Well, the Zeiss 100mm f2.0 I have been using a lot this past year is a totally brilliant performer (said by many to be one of the very best pieces of glass ever). I have not seen a comparison between the 100 and 135, but I would be surprised if they were very far appart.

Quote from: aaykay
Clearly, it seems like Michael and Erwin have differing views on what they find to be "better" and you find Erwin's view more convenient and believable.    I personally am pretty neutral on this point.

What we have here are 2 different views on 2 differents things. The proposition that these views can be compared in absolute terms only stands on the assumption that the A900 and D3x have the same level of performance. That is the point I cannot agree with.

(A>B, C>D) only yields relevant information on the relative value of A and D if the relationship between B and C is known.

I couldn't care less if the M9 were better than the d3x, that doesn't affect me and I have no way to verify this myself. I am only glad to see more valuable photographic options on the market. But it should be clear that Michael's assessement that the M9 is better than the A900 doesn't establish any clear relative ranking between the M9 and the D3x, one way or the other.

Having worked with files from the D3x (an awful lot) and the A900 (a lot less but still enough to form an opinion), I see differences that are as significant as it gets between DSLRs of similar pixel count. Put it otherwise, if the A900 and D3x are the same for Michael, then he doesn't need bother looking at any new DSLR in next 2 or 3 years, they will be the same as well.

I could very well understand that compared to a P65+ that he will use for all his critical work anyway, both are poor resolutionwise. Now for those of us looking at extracting the best possible quality from DSLRs, much cleaner shadows and the ability to apply a lot more low radius sharpening are very relevant and important factors.

Final comment from me on this.  

Cheers,
Bernard
Title: AA-filtering CCD and CMOS
Post by: thierrylegros396 on October 15, 2009, 04:00:34 am
Quote from: ErikKaffehr
Hi,

Sorry for the term mush, may be my english. I mean a region without resolved detail.

I include the region which contains artifacts in my view: [attachment=17192:aliasing.jpg]

Seen as discontinous hairs and brightness variations along strains.

According to imaging theory there should be aliasing artifacts if the lens resolves above the Nyquist limit, so it just may prove that you have a good lens and technique. I won't argue that they are harmful.

Best regards
Erik

Very interresting artifacts image.

But from where comes those artifacts (RAW conversion, filtering, jpeg) ?!

It would be kind to have acces to the RAW file.

Hav a Nice Day.

Thierry
Title: AA-filtering CCD and CMOS
Post by: ErikKaffehr on October 15, 2009, 08:11:55 am
Two comments.

Regarding the part of the image beeing "mushy" I don't suggest it's your techique. The aliasing I'm certain to see certainly indicates that your technique works very well. The area just below the eye lacks detail and I don't think it's a DOF issue, more like simply that the detail is smaller than that can be resolved.

Regarding testing I would say that using the same converter for all tested cameras is good science, I may even add that using a known, well documented and basic converter is optimal. After basic testing have been done we can explore different converters, but are you testing converter or camera/lens/sensor in that case?

Imatest that Erwin Puts use does have a built in RAW converter called DC-Raw.

I have some writeup comparing scanned 67 Velvia with my Alpha 900 and see a similar problem. Both the scanned Velvia and the Alpha images need sharpening. So I could tweak sharpening and try to achieve optimal sharpening, but optimal sharpening to one observer may be to much for another observer.

So what I did was to use the normal settings I would start with, Photokit Sharpener with settings for MF film and "Landscape" preset in Lightroom.

My article on MF Velvia vs. Sony Alpha 900 is here: http://83.177.178.241/ekr/index.php/photoa...-sony-alpha-900 (http://83.177.178.241/ekr/index.php/photoarticles/16-pentax67velvia-vs-sony-alpha-900)


Best regards
Erik


Quote from: telyt
It might be my technique.  After all 1/125 sec on a monopod & shoulder stock with no mirror pre-release or ST @ full aperture isn't quite optimum.



My thoughts too.  He also has compared cameras' raw files after processing all samples with the same raw developer.  I don't think he'd have done that when comparing films.  IMHO he's only started on the digital learning curve.
Title: AA-filtering CCD and CMOS
Post by: bjanes on October 15, 2009, 08:48:29 am
Quote from: ErikKaffehr
I'm actually a bit confused by Erwin Puts test. For instance he has a picture of a couple of combs. In this case I certainly perceive the M9 image to be sharper than the 3DX image although Erwin says that the D3X is sharper.

