Luminous Landscape Forum
Equipment & Techniques => Medium Format / Film / Digital Backs – and Large Sensor Photography => Topic started by: Chris Barrett on December 07, 2013, 10:17:34 am
-
So, I recently picked up this little 36 MegaPixel toy from Sony. I call it a toy because I just can't get over it's minuscule size. I've been testing it very casually, just going out and shooting and seeing what I get. Nothing very scientific. I'm using EOS mount lenses on it via the Metabones Smart Adapter III. It performs beautifully with the Canon Tilt / Shifts... allowing it to be what I wish the D800 had been (a high MP little camera with excellent T/S glass). The A7r actually has the same chip as the D800 (as I understand it) which Sony manufactures. Since it's mirrorless, it has a super short flange distance and you can adapt practically any lens to it, including Leica M glass.
Anyway, here are some shots from my living room:
Phase One IQ 260 / Arca Swiss Rm3d / Schneider Apo-Digitar 35mm/f5.6 XL *Processed in C1 Pro
(http://christopherbarrett.net/forum_images/CameraTests_Phase1_IQ260.jpg)
Sony A7r / Canon 24mm TS-E II *Processed in LightRoom 5
(http://christopherbarrett.net/forum_images/CameraTests_Sony_A7r.jpg)
Canon 5d2 / 24mm TS-E II *Processed in C1 Pro
(http://christopherbarrett.net/forum_images/CameraTests_Canon_5d2.jpg)
Red Epic / Canon 17mm TS-E *Processed in RedCine-X
(http://christopherbarrett.net/forum_images/CameraTests_Red_Epic.jpg)
Right away I can see the difference in the C1P vs LR processing. C1P always seems to have snappier color to me whereas LR can usually dig out a little more Hilight info. I set the hilight recovery sliders at 30 in C1P and LR for all the still cameras. All three still cameras have comparable sharpness with the IQ back being a clear winner upon pixel peeping (no surprise). I turned off all in-software sharpening and the little cameras have neutral profiles in camera. I don't find the Red that useful for my stills work. It possesses much more shadow noise and much less sharpness than the still cameras but It's fun to see how it blows them away on DR. I think it was really designed to feel like Neg Stock.
You can download all the RAW files or my processed TIFFs via DropBox and draw your own conclusions: https://www.dropbox.com/sh/coeaopdb1srglbe/jdi7ZiT4IX (https://www.dropbox.com/sh/coeaopdb1srglbe/jdi7ZiT4IX)
In conclusion... this was really all about evolving my own workflow and making life a little simpler and well... hell I'm a big gear nerd (but you knew that). Speaking of workflow, the only way I've found to shoot the Sony tethered is to use their Remote Camera Control software and setup LightRoom to auto import the files from a watched folder. Lame, but usable. C1P support is coming, they tell me.
CB
-
Hi Chris,
I'm not an expert on procesing Phase One files nor A7R, but I did a C1 pro test. Here is a parcial of the image to 100%.
(https://dl.dropboxusercontent.com/u/10120389/Screen-Shot-2013.jpg)
The left IQ in the screen shows standard noise reduction and the right one strong noise reduction. I don't know if I'm doing something wrong, but the IQ is really noisy.
As I can see the A7R (in the middle) shadows and extreme high lights are easier to handle. The IQ 260 is a lot crispier as supposed.
Please comment as I can recall some adjustments.
Thanks. ACH
-
The RED image looks the best. It would be interesting to see how much highlight recovery one can do with the other cameras but right off the bat the RED image had the best (visible) dynamic range by far. Handled the scene beautifully.
From best to worst (image posted which is obviously a jpg and low res):
RED-->IQ260-->A7R-->5D2
Kinda what you would expect although I wish the IQ260 would do better in DR given its a dedicated stills camera and more expensive than the RED.
-
Hi Chris,
I'm not an expert on procesing Phase One files nor A7R, but I did a C1 pro test. Here is a parcial of the image to 100%.
(https://dl.dropboxusercontent.com/u/10120389/Screen-Shot-2013.jpg)
The left IQ in the screen shows standard noise reduction and the right one strong noise reduction. I don't know if I'm doing something wrong, but the IQ is really noisy.
As I can see the A7R (in the middle) shadows and extreme high lights are easier to handle. The IQ 260 is a lot crispier as supposed.
Please comment as I can recall some adjustments.
Thanks. ACH
The A7R image looks very low contrast, hence the softer transition to highlights which are not really that light and more gray, while the IQ260 image looks much more contrasty. I would check default conversion curves.
-
Deleted. Apologies.
-
Hmm, less noisy on my system...
(http://christopherbarrett.net/forum_images/Screen%20Shot%202013-12-07%20at%209.42.48%20AM.png)
Keep in mind this is a 200% magnification, so I don't expect the IQ file to look flawless. Other contributing factors: ISO was 100 across the board (50 would be smoother on the IQ). Also, WB was around 3600 and I find the Phase 1 products to be challenged when pushing the blue channel. That's made me consider shooting through an 80D on many occasions. No doubt the A7r feels smoother and flatter. I'd chalk that up to the difference in converters.
Capture 1 profiles were set at "Film Standard"
-
Hi Chris,
an additional question: How much shift is possible with the 17TSE and the 24TSE on the A7R via the metabones adapter before vignetting of the adapter or the camera mount?
Thx Christoph
-
17 & 24 Full shift in all directions; no cutoff.
-
Keep in mind this is a 200% magnification, so I don't expect the IQ file to look flawless. Other contributing factors: ISO was 100 across the board (50 would be smoother on the IQ). Also, WB was around 3600 and I find the Phase 1 products to be challenged when pushing the blue channel. That's made me consider shooting through an 80D on many occasions. No doubt the A7r feels smoother and flatter. I'd chalk that up to the difference in converters.
Capture 1 profiles were set at "Film Standard"
The use of the native ISO 50 on the iq260 would have also increased the ability to retain highlights.