Also, the MTF he shows for Nikon has excessive sharpening, I have done a lot of testing with Imatest, although not professionally, and I never had that level of sharpening Erwin shows.

Erwin's stuff is interesting and worth the read, but I don't really feel that the information is complete, especially it is hard to know exactly what he means.

Best regards
Erik

I'm glad to hear that someone else is confused by Erwin's tests, because I too have done a fair amount of non-professional work with Imatest, and was shocked and perplexed by his MTFs of 250% on many images. His Leica results also are perplexing. At f/2.8 without sharpening (Image L1005878.jpg) shows an MTF of 250%! I was under the impression that MTFs of greater than 100% are caused by sharpening (Norman Koren (http://forums.imatest.com/index.php?topic=708.0)). In my experience such MTFs result from gross oversharpening and one is likely looking at sharpening artifacts and not true resolution. With Imatest, over-sharpening can result in resolution above Nyquist. Personally, I would not trust such results, but perhaps Erwin knows something that I don't.
Title: AA-filtering CCD and CMOS
Post by: telyt on October 15, 2009, 09:34:33 am
Quote from: ErikKaffehr
Regarding testing I would say that using the same converter for all tested cameras is good science, I may even add that using a known, well documented and basic converter is optimal.

Different converters will provide different results with different cameras' raw files and if you want to test the potential of the system then the optimum for that system should be used.  By using the same converter with all raw files all you're testing is a particular set of tools which might not be optimum for any of the tested systems.  In my experience with the DMR for example optimum results are with Imacon Flexcolor, next would be C1, then ACR.
Title: AA-filtering CCD and CMOS
Post by: ejmartin on October 15, 2009, 10:07:33 am
Quote from: telyt
Different converters will provide different results with different cameras' raw files and if you want to test the potential of the system then the optimum for that system should be used.  By using the same converter with all raw files all you're testing is a particular set of tools which might not be optimum for any of the tested systems.  In my experience with the DMR for example optimum results are with Imacon Flexcolor, next would be C1, then ACR.

Depends on what you're trying to achieve.  If the goal is to get the most out of a given camera's RAW files then indeed different converters might be appropriate.  If the goal is to discern what of the output can be attributed to the effects of an AA filter and lack or presence of aliasing, then the use of different converters throws an additional set of uncontrolled variables into the testing, that make it hard to know whether what one is seeing is due to say different demosaic algorithms vs the effect of aliasing.
Title: AA-filtering CCD and CMOS
Post by: bjanes on October 15, 2009, 10:16:28 am
Quote from: telyt
Different converters will provide different results with different cameras' raw files and if you want to test the potential of the system then the optimum for that system should be used.  By using the same converter with all raw files all you're testing is a particular set of tools which might not be optimum for any of the tested systems.  In my experience with the DMR for example optimum results are with Imacon Flexcolor, next would be C1, then ACR.


What you are saying is true, but one can not necessarily assume that the camera maker's raw converter is the best for their camera. One would think that the camera maker should know how to get the best results from their cameras, but this is not necessarily the case. The maker may use a Sony or Kodak chip and have little expertise in software development. One would then have to test with all available raw converters, and this would greatly complicate the testing, especially when one considers that each converter has multiple parameters. Sharpening is not standardized. Would you omit sharpening (and penalize cameras with a blur filter) or try to determine optimal sharpening?

Since the tester might not want to buy all possible raw converters, many choose to use the camera's JPEG engine or a well regarded raw converter that supports multiple cameras such as C1, DCRaw or ACR, often with default settings. In any case, the results will be subject to criticism by some.
Title: AA-filtering CCD and CMOS
Post by: aaykay on October 15, 2009, 12:35:33 pm
Quote from: BernardLanguillier
Having worked with files from the D3x (an awful lot) and the A900 (a lot less but still enough to form an opinion), I see differences that are as significant as it gets between DSLRs of similar pixel count. Put it otherwise, if the A900 and D3x are the same for Michael, then he doesn't need bother looking at any new DSLR in next 2 or 3 years, they will be the same as well.

The default JPEGs that the A900 spits out (even in the "Extra Fine" mode) are pretty lousy, for any kind of comparison with the output of any other camera.  I personally never shoot JPEG with the A900, other than the throwaway lowest resolution JPEG that is shot along with the RAW+JPEG shooting mode.

Also, when you shoot RAW, a common finding with Sony files when processed with any of the Adobe-based products like CS4 or Lightroom is that they just mangle the Sony RAW data, resulting in disgusting results.  Don't have any idea why.  Is there something within the Adobe RAW processing engine that are not optimized for Sony files ? Why is it destroying fine detail that the other RAW processors have no difficulty in revealing ?  No idea at all, but I would not form any judgement on the quality of the A900 files, based on any kind of processing done via Adobe products.

Now, what else ?  I find DXO (especially the latest 5.3.5 release) to be the best RAW processing tool for A900 files.  Some people find Capture One Pro to be good but I found DXO to be better.  Then there are other people who swear by packages like Rawtherapee etc for A900 files and if you use Apple Mac, then RPP (Raw Photo Processor) is supposed to be top-notch.

Bottomline, any opinions formed by processing A900 files via a sub-optimal RAW processor, is not revealing what the product can do, especially when you stick some of the best lenses in dSLR-dom on the camera.  
Title: AA-filtering CCD and CMOS
Post by: BernardLanguillier on October 15, 2009, 06:32:04 pm
Quote from: aaykay
Bottomline, any opinions formed by processing A900 files via a sub-optimal RAW processor, is not revealing what the product can do, especially when you stick some of the best lenses in dSLR-dom on the camera.  

My tests were done with C1 Pro 4.8 with the exact same settings I use for my d3x files.

Cheers,
Bernard
Title: AA-filtering CCD and CMOS
Post by: telyt on October 15, 2009, 09:29:02 pm
Quote from: bjanes
Since the tester might not want to buy all possible raw converters, many choose to use the camera's JPEG engine or a well regarded raw converter that supports multiple cameras such as C1, DCRaw or ACR, often with default settings. In any case, the results will be subject to criticism by some.

And it will be an incomplete test.  However, the internet buzz would turn into "Tester A says Camera C sux".
Title: AA-filtering CCD and CMOS
Post by: telyt on October 15, 2009, 09:31:14 pm
Quote from: ejmartin
Depends on what you're trying to achieve.  If the goal is to get the most out of a given camera's RAW files then indeed different converters might be appropriate.  If the goal is to discern what of the output can be attributed to the effects of an AA filter and lack or presence of aliasing, then the use of different converters throws an additional set of uncontrolled variables into the testing, that make it hard to know whether what one is seeing is due to say different demosaic algorithms vs the effect of aliasing.

If you want to see what the effect of the AA filter is take two copies of a camera, verify that their output is identical, then have the AA filter removed from one of them and re-test.  The test would be valid only for that particular camera model, otherwise you have too many uncontrolled variables.
Title: AA-filtering CCD and CMOS
Post by: ErikKaffehr on October 15, 2009, 09:37:16 pm
Hi,

I don't want to engage in any Nikon versus Alpha wars. I would just say that:

1) The lens that Bernard uses is of stellar reputation, at least if it's  the 100/2.8 Macro Planar, It's probably one the three or four best lenses ever made for 135 format cameras.

2) Comparing systems with the same converter is what any scientist would consider good science.

3) I have only the 24-70/2.8 ZA lens my self, and I may not consider it my sharpest zoom. The 135/1.8ZA is said to be stellar but the 85/1.4 less so.

4) I have downloaded testshots for the D3X and the Alpha 900 from Imaging Resource and converted myself using Lightroom and also measured using Imatest. There was not a lot of difference.

I'd also add that I made some experiments comparing MFDB Imags dlownloaded from the net with DSLRs. The experiments were based on A2-prints. A verbal description of the outcome is here

http://83.177.178.241/ekr/index.php/photoa...ed#Experiment_1 (http://83.177.178.241/ekr/index.php/photoarticles/24-how-many-megapixels-do-we-need#Experiment_1)
and
http://83.177.178.241/ekr/index.php/photoa...ed#Experiment_2 (http://83.177.178.241/ekr/index.php/photoarticles/24-how-many-megapixels-do-we-need#Experiment_2)

Two interesting observations are:

1) It's difficult to tell apart Nikon 3DX and MFDBs at A2 size or smaller (16.5 x 23.4 in)
2) Stopping down the lens to f/22 on MFDB by and significantly reduces the advantage of MFDB over DSLR stopped down to f/8 (this comparison was based on larger print size, around 24x34"). This depends solely on diffraction.