Not a criticism, just some additional info for anyone not familiar with the iq260.
Thanks for sharing!
-
thanks for putting up the raws.
-
The leather in the Eames chair is black in the 260 and the D5, but is brown(isn) in the SONY image - that is what I see any way.
Jerry
-
Interesting to see the difference in the noise, but I wonder if the 260 files would look better after they have been processed from raw into tiff file. I often find that the raw preview in C1 is a little rougher (especially when viewed at 100% or closer) then after the files has been processed.
-
Great stuff, thanks for sharing.
-
I don't find the Red that useful for my stills work. It possesses much more shadow noise and much less sharpness than the still cameras but It's fun to see how it blows them away on DR. I think it was really designed to feel like Neg Stock.
But for film work, where combining different frames/exposures as one can easily do with stills to be able to cope with bright lights or windows like in your demo shot, it's pretty amazing.
-
I don't find the Red that useful for my stills work. It possesses much more shadow noise and much less sharpness than the still cameras but It's fun to see how it blows them away on DR. I think it was really designed to feel like Neg Stock.
If you look at Kodak vision RED files seem to match, kind of the ants crawling up a wall look.
I never noticed it until I pulled a green screen key and first thought the key was off as I could see holes in dark areas like hair, until I realized that's just the RED look.
I had to mask the dark areas back into the clip to kill the noise in the darks. It's a pain, but most things with motion are a pain.
In my REDs the noise increases with heat and unless we're running multi cam I switch out cameras as we work to keep the noise down.
I learned this the hard way.
Anyway.
One thing I've found is the Quicktime Files from the RED 1 are sharper than the cinex files and though I've shown this image before, I processed out in cinex, then matched the size from the 4k quicktime clip and laid them in which gives the appearance of more sharpness.
(http://www.russellrutherford.com/rr_red1_file.jpg)
I also shot this medium format and technically the image was better, though I liked this pose from the RED, so I worked it for print.
BTW: Thanks for the test. I think the A7 file looks workable but as all things will have to test myself.
BC
-
Corrections made. Notice the files were saved as tiff. The one from IQ was C1 Processed and A7R was Camera Raw. Both open in photoshop and screen shot to sRGB for web viewing. ACH
(https://dl.dropboxusercontent.com/u/10120389/another.jpg)
-
IQ260
(https://dl.dropboxusercontent.com/u/10120389/IQ_260.jpg)
A7R
(https://dl.dropboxusercontent.com/u/10120389/Sony_A7r.jpg)
-
Hi Antonio,
I prefer the Sony.
Just look at all the noise in the blacks and the what I call "oil on water effect" in the blacks from the Phase back.
That is why I sold my Leaf Aptus back cause of the noise in the shadows. Made it useless for shooting cars.
Can’t wait for the Nikon D4x when it comes out next year.
Simon
-
Hi,
Thanks for sharing!
Best regards
Erik
So, I recently picked up this little 36 MegaPixel toy from Sony. I call it a toy because I just can't get over it's minuscule size. I've been testing it very casually, just going out and shooting and seeing what I get. Nothing very scientific. I'm using EOS mount lenses on it via the Metabones Smart Adapter III. It performs beautifully with the Canon Tilt / Shifts... allowing it to be what I wish the D800 had been (a high MP little camera with excellent T/S glass). The A7r actually has the same chip as the D800 (as I understand it) which Sony manufactures. Since it's mirrorless, it has a super short flange distance and you can adapt practically any lens to it, including Leica M glass.
Anyway, here are some shots from my living room:
Phase One IQ 260 / Arca Swiss Rm3d / Schneider Apo-Digitar 35mm/f5.6 XL *Processed in C1 Pro
(http://christopherbarrett.net/forum_images/CameraTests_Phase1_IQ260.jpg)
Sony A7r / Canon 24mm TS-E II *Processed in LightRoom 5
(http://christopherbarrett.net/forum_images/CameraTests_Sony_A7r.jpg)
Canon 5d2 / 24mm TS-E II *Processed in C1 Pro
(http://christopherbarrett.net/forum_images/CameraTests_Canon_5d2.jpg)
Red Epic / Canon 17mm TS-E *Processed in RedCine-X
(http://christopherbarrett.net/forum_images/CameraTests_Red_Epic.jpg)
Right away I can see the difference in the C1P vs LR processing. C1P always seems to have snappier color to me whereas LR can usually dig out a little more Hilight info. I set the hilight recovery sliders at 30 in C1P and LR for all the still cameras. All three still cameras have comparable sharpness with the IQ back being a clear winner upon pixel peeping (no surprise). I turned off all in-software sharpening and the little cameras have neutral profiles in camera. I don't find the Red that useful for my stills work. It possesses much more shadow noise and much less sharpness than the still cameras but It's fun to see how it blows them away on DR. I think it was really designed to feel like Neg Stock.
You can download all the RAW files or my processed TIFFs via DropBox and draw your own conclusions: https://www.dropbox.com/sh/coeaopdb1srglbe/jdi7ZiT4IX (https://www.dropbox.com/sh/coeaopdb1srglbe/jdi7ZiT4IX)
In conclusion... this was really all about evolving my own workflow and making life a little simpler and well... hell I'm a big gear nerd (but you knew that). Speaking of workflow, the only way I've found to shoot the Sony tethered is to use their Remote Camera Control software and setup LightRoom to auto import the files from a watched folder. Lame, but usable. C1P support is coming, they tell me.
CB
-
Chris you have some really good books in there! Thanks for sharing!
-
If you look at the raw data, IQ260 is exposed 2 stops lower compared to Sony and Canon.
-
Chris you have some really good books in there! Thanks for sharing!
And some dubious DVD's ;D
-
If you look at the raw data, IQ260 is exposed 2 stops lower compared to Sony and Canon.
Iliah,
I'm not sure I understand - the Phase back has a single ISO, right? So if set to 100 it should anyway be underexposed at least 1 stop in raw?