Best regards
Erik

Quote from: BernardLanguillier
My tests were done with C1 Pro 4.8 with the exact same settings I use for my d3x files.

Cheers,
Bernard
Title: AA-filtering CCD and CMOS
Post by: telyt on October 15, 2009, 10:55:55 pm
Quote from: ErikKaffehr
2) Comparing systems with the same converter is what any scientist would consider good science.

pardon my language, but speaking as a scientist that's bullshit.  The only reason to do that is if the one converter you're using is the only one you ever expect to use.  This test would be "how do cameras B and C compare when using only one converter?", not "how do Camera B and Camera C compare?"  The test with only one converter would transmogrify on the internet to "Tester A says Camera C sux" when it's really Camera C doesn't do as well as Camera B when files are converted with Converter D.  Are you only interested in Converter D?  If so then the test you propose is fine.  If you want to see how well each system performs, the test with one converter doesn't tell you the whole story.
Title: AA-filtering CCD and CMOS
Post by: ejmartin on October 15, 2009, 11:35:23 pm
Quote from: telyt
If you want to see what the effect of the AA filter is take two copies of a camera, verify that their output is identical, then have the AA filter removed from one of them and re-test.  The test would be valid only for that particular camera model, otherwise you have too many uncontrolled variables.

And this is supposed to be a justification for changing both the camera and the converter, as you originally suggested?  
Title: AA-filtering CCD and CMOS
Post by: ejmartin on October 15, 2009, 11:50:40 pm
Quote from: telyt
pardon my language, but speaking as a scientist that's bullshit.  The only reason to do that is if the one converter you're using is the only one you ever expect to use.  This test would be "how do cameras B and C compare when using only one converter?", not "how do Camera B and Camera C compare?"  The test with only one converter would transmogrify on the internet to "Tester A says Camera C sux" when it's really Camera C doesn't do as well as Camera B when files are converted with Converter D.  Are you only interested in Converter D?  If so then the test you propose is fine.  If you want to see how well each system performs, the test with one converter doesn't tell you the whole story.

Speaking as a scientist, that's bullshit  

The test Erik proposes would be, "how do cameras B and C compare when their RAW data are subjected to identical manipulations", not "how do cameras B and C compare when treated in uncontrolled different ways?" as you seem to prefer.   If the converter is a potential source of bias, then it is better to use a family of converters on each camera to eliminate that bias, so long as it is known that any converter used applies the same process to any given RAW file.  However, since only dcraw is open source, that's hard to verify for any other converter.  RAW conversion is not a black art, which needs sympathetic vibes between conversion algorithm and camera.
Title: AA-filtering CCD and CMOS
Post by: telyt on October 16, 2009, 12:36:31 am
Quote from: ejmartin
Speaking as a scientist, that's bullshit  

The test Erik proposes would be, "how do cameras B and C compare when their RAW data are subjected to identical manipulations", not "how do cameras B and C compare when treated in uncontrolled different ways?" as you seem to prefer.   If the converter is a potential source of bias, then it is better to use a family of converters, so long as it is known that any converter used applies the same process to any given RAW file.  However, since only dcraw is open source, that's hard to verify for any other converter.  RAW conversion is not a black art, which needs sympathetic vibes between conversion algorithm and camera.

Would you compare an E-6 film and a C-41 film using XTol as a developer, and then conclude anything meaningful about the films' performance?

The problem with the test Eric proposes is that different camera systems require different manipulations to get the most out of the system.  different color profiles, and different capture sharpening for example.  What's the purpose of the test?  It it to compare camera systems?  if so limiting the test to a single converter provides only scanty information, and unfortunately the unscientific voices on the internet (and there are a few of these   ) don't recognize the difference between a test limited to a single converter and a comprehensive test.

An uncontrolled test, BTW, would involve applying random converters to equally random cameras' files.  You have mis-represented my position.  What I'm suggesting is that in order to compare optimum performance of two or more camera systems, optimize each system first then compare, otherwise the test's conclusion will more than likely be erroneous.
Title: AA-filtering CCD and CMOS
Post by: Slough on October 16, 2009, 05:05:44 am
Quote from: ejmartin
Speaking as a scientist, that's bullshit  

The test Erik proposes would be, "how do cameras B and C compare when their RAW data are subjected to identical manipulations", not "how do cameras B and C compare when treated in uncontrolled different ways?" as you seem to prefer.   If the converter is a potential source of bias, then it is better to use a family of converters on each camera to eliminate that bias, so long as it is known that any converter used applies the same process to any given RAW file.  However, since only dcraw is open source, that's hard to verify for any other converter.  RAW conversion is not a black art, which needs sympathetic vibes between conversion algorithm and camera.

Depends. If you are trying to analyse something inherent in each camera, then yes use the same RAW converter. That way you minimise the number of different variables. But if you want an answer that applies to the real world, such as "Which camera is capable of the best resolution and lowest artifacts", then choose the optimal RAW converter for each. After all, when someone buys camera A, will the clouds part to reveal an elderly gentleman who goes on to say in a deep voice "Though shalt only use converter X with that camera, as should have been decreed by Moses, but he missed that one off his list due to a poor retention span and forgetting to make notes, even though I warned him 11 commandments was an awful lot to remember, what with the excitement of meeting his God for the first time, oh well, now where was I, oh yes, I must get back to that game of mah jong with Beelzebub".
Title: AA-filtering CCD and CMOS
Post by: bjanes on October 16, 2009, 08:54:40 am
Quote from: telyt
Would you compare an E-6 film and a C-41 film using XTol as a developer, and then conclude anything meaningful about the films' performance?

The problem with the test Eric proposes is that different camera systems require different manipulations to get the most out of the system.  different color profiles, and different capture sharpening for example.  What's the purpose of the test?  It it to compare camera systems?  if so limiting the test to a single converter provides only scanty information, and unfortunately the unscientific voices on the internet (and there are a few of these   ) don't recognize the difference between a test limited to a single converter and a comprehensive test.

An uncontrolled test, BTW, would involve applying random converters to equally random cameras' files.  You have mis-represented my position.  What I'm suggesting is that in order to compare optimum performance of two or more camera systems, optimize each system first then compare, otherwise the test's conclusion will more than likely be erroneous.

Doug,

Have you actually used your extended optimization procedure for any two or more cameras with systematic manipulation of the raw converter settings for each of multiple raw converters? If so, I'm surprised that you have time to participate in this thread. As you know, raw converters usually have defaults that differ for each individual camera, and these may not be optimum for the given camera. Controlling all these variables would be a daunting task and you would still have uncontrolled variables. We are discussing aliasing, and the lens is an important factor: if the lens can't resolve above Nyquist, there will be no aliasing. What is the optimum f/stop for each lens? Do you use the same lens for all cameras? Some raw converters correct lens aberrations for their brand of camera with certain of their lenses. Sharpening is also critical, since cameras with blur filters require more sharpening than cameras lacking them. Different raw converters have different sharpening.

I don't know your background, and your understanding of the scientific method, but I would remind you that Emil is professor of physics at one of our best research institutions (the University of Chicago) and he knows more about the scientific method and experimental design than most of us. Personally, I give his posts a great deal of weight as compared to those by persons unknown to me.

As Emil suggests, it might be prudent to test with a number of well regarded raw converters to determine how significant this variable is. If the results are similar or can be correlated, you might decide to settle on one raw converter. This is the approach of most testers who post results on the net. If you get the best results with the camera maker's raw converter, you still do not know if these good results are due to the camera or raw converter (which does not support any other brand of camera).



Title: AA-filtering CCD and CMOS
Post by: thierrylegros396 on October 16, 2009, 09:19:31 am
To make good comparisons, we need first a well accepted definition of "optimum sharpening" and means to check it.

Sometimes I'm amazed with definition graph with a lot of "moiré" and other artifacts, and people who are able to tell : 2300 LPH absolute resolution.

So yes we first need a well accepted repetitive procedure !


Have a Nice Day.

Thierry
Title: AA-filtering CCD and CMOS
Post by: bjanes on October 16, 2009, 09:22:44 am
Quote from: Slough
After all, when someone buys camera A, will the clouds part to reveal an elderly gentleman who goes on to say in a deep voice "Though shalt only use converter X with that camera, as should have been decreed by Moses, but he missed that one off his list due to a poor retention span and forgetting to make notes, even though I warned him 11 commandments was an awful lot to remember, what with the excitement of meeting his God for the first time, oh well, now where was I, oh yes, I must get back to that game of mah jong with Beelzebub".