Could you please post the raw exposure graphs? This might also show us where the "ants" come from ...
Edmund
-
Hi Edmund,
Let's select the same black area on both Sony and IQ260 shots, I used the black screen on the chimney:
(http://cl.ly/image/140i1s1H0S3q/Sony_A7r-20131207-182348-RawDigger-ScreenShot.png)
EV0 on the following histograms is set at the respective sensor saturation points.
Here is the histogram of that area for IQ260, you can see blue is clipped:
(http://cl.ly/image/0o353A3b2u3n/IQ_260-Sel-4109-3846-339x446.png)
Here is the histogram for the same area for Sony A7r:
(http://cl.ly/image/2V313i0C3k3Y/Sony_A7r-Sel-3446-2962-277x367.png)
You can see that the peak on the green channel for Sony is at about -9 EV, while for IQ260 it is at about -10.5 EV; that makes for 1.5 EV difference in the green channel alone.
It may be interesting to compare the flare on both setups.
-
all I can say is wow, this test is very telling. thanks. I too will have to test that sony out, but at this point it looks like another game changer.
Chris have you shot any motion with the 7r yet?
-
One would think if you were going to go through the effort of careful numerical analysis and judging relative image performance you'd do so with a set of images that were shot at base ISO for each camera. No?
Chris was nice enough to post a casual test from which you can glean some good general conclusions (like they are all very capable cameras). But it's probably misleading to draw detailed conclusions from them.
-
I'm just looking at the images posted, but IMHO the IQ260 seems to have the most depth. The A7r and the Red the least.
-
Hi,
Well, both cameras have base ISO at 50. At ISO 100 DxO-mark measured 29 ISO for the IQ180 and 73 for. That corresponds to one stop. Phase backs seem to have a high ISO to protect the highlights.
Best regards
Erik
One would think if you were going to go through the effort of careful numerical analysis and judging relative image performance you'd do so with a set of images that were shot at base ISO for each camera. No?
Chris was nice enough to post a casual test from which you can glean some good general conclusions (like they are all very capable cameras). But it's probably misleading to draw detailed conclusions from them.
-
Hi,
I happened to see some pretty bad staircase aliasing on the A7r that is not visible on the IQ 260. Check the top of the lamp cord.
Regarding the images I would say the dark areas on the Sony are clear and I would say the IQ is noisy, but as it has been pointed out by Ilias there is 1+ step exposure difference. The IQ 260 is significantly sharper, but I would assume this comes from megapixels and perhaps also from the pixels being larger.
I checked out the images in Capture 1, too. I still feel the IQ260 image is noisy.
Best regards
Erik
-
When you expose to the left AND use non base iso on the phase backs that is what you get. Do the same test using iso 50 (on the IQ260, it is iso 35 on the 80mp models) and expose a bit more to the right (not much) and the shadows will be MUCH cleaner. The IQ image looks like what one would get with long exposures (in normal mode). Also which aperture did you use on the schneider 35? When you use iso 100 on the IQ backs its like cutting out 1 stop of data on the shadows side.
Color separation is best on the IQ file. The Canon is over saturating the reds everywhere it seems.
-
What about the cost factor?
A7r, metabones and a trio of 17,24,90 Tse lens (plus a Zeiss 50? Sigma 35 1.4 or 24-70 2.8?) =£?
Compared to iQ260, tech cam and a quartet of lenses =£?
-
This would be an interesting test for the IQ260 and the long exposure ISO setting of 140 which is the base iso for the long exposure setting. Then compare it to the ISO 100 shot.
Paul Caldwell
-
It is fascinating to see how an increasingly wider audience gets to realize the value of the Exmor technology as cameras equipped with those sensors decrease in price and become lens brand agnostic.
Cheers,
Bernard
-
Hi Bernard,
Not sure a camera that records 1775 values per channel (less than 11 bits) can replace a digital back in all cases.
-
Something else to point out is that I have found that MFDB increase in color noise much more quickly than luminous noise when ISO is increased. Whereas with CMOS based cameras, the opposite is true. We tend to notice color noise more than luminous noise because it is, well, more noticeable.
I am assuming we could increase noise filters in C1 and get ride of it, however this would decrease sharpness to an extent but probably not to the point where it is less sharp than the aR7 file.
-
Not sure a camera that records 1775 values per channel (less than 11 bits) can replace a digital back in all cases.
Hi Iliah,
Out of curiosity, how does that compare to a D800E/IQ280?
Cheers,
Bernard
-
Firstly, I agree with Doug; some of the close analysis attempted here would be best reserved for a comparison done with each camera at its best (base?) ISO speed setting. But thanks for this comparison too.
Anyway: Illiah, how is the EV scale defined on those histograms, and how does it handle raw output with different bit depths? If the histograms are counting EV's from maximum level down, the difference between 14-bit and 16-bit data would report the same numerical level with an EV difference of two stops. For example, a numerical level of 8 = 2^3 would be EV-11 relative to the top of a 14-bit scale but EV-13 relative to the top of a 16-bit scale.
Are the min and max values at right referring to the raw numerical levels? In that case, the IQ180 is actually giving absolute higher numerical levels, and the EV difference is merely a consequence of its 16-bit raw data having an additional two most significant bits, which of no relevance to the low-light region described in these histograms.
[EDIT: I do not understand that "blue clipping" at EV-14, from what AFAIK is a 16-bit signal: is that the existence of pixels with "0" value?]
-
This is definitely not a scientific test, but rather something that correlates more closely to my working methods. In many cases it's just not practical for me to shoot ISO 50. We often have people in our shots, are using strobe where multi-popping is counterproductive and occasionally shooting through a polarizer.
The Sony and Canon were shot at f/11 2 Sec Iso 100. The IQ 260 was shot at f/8 2/3. My initial exposure of 2 Sec just felt too bright on the IQ 260, so I used my 1 Sec exposure for the test. When I process the 2Sec exposure to match the tonality of the group it DOES have better shadow noise, but the hilights don't clip as nicely. I usually shoot to protect the hilights as much as possible (probably because I shot transparencies for 15 years).
Of course each of these images would only be at the very beginning of my postproduction flow and different techniques would be used to help each along. All would have darker windows blended in. I might revisit the brighter IQ260 file to use as a base and just as you can use darker exposures to recover hi lights, I sometimes use brighter exposures to clean up shadow noise (by processing them @ -X to match the base exposure and masking for just the shadows).
Perhaps the best test, for me, would be to proceed and retouch each to be the best possible image.
For now this tests has shown me exactly what I wanted to know... The A7r can and will replace my 5d2. While it won't replace my MFDB system altogether, it will take over some of that shooting.
Coming from years locked into shooting 4x5 Transparency film as the only possible workflow, I gotta tell ya, I'm damn giddy to have so many excellent tools at my disposal.
ps I've added the brighter IQ260 RAW file to the DropBox if anyone wants to play with that.
CB
-
As others have noted, This explains the 2 stops underexposure which Iliah is reading from the data for the Phase back.
In fact 2 stops underexposure will occur every time a back is used when set at 100 ISO (backs have no amplification, just use digital amplification). So 1 stop is lost from the "real 29" to the "fake "50 base ISO, and another is lost from ISO 50 to 100.
Of course an advantage of the back is that it can cleanly recover two stops of highlights in Raw.
Edmund
PS sorry about the redundant post I wanted to write this yesterday but sleep occurred.
Hi,
Well, both cameras have base ISO at 50. At ISO 100 DxO-mark measured 29 ISO for the IQ180 and 73 for. That corresponds to one stop. Phase backs seem to have a high ISO to protect the highlights.
Best regards
Erik
-
Hi Chris,
Thanks for posting!
Regarding scientific tests, DxO does that better than most of us.
Best regards
Erik
This is definitely not a scientific test, but rather something that correlates more closely to my working methods. In many cases it's just not practical for me to shoot ISO 50. We often have people in our shots, are using strobe where multi-popping is counterproductive and occasionally shooting through a polarizer.
The Sony and Canon were shot at f/11 2 Sec Iso 100. The IQ 260 was shot at f/8 2/3. My initial exposure of 2 Sec just felt too bright on the IQ 260, so I used my 1 Sec exposure for the test. When I process the 2Sec exposure to match the tonality of the group it DOES have better shadow noise, but the hilights don't clip as nicely. I usually shoot to protect the hilights as much as possible (probably because I shot transparencies for 15 years).
Of course each of these images would only be at the very beginning of my postproduction flow and different techniques would be used to help each along. All would have darker windows blended in. I might revisit the brighter IQ260 file to use as a base and just as you can use darker exposures to recover hi lights, I sometimes use brighter exposures to clean up shadow noise (by processing them @ -X to match the base exposure and masking for just the shadows).
Perhaps the best test, for me, would be to proceed and retouch each to be the best possible image.
For now this tests has shown me exactly what I wanted to know... The A7r can and will replace my 5d2. While it won't replace my MFDB system altogether, it will take over some of that shooting.
Coming from years locked into shooting 4x5 Transparency film as the only possible workflow, I gotta tell ya, I'm damn giddy to have so many excellent tools at my disposal.
ps I've added the brighter IQ260 RAW file to the DropBox if anyone wants to play with that.
CB
-
Hi Illiah,
I am somewhat surprised that A7r exif says 14 bit but RawDigger shows 12 bit data (going up to 1:4096) .
If we assume true 14 bit we would have 16384 values, but the significance of the LSBs (least significant bits) at high numbers would be very low. If we assume that FWC corresponds to 16384 the standard of deviation would be 128. It is quite obvious the last few bits only represent noise. On the low end more bits are used but the last two bits (or so) are still fairly stochastic, are they not?
Best regards
Erik
Hi Bernard,
Not sure a camera that records 1775 values per channel (less than 11 bits) can replace a digital back in all cases.
-
Hi Chris,
Thanks for posting!
Regarding scientific tests, DxO does that better than most of us.
Best regards
Erik
I feel like number 90 on the tennis player list when someone says "Federer plays better than most of us".
Edmund
-
Hi Erik,
What Sony are doing is applying a lossy compression curve - you can dump it and inspect if you wish. Data before compression is not truely 14-bit too, it is 13 bit. Here is the histogram of raw data for Sony as it is recorded in the file, before applying the decompression curve.
(http://cl.ly/image/0m2m1s3Y2G19/Sony_A7r-Full-7368x4920.png)
Looking at the right side you can see that the value range is 0 to 3546 while the number of levels is 1773, meaning we have missing values.
Now if we zoom to the data we can see that each second level is indeed missing
(http://cl.ly/image/3U2L173D3m19/Sony_A7r-Full-7368x4920Zoom.png)
If we look at the result after decompression (linearization) we can see that the gaps are wider now:
(http://cl.ly/image/3K0e0c2F2C13/Sony_A7r-Full-7368x4920ZoomDecompressed.png)
Inspecting different histogram region, it becomes obvious that for the shadow portion the width of the gaps is 1 level, and they become progressively wider towards highlights.
When EXIF says 14-bits, it does not mean all 14 bits are in use, only that the resulting data maximum corresponds to 14 bits. It is like a staircase, 1 meter staircase consisting of 7 steps each of 15 cm. It is true that the total height is 1 meter, but it does not mean the levels are each centimeter.
-
Hi Bernard,
If D800 is in lossless compression mode, the gaps are nearly non-existent, only white balance pre-conditioning adds them, and those are very few.
(http://cl.ly/image/1z2A361E0L1q/d800_post-Full-7378x4924.png)
The results with digital backs strongly depend on exposure, which includes the colour of light. For a relatively well-exposed regions the histogram is totally gapless:
(http://cl.ly/image/283O0g0M3o1Z/IQ_260-Full-8984x6732.png)
-
If the histograms are counting EV's from maximum level down, the difference between 14-bit and 16-bit data would report the same numerical level with an EV difference of two stops.