Usually I steer clear of religion when discussing photography, but couldn't resist the following analogy. Actually, the Torah contains not 10 but 613 commands. I believe it was St. Paul who taught that trying to follow all 613 commands forced one to sin, since no human could possibly fulfill all of those commands. Accordingly, St. Paul and St. Peter condensed Mosiac law into simpler constructs. Some photographers are attempting to control all variables, but inevitably fail.
Title: AA-filtering CCD and CMOS
Post by: telyt on October 16, 2009, 11:22:11 am
Quote from: bjanes
As you know, raw converters usually have defaults that differ for each individual camera, and these may not be optimum for the given camera. Controlling all these variables would be a daunting task and you would still have uncontrolled variables. We are discussing aliasing, and the lens is an important factor: if the lens can't resolve above Nyquist, there will be no aliasing. What is the optimum f/stop for each lens? Do you use the same lens for all cameras? Some raw converters correct lens aberrations for their brand of camera with certain of their lenses. Sharpening is also critical, since cameras with blur filters require more sharpening than cameras lacking them. Different raw converters have different sharpening.

You've expressed exactly why a single raw converter should not be used to compare camera systems.

Quote from: bjanes
I don't know your background, and your understanding of the scientific method

degrees in biological sciences and engineering, working in engineering software development.  I understand all to well how a single software package can and will treat different data sources differently.  Using a single raw converter in no way eliminates any variables whatsoever, it's only value is the tester's convenience.

Quote from: bjanes
it might be prudent to test with a number of well regarded raw converters to determine how significant this variable is. If the results are similar or can be correlated, you might decide to settle on one raw converter. This is the approach of most testers who post results on the net. If you get the best results with the camera maker's raw converter, you still do not know if these good results are due to the camera or raw converter (which does not support any other brand of camera).

Does it matter whether the best results are from the software or from the hardware?  My impression is that the goal is optimum results, not optimum hardware by itself or optimum software by itself.  The two are inseparable, so if you are really interested in results, i.e., the photograph, find the optimum of each combination.  If the goal is optimal overall results, it doesn't matter that "most  testers who post results on the net" use a singe raw converter, it's a lazy suboptimal shortcut no matter how many people do it.
Title: AA-filtering CCD and CMOS
Post by: bjanes on October 16, 2009, 11:30:48 am
Quote from: telyt
Does it matter whether the best results are from the software or from the hardware?  My impression is that the goal is optimum results, not optimum hardware by itself or optimum software by itself.  The two are inseparable, so if you are really interested in results, i.e., the photograph, find the optimum of each combination.  If the goal is optimal overall results, it doesn't matter that "most  testers who post results on the net" use a singe raw converter, it's a lazy suboptimal shortcut no matter how many people do it.

I'm waiting to see your optimal set of tests.
Title: AA-filtering CCD and CMOS
Post by: Slough on October 16, 2009, 11:42:18 am
Quote from: bjanes
Usually I steer clear of religion when discussing photography, but couldn't resist the following analogy. Actually, the Torah contains not 10 but 613 commands. I believe it was St. Paul who taught that trying to follow all 613 commands forced one to sin, since no human could possibly fulfill all of those commands. Accordingly, St. Paul and St. Peter condensed Mosiac law into simpler constructs. Some photographers are attempting to control all variables, but inevitably fail.

   So Christianity is Judaism 'lite'? (And Church of England is for people who like coffee mornings, and songs, but aren't so fond of the religious bits.) I'd better stop there before I offend too many people ...
Title: AA-filtering CCD and CMOS
Post by: ejmartin on October 16, 2009, 01:01:57 pm
telyt,  I'm curious to know how a demosaic algorithm can be tuned to do better with an image depending on whether it has been taken with or without an AA filter on the sensor.
Title: AA-filtering CCD and CMOS
Post by: aaykay on October 16, 2009, 01:02:30 pm
Quote from: telyt
pardon my language, but speaking as a scientist that's bullshit.  The only reason to do that is if the one converter you're using is the only one you ever expect to use.  This test would be "how do cameras B and C compare when using only one converter?", not "how do Camera B and Camera C compare?"  The test with only one converter would transmogrify on the internet to "Tester A says Camera C sux" when it's really Camera C doesn't do as well as Camera B when files are converted with Converter D.  Are you only interested in Converter D?  If so then the test you propose is fine.  If you want to see how well each system performs, the test with one converter doesn't tell you the whole story.