The maximum is normalized to "1". The absolute maximum value is not of importance anymore after normalization. The value of 0.5 is 1 stop below maximum; in absolute terms it means 16116/2= 8008 for Sony is equivalent to 65535/2= 32762 for IQ260.
You may want to look at this. Let's select two relatively close midrange regions on both shots and compute the difference.
Sony:
(http://cl.ly/image/1g0p2i0r2j1X/Sony_A7r-20131208-122315-RawDigger-ScreenShot.png)
Average values for the green channel for selected areas are 2269.36 and 1273.61 and the difference between 2 samples is log2(2269.36/1273.61) = 0.83EV
IQ260:
(http://cl.ly/image/162g3j2b0r1y/IQ_260-20131208-122455-RawDigger-ScreenShot.png)
Average values are 4330.84 and 1685.09 and the difference is log2(4330.84/1685.09) = 1.36 EV.
That is pretty significant change in light.
Now, taking into account maximum value for Sony is 16116, and maximum value for IQ260 is 65535 we can compute the difference in absolute terms. Lets' consider the brighter sample, 2269.36 for Sony and 4330.84 for IQ260. log2(16116/2269.36) = 2.83EV from saturation; log2(65535/4330.84) = 3.92 from saturation. That is they are 1 stop apart, and the difference seems to increase towards shadows: 3.66 EV for the darker sample on Sony shot, 5.28EV on the IQ260 shot.
The cause may be flare in metabones adaptor.
-
The maximum is normalized to "1". The absolute maximum value is not of importance anymore after normalization. The value of 0.5 is 1 stop below maximum; in absolute terms it means 16116/2= 8008 for Sony is equivalent to 65535/2= 32762 for IQ260.
You may want to look at this. Let's select two relatively close midrange regions on both shots and compute the difference.
Average values are 4330.84 and 1685.09 and the difference is log2(4330.84/1685.09) = 1.36 EV.
That is pretty significant change in light.
Thanks, but I still do not fully understand. As far as I now, these numbers are ADC output levels after the different cameras have likely applied different levels of amplification to the signal and so different conversion factors, whether measured in "photons counted per raw level" or the ratio between "photon count as a fraction of full well capacity" and "output level as a fraction of maximum level".
So how do these differently amplified output values tell us anything about the relative amount of light received by the sensors? We seem to need information relating back to photo-electron counts and well capacities in order to compare the amount of light received, either in absolute terms (photons counted) or in relative terms (fraction of full well capacity). Maybe that information is encoded in the ratio of DxO's sensitivity measurement to the ISO speed setting used.
Thinking about lens flare is very interesting though: having seen some measurements of lens flare with real world images which suggested that it is very hard to keep the darkest parts of the recorded image as far as 12 stops below the brightest, I am skeptical about the practical relevance of some extremes of DR measured in unnatural lab situations. For example, a DR test target with very bright areas adjacent to very dark ones might give different results that the usual graduated strip with the brightest and darkest regions at opposite ends.
-
Oh my God, my head hurts. These graphs make my IQ260 look really impressive. All I really care about in the end, though, is how my photographs look and whether my clients are happy or not. Over the next couple weeks I'll shoot the Sony along side my Arca and take real world samples all the way through post. That'll be interesting.
-
these numbers are ADC output levels after the different cameras have likely applied levels of amplification to the signal
Yes, but the level of amplification (gain) does not change the linear proportion between input count and linear output number. Say, well is 32,000 electrons, gain is 8, maximum output count is 4,000 (12 bits). We collected half well, 1 EV less than full, 16,000; so it is 2,000 output counts, once again 1 EV less than full. We can just consider the normalized count, dividing by 32,000 at input, or divided by 4,000 at output. Normalized counts at input and at output are identical (less noise introduced by the amplifier and ADC).
The ADC and amplifier are optimized so that "full well" closely corresponds to the maximum output. Photon count is proportional to electron count through quantum efficiency. To the effect, we have a very linear system; and where it is not linear (deep shadows and extreme highlights) we want to clip to avoid poor colour, blotches, other artifacts.
Flare reduces the acceptably linear portion of the characteristic curve to 11 stops with absolute best prime lenses. Adaptors, filters, any tiny amount of dust reduce it further.
-
Thanks, but I still do not fully understand. As far as I now, these numbers are ADC output levels after the different cameras have likely applied different levels of amplification to the signal and so different conversion factors, whether measured in "photons counted per raw level" or the ratio between "photon count as a fraction of full well capacity" and "output level as a fraction of maximum level".
So how do these differently amplified output values tell us anything about the relative amount of light received by the sensors? We seem to need information relating back to photo-electron counts and well capacities in order to compare the amount of light received, either in absolute terms (photons counted) or in relative terms (fraction of full well capacity). Maybe that information is encoded in the ratio of DxO's sensitivity measurement to the ISO speed setting used.
Thinking about lens flare is very interesting though: having seen some measurements of lens flare with real world images which suggested that it is very hard to keep the darkest parts of the recorded image as far as 12 stops below the brightest, I am skeptical about the practical relevance of some extremes of DR measured in unnatural lab situations. For example, a DR test target with very bright areas adjacent to very dark ones might give different results that the usual graduated strip with the brightest and darkest regions at opposite ends.
12 bits non-flared may be the max, but that's 12 bits per channel, going down from the exposure max which is never set perfectly.
Add a couple of stops to deal with mismatched light balances, and another couple of stops for exposure headroom and you can see that 14 bits lab DR is the least you really want -you cannot expose perfectly so the top is not at the top, and anyway there's a non-linear break near the top- in the case of a digital back you also have digital ISO setting.
I disagree with the fact that we have enough practical DR; in fact we have BARELY ENOUGH for realistic non-lab use. Of course people who work in studios with lightmeters and controlled lighting won't run out of DR as quickly as those in the field who deal with unmatched lighting, ISO set a bit higher and underexposed images who will be in purgatory quicker than they can shout "Beatrice!" .
If you wish we can try and do the numbers exactly.
Edmund
-
The ADC and amplifier are optimized so that "full well" closely corresponds to the maximum output.
How do we know that? It is certainly not true at ISO speed settings higher than base-ISO speed in cameras that apply increased analog gain in that case, and so can convert a signal far less that full well charge to maximum ADC level. It probably should be true with those CCD MF backs that apply a fixed analog gain and leave output adjustments for exposure at higher EI to be done in conversion from raw. But with the Sony A7R offering an ISO speed setting lower than the one used in this comparison, and with its base ISO speed apparently less than 100, I would not be so sure in that case.
Flare reduces the acceptably linear portion of the characteristic curve to 11 stops with absolute best prime lenses. Adaptors, filters, any tiny amount of dust reduce it further.
Thanks, that fits with my recollection. So when and how is 13 stops of sensor DR better than 12? This question is to everyone who enjoys discussing DR, not Iliah in particular.
-
Hi Edmund,
Problem, plainly, there is no way to accurately compute quality DR from lab DR. With current cameras it looks like -5 EV, more or less; empirically.
-
Hi Edmund,
Problem, plainly, there is no way to accurately compute quality DR from lab DR. With current cameras it looks like -5 EV, more or less; empirically.
Indeed: that is one reason that I am much more interested in measuring the range from maximum down to a point where the local SNR is something like 10:1 or 5:1 Anything much darker than that is probably only of interest to astronomy and surveillance, because lifting it enough to display as even dark shadows rather than pure black is going to look ugly.
Photon shot noise alone means that 5:1 SNR requires at least a 25 photon count, about three or four stops above the engineering DR floor of a good modern sensor with a dark/read noise level of about one or two electrons.
For example, with a well capacity of 50,000, that 5:1 floor is 11 stops down, and the traditional 10:1 guideline for "barely acceptable" requires a 100 count, so nine stops down from full well.
-
How do we know that?
Because it is what designers state, also on the international conferences where the sensors and digital imaging are main topics. Because Chipworks made available some independent studies, though those are expensive. Finally, because this is how the system optimizations are performed, and not only for sensors.
It is certainly not true at ISO speed settings higher than base-ISO speed in cameras that apply increased analog gain in that case, and so can convert a signal far less that full well charge to maximum ADC level.
Yes, I was referring to "base ISO" mostly. However it is worth mentioning that the industry sometimes puts different meanings to familiar terms. Here is what happens with the "well" term. For each gain full well is defined accordingly. The most practically used definition of full well is the amount that can be converted in linear fashion. At "base" ISO it is usually a bit (literally) smaller than the holding capacity of well.
with the Sony A7R offering an ISO speed setting lower than the one used in this comparison, and with its base ISO speed apparently less than 100
You can do direct experimentation to see what actually the lowest "honest" ISO setting on each camera. Just expose a relatively uniformly lit featureless surface strongly out of focus according to spot meter at different ISO settings (including "intermediate", like 130 and 160) and examine raw histogram of a portion close to sensor centre.
-
Any SNR on raw data can't be related directly to converted image, because demosaicking and colour conversion add significant noise and artifacts. 10:1 easily ages to 3:1 in the processed image, shadows are damaged much easier than midtones.
-
Yes, I was referring to "base ISO" mostly.
Agreed about base ISO speed, but that is my point: neither of these cameras is at base ISO speed in these examples, and they are above it by substantially different amounts: the ratio of their base ISO sensitivities seems to be about 3:1 (73:29), or about 1.3 stops. With the IQ180 probably amplified for that base ISO speed of 29, while the A7R is maybe amplified as if for ISO speed 73, if I am interpreting the DxO measurements correctly.
On 10:1 SNR degrading to 3:1 due to demosaicing and such: are you suggesting that the practical lower limit on the ratio of signal to shot noise might be even higher than 10:1 due to those effects, so needing more than a 100 photon count for pixel to be "photographically useful"? That makes some sense, but I would have to think about the benefits in the case where one then down-samples, improving the "per pixel" SNR, or with print dithering, which has a similar visual effect.
P. S. Chris: sorry if this was not the discussion you were expecting; no good deed goes unpunished in this forum! Thanks again for the samples.
-
Hi,
I have found very few images with more than 9 steps of DR. When I got my Alpha 99 I wanted to see if I could find a difference in DR to my Alpha 900. It took me something like three months, duping a high contrast Velvia slide in a "totally" dark room.
I would say that lens flare and other flare are a limiting factor.
On the other hand, I would suggest that we are quite tolerant on noise in the shadows. The shadows are compressed and hard see. The spectrum of the noise also matters. I had seldom issues with noise on my Sony's but the P45+ seems to be a more bit challenged. On the Sonys I can pull the shadows with confidence, on the P45+ the confidence is not there.
Best regards
Erik
Yes, but the level of amplification (gain) does not change the linear proportion between input count and linear output number. Say, well is 32,000 electrons, gain is 8, maximum output count is 4,000 (12 bits). We collected half well, 1 EV less than full, 16,000; so it is 2,000 output counts, once again 1 EV less than full. We can just consider the normalized count, dividing by 32,000 at input, or divided by 4,000 at output. Normalized counts at input and at output are identical (less noise introduced by the amplifier and ADC).
The ADC and amplifier are optimized so that "full well" closely corresponds to the maximum output. Photon count is proportional to electron count through quantum efficiency. To the effect, we have a very linear system; and where it is not linear (deep shadows and extreme highlights) we want to clip to avoid poor colour, blotches, other artifacts.
Flare reduces the acceptably linear portion of the characteristic curve to 11 stops with absolute best prime lenses. Adaptors, filters, any tiny amount of dust reduce it further.
-
P. S. Chris: sorry if this was not the discussion you were expecting; no good deed goes unpunished in this forum! Thanks again for the samples.