I agree with this completely.  Many review sites make the mistake of using a "common converter" (invariably using Adobe based products for this purpose) with "identical settings" to process RAW images from different cameras from different makers, without realizing that there are differences "under the hood" in the way the "common converter" deals with these different RAW files.

Using a "common converter with identical settings" to determine how Camera A compares to Camera B is the commonest mistake that several review sites (even reputed ones) seem to be making, since using a "common converter with identical settings" will only prove how the 'common converter' handles files from Camera A and Camera B - not what Camera A or Camera B can do.  

Bottomline, the answer to how "Camera A compares to Camera B" can only be determined by choosing the optimum converter for Camera A (with settings appropriate for Camera A) and choosing the Optimum converter for Camera B (with settings appropriate for Camera  and then comparing the resulting outputs.   Obviouslly there is still the wiggle room that even the best RAW converters of today may be bettered by the ones available tomorrow and the results might be different tomorrow.


Title: AA-filtering CCD and CMOS
Post by: nma on October 16, 2009, 02:13:17 pm
I have had a late start but I am trying to follow the twists and turns of this thread.

The virtues of the AA filter have been affirmed and denied, with religious certainty on both sides. One argument in favor of removing or omitting the AA filter in landscape photography is that the fine geometric patterns, such as in fabrics, are not present in the natural scene and thus aliasing is negligible.  However, sampling theory, as well as simulation, shows that this statement is false.  When an image is under sampled, components above the Nyquist frequency are erroneously used in forming the digital image, appearing as lower frequency information.  This is aliasing. Even if one doesn’t recognize it as such, it is a distortion of the image.  These distortions might be pleasing in some case and lead some to prefer a camera without an AA filter. But one cannot be certain that the camera without the AA filter will always yield the more pleasing result; it will depend strongly on the scene being photographed.  

I think it is fair to say that the digital sampling of the sensor puts a limit to the resolution that can be obtained. One should not be surprised to learn that it is impossible to get perfect fidelity in rendering digital images formed with a finite number of samples. The sampling frequency of the sensor and the MTF of the lens must be consistent or the results might not be acceptable. Some very high resolution lenses may not be well matched to every sensor design sans AA filter.

This argument could also  be turned around in favor of designs without an AA filter. By using a sensor with very high sampling, one could in principle, set the Nyquist frequency to some reasonable level, and filter out the information above Nyquist. Then the image could be downsampled as desired. In principal the sensor with high sampling could record the same total number of photons as one with lower sampling (that’s an engineering challenge for sure), so the ultimate IQ would not be worse.



Title: AA-filtering CCD and CMOS
Post by: bjanes on October 16, 2009, 02:30:17 pm
Quote from: aaykay
Bottomline, the answer to how "Camera A compares to Camera B" can only be determined by choosing the optimum converter for Camera A (with settings appropriate for Camera A) and choosing the Optimum converter for Camera B (with settings appropriate for Camera  and then comparing the resulting outputs.   Obviouslly there is still the wiggle room that even the best RAW converters of today may be bettered by the ones available tomorrow and the results might be different tomorrow.

So now for a proper comparison of camera A to camera B we must test both with all known raw converters with all possible settings and then mark the results as tentative, awaiting the next generation of raw converters.

To get anywhere, I think that it will necessary to streamline the process so as to control the most important variables as determined by preliminary testing.
Title: AA-filtering CCD and CMOS
Post by: telyt on October 16, 2009, 02:42:52 pm
Quote from: bjanes
So now for a proper comparison of camera A to camera B we must test both with all known raw converters with all possible settings and then mark the results as tentative, awaiting the next generation of raw converters.

Yes.  It's always been that way, but most in the general population have too weak a grasp of the variables involved to understand how flawed the single-converter tests are.
Title: AA-filtering CCD and CMOS
Post by: jing q on October 16, 2009, 04:23:44 pm
Quote from: nma
I have had a late start but I am trying to follow the twists and turns of this thread.