+1
The geeks in this forum have descended on those files like locusts ;)
Edmund
-
Thanks for these test samples. That is very impressive from the A7r. I shoot a lot of interiors using TSE lenses with the 5DIII tethered to C-1 but with DXO Optics for for my conversions. (I think I can control shadow fill light and look better in DXO than with any other program I have tried.)
My clients and I are happy with the Canon and I like my workflow. But I will serious consider going to the A7r for use with my Canon lenses once there is DXO raw conversion support for it. Hopefully C1 tethering will be supported soon too. Until then, I think it might be too much hassle for me to use it even if it is ca[able of slightly better results.
-
Comparing pure channels tell us more than pure numbers :)
-
Comparing pure channels tell us more than pure numbers :)
But you are not looking at pure channels as it seems, what you are looking at seems as processed channels?
-
Hi,
One of those geeks happen to develop the possibly best software for raw conversion, have some respect…
That aside, I think Chris has sound approach. Shoot to find out what works best in which situation.
Best regards
Erik
+1
The geeks in this forum have descended on those files like locusts ;)
Edmund
-
Hi Erik,
I guess Edmund is joking.
That aside ;) - yes, shoot and see; but exposure is a huge issue.
-
Hi Erik,
I guess Edmund is joking.
That aside ;) - yes, shoot and see; but exposure is a huge issue.
As always, Iliah, I agree with you :)
I used to think exposure was a solved problem because of Raw and huge DR, and color was "the" issue, and am now discovering that exposure is still critical.
Edmund
-
Well I think the thread is turning more an engineering discussion rather than photographic.
At the end, what we see in these examples is the fact that for a given job, with some cameras, you will have to work harder in order to get highlights and shadows to work. Which means more files with different exposures to blend the desired image in photoshop.
That's my thought.
Two years ago, I went through the process of giving up on my Canon 5DMKII for the D800 which gave me the ability to obtain cleaner files, with more DR and cleaner shadows. All this reflected on me working less and my clients being happier.
ACH
-
So is there a consensus that because of flare more than 11 steps of practical DR is not possible with current camera designs (not sensors) ?
How much practical DR can we already get from the better camera/sensor combos in the absence of strong flare from outside the image field ?
Edmund
-
The A7r file converts very well in RT. The shadows are smooth. The window light and lamp light have smooth gradations. It looks decent with sharpening off, NR off.
Rather than attach a crop with large file size, here are the settings to try.
I didn't look at the other files. I can give a family a whole new kitchen for the price of the camera so they don't matter to non-pros.
-
Ach,
Thanks for putting things in perspective ;-)
Personally I find the theoretical part interesting, too, but you are right that cameras are just tools. I think the A7r is an interesting tool, specially for one lusting for DR and resolution. Personally I am not a potential buyer, right now.
Best regards
Erik
Well I think the thread is turning more an engineering discussion rather than photographic.
At the end, what we see in these examples is the fact that for a given job, with some cameras, you will have to work harder in order to get highlights and shadows to work. Which means more files with different exposures to blend the desired image in photoshop.
That's my thought.
Two years ago, I went through the process of giving up on my Canon 5DMKII for the D800 which gave me the ability to obtain cleaner files, with more DR and cleaner shadows. All this reflected on me working less and my clients being happier.
ACH
-
Dear Erik,
I'm not sure that exposure is a theoretical part of photography. Sorry for being so stubborn ;)
-
Perhaps the best test, for me, would be to proceed and retouch each to be the best possible image.
That's the only way any test of camera ability should be done.
-
Hi Bernard,
If D800 is in lossless compression mode, the gaps are nearly non-existent, only white balance pre-conditioning adds them, and those are very few.
The results with digital backs strongly depend on exposure, which includes the colour of light. For a relatively well-exposed regions the histogram is totally gapless:
Thanks Iliah,
So, are you saying that the raw files of the a7r, due to their lossy nature, contain significantly less information that those of the D800, but that the gap between the D800 and the latest backs is small?
Cheers,
Bernard
-
Hi Iliah,
I don't argue with an expert ;)
Regarding exposure I see your point. Personally I am a base ISO guy and I try to expose for highlights, getting maximum exposure without clipping. It seems that Chris find 100 ISO the lowest he wants to go and exposes to protect highlights. Life is often a compromise.
I am quite interested in this stuff, but I guess some of the discussion is a bit of topic for some of our friends.
I am thankful for your explanation of the Sony ARW format. I guess it could be that the Bionz is only twelve bit wide but they want to push 14 bits of data trough it. It sort of makes sense to me. What about the Alpha 99, that camera shows up with 14 bits in RawDigger, as far as I know? Do they use the same solution?
Best regards
Erik
Dear Erik,
I'm not sure that exposure is a theoretical part of photography. Sorry for being so stubborn ;)
-
Hi Bernard,
> are you saying that the raw files of the a7r, due to their lossy nature, contain significantly less information that those of the D800
Yes.
> but that the gap between the D800 and the latest backs is small?
No, I see better colour from the digital backs, more balance of the system towards high resolution, but I also see that DB manufactures may not be doing enough, explaining exposure, vibration, lenses, and filters.
-
Hi Erik,
Without going any further into Sony innards, I would say they are capable of creating very decent digital backs, even if 135 format, but to do so they need to drop lossy compression and design it as a back, not as a camera.
-
Iliah,
Thanks for all your answers to my persistent questions and quibbles; I think I understand the situation better now.
About the A7R being "only 13-bit even before compression": do you think that its ADCs themselves are only 13-bit, perhaps using 14-bit output format just because for whatever reason, an even number of bits is always used for output? And yet the D800 has a true 14-bit mode, probably from the same sensor? That reminds me of an earlier case, where it was speculated that a Nikon body using a Sony EXMOR sensor was doing multiple non-destructive ADCs at each pixel and averaging to get an extra bit or two in its low frame-rate mode, whereas a Sony body using the same sensor only offered 12-bit.