The virtues of the AA filter have been affirmed and denied, with religious certainty on both sides. One argument in favor of removing or omitting the AA filter in landscape photography is that the fine geometric patterns, such as in fabrics, are not present in the natural scene and thus aliasing is negligible.  However, sampling theory, as well as simulation, shows that this statement is false.  When an image is under sampled, components above the Nyquist frequency are erroneously used in forming the digital image, appearing as lower frequency information.  This is aliasing. Even if one doesn’t recognize it as such, it is a distortion of the image.  These distortions might be pleasing in some case and lead some to prefer a camera without an AA filter. But one cannot be certain that the camera without the AA filter will always yield the more pleasing result; it will depend strongly on the scene being photographed.

thank you, this could possibly sum everything up.
perhaps having no AA filter leads to "distortions", but in photography most of us never strove to capture a scene exactly as it is.
There is something to be said for the intangible sense of satisfaction one gets from seeing an image straight out of the camera looking "right" without having to post process the hell out of it.

Many people shoot not for scientific reasons, but we shoot to get images which please us, looking for the right cameras which render images the most pleasing way possible. Most people are lazy and prefer not to have to fiddle around so much with images to get them to look good.
And there is something pleasing about the way an AA filterless camera renders images.

I think the vehement disagreement regarding this topic boils down to a few things:
we have a few very scientific minded people here who regard it sacrilegious that some photographers would actually think that aliasing is acceptable and pleasing, and believe that one should strive for a theoretically perfect camera output with the best lens and raw converter possible.

there are others who simply just prefer the look of an AA filterless picture, and perhaps use scientifically inaccurate terms to describe the reason why they feel these images are better.Perhaps *shock* we don't really care for accurate photographs!

Thus the twain shall never meet. I'm happy to stay in the latter camp as long as my images look good.

I'll leave the theoretical debating to you guys  
Title: AA-filtering CCD and CMOS
Post by: madmanchan on October 17, 2009, 11:41:05 am
Even single raw converters will treat different camera models differently. For example, a given raw converter may apply different demosaic, sharpening, and NR methods to different models.
Title: AA-filtering CCD and CMOS
Post by: Slough on October 17, 2009, 12:10:40 pm
Quote from: madmanchan
Even single raw converters will treat different camera models differently. For example, a given raw converter may apply different demosaic, sharpening, and NR methods to different models.

Exactly. So using one converter for all cameras is not much different from using different converters. And in any case, it seems to be, rightly or wrongly, that the difference between different RAW converters are often small as far as resolution goes. Where they differ is high ISO performance (noise), and colour rendition (which is why I use Nikon Capture despite the horrendous bugs that any self respecting software house would fix). If the differences in resolution are so small, then what size do you have to print to such that they become readily discernible to ordinary people as well as the socially challenged?

By the way, what was the question again? Or rather the point of this intellectual navel gazing?
Title: AA-filtering CCD and CMOS
Post by: madmanchan on October 17, 2009, 04:42:34 pm
I think the larger point is that the whole imaging chain is complex. It is hard to isolate variables. Even when you can, it can be hard to understand how a given design point (e.g., a sensor's microlenses, or lack thereof) affect the final result. As an analogy, scientists try to attribute certain types of health benefits to specific nutrients (or types of foods) but it's very hard to understand the real effects because they are only one part of a much larger, very complex system (e.g., overall health of the individual, interaction with other foods, daily activities, etc.).
Title: AA-filtering CCD and CMOS
Post by: ejmartin on October 17, 2009, 05:18:41 pm
Quote from: madmanchan
Even single raw converters will treat different camera models differently. For example, a given raw converter may apply different demosaic, sharpening, and NR methods to different models.

Yes, but there are others that don't.
Title: AA-filtering CCD and CMOS
Post by: madmanchan on October 17, 2009, 10:41:41 pm
They probably should, though, Emil. For example, I would be surprised if a demosaic method that works on a 3-channel Bayer camera works equally well on a 4-channel camera. There are also different processing requirements for Foveon, and Fuji cameras. All of these affect the initial level of quality. There's also a ton of pre-demosaic processing that needs to happen on many medium format files; otherwise the image is unusable.

In any case, the point I'm trying to make is that users conducting detailed tests are generally not able to tell (at least, not easily) whether a given raw converter treats all the models the same way ...
Title: AA-filtering CCD and CMOS
Post by: JeffKohn on October 18, 2009, 12:32:51 am
Hey Eric, when are you guys going to give us the option to disable the hidden 'basline' adjustments. ACR wouldn't get such a bad rap if it wasn't for this, IMHO. The Exposure adjustment can easily be reversed, but there's really no way to defeat the noise reduction and it hurts ACR in comparison to other converters.