-
do you think that its ADCs themselves are only 13-bit, perhaps using 14-bit output format just because for whatever reason, an even number of bits is always used for output?
The ADCs used in those sensors are comparator-based, one can extract very high precision from those if the time to raise the ramp in fine steps is enough. A short description of the ramp-compare ADC is on http://en.wikipedia.org/wiki/Analog-to-digital_converter The design decision is "what is good enough", to keep acquisition time short to ensure the target frame rate, and to have good enough precision.
And yet the D800 has a true 14-bit mode, probably from the same sensor?
Sensor is very similar, but the supporting circuitry, even that on the sensor chip, is somewhat different, and differently timed and sequenced. Looking at the dump of the raw data one can see it - say, A7r and D800 have different optical blacks around them. One can do quite a lot by just programming internal sensor registers in a different manner.
That reminds me of an earlier case, where it was speculated that a Nikon body using a Sony EXMOR sensor was doing multiple non-destructive ADCs at each pixel and averaging to get an extra bit or two in its low frame-rate mode, whereas a Sony body using the same sensor only offered 12-bit.
It was the same different sawtooth programming, for slower frame rate and higher bitness the steps were finer.
-
The ADCs used in those sensors are comparator-based, one can extract very high precision from those if the time to raise the ramp in fine steps is enough. A short description of the ramp-compare ADC is on http://en.wikipedia.org/wiki/Analog-to-digital_converter The design decision is "what is good enough", to keep acquisition time short to ensure the target frame rate, and to have good enough precision.
Sensor is very similar, but the supporting circuitry, even that on the sensor chip, is somewhat different, and differently timed and sequenced. Looking at the dump of the raw data one can see it - say, A7r and D800 have different optical blacks around them. One can do quite a lot by just programming internal sensor registers in a different manner.
It was the same different sawtooth programming, for slower frame rate and higher bitness the steps were finer.
Arghh, this ADC was a standard electronics course design example; I must have spent hours calculating their parameters when young.
Edmund
-
The ADCs used in those sensors are comparator-based, one can extract very high precision from those if the time to raise the ramp in fine steps is enough.
...
It was the same different sawtooth programming, for slower frame rate and higher bitness the steps were finer.
Thanks! A much better and simpler answer than the speculation I had heard before.
-
Hi,
I guess that I was one of those suggesting the double readout idea, but since than I arrived at what Iliah suggests. Nice to have it confirmed.
Best regards
Erik
Thanks! A much better and simpler answer than the speculation I had heard before.
The ADCs used in those sensors are comparator-based, one can extract very high precision from those if the time to raise the ramp in fine steps is enough.
...
It was the same different sawtooth programming, for slower frame rate and higher bitness the steps were finer.
-
So do you think the compressed raw is hard-coded into the hardware, or can we expect a future firmware update to provide uncompressed raw (with some limitations for example on the read-out speed) ?
-
Sorry to add to the geeky part of the discussion.
Mentioning the loss of resolution due to the more or less logarithmic conversion of Sony sounds bad. But once it is realized that the dominant noise source at higher illumination level is shot noise, a noise source that grows as a function of the light level, it can be understood that the full linear resolution as provided by Nikon and many back makers, is of little practical use. The extra resolution in the highlights will be completely swamped by shot noise.
A linear analog to digital converter provides half its resolution to the highest full stop of its dynamic range. So in the case of a 14 bit ADC, the highest EV of the total range is divided in 8192 values. One EV down the resolution is 4096 etc, until the last bit, the one in the deepest shadows, where a single EV is only expressed with a single bit. Apart from the rather ridiculous resolution the highest EV gets, this amount of resolution is not present in the signal going into the ADC. Shot noise will reduce the SNR, at the top of the illumination range, to say 45dB, which is a ratio of 1:178. So most of the ADC resolution is used to express noise, which is of little practical value. Sony is no doubt aware of this, and reduces the resolution to a more useful range, reducing file size.
As long as shot noise is high enough to dominate the resolution of the Sony approach at each level of illumination, dither will ensure that all tonal gradations are smooth.
The Sony a900 had both the same compression algorithm and uncompressed RAW. I was never able to find any visible difference.
-
Perhaps my reply above is a bit abstract. Maybe better to show some examples to try and clarify what (I think) is happening.
To start with I made a completely out of focus shot of a evenly lit white subject. I did this for several exposure setting. One rather under exposed and one over exposed. I opened the resulting files (made with a Sony a99, same compression as a7(R)) in RawDigger. I selected a small part of the image in the center. I made a histogram of the selected area. The results are attached. In the underexposed example one can clearly see the lack of resolution, there are clear discrete values. This is due to the linear nature of the ADC. At these low values there is no compression yet and we simply see the resolution of the ADC. And there is not a single value, as one would expect, but a range of values. This variation is caused by noise. Maybe a part is caused by differences in the sensitivity of each individual pixel, but the bulk is most likely caused by a mix of noise sources.
The second attachment shows the result for a high light intensity. The maximum of 16384 is almost reached. RawDigger does not show all values in this histogram, the results are grouped in 'bins' of 1/96th EV. Effectively, RAW digger bins the result in a logarithmic fashion (not totally different from the way the Sony compression algorithm works). So all of the theoretical 8192 values in the highest EV, are reduced to 96 values. One would expect to see only a single value in the histogram. All pixels receive the same amount of light, so the values should be the same. Read noise no longer plays a role at this high level of illumination. But instead of a single value, there is a group of values. This is (mostly) caused by shot noise. One could do the same experiment for a D800, and the result should be comparable. In other words, the noise completely dominates the ADC resolution and there is really no need for the excess of resolution in the higher part of the EV range.
-
Perhaps my reply above is a bit abstract. Maybe better to show some examples to try and clarify what (I think) is happening.
Hans,
Thank you for a clear exposition.
Edmund
-
Reading this post and rising the question about formats, sensors and software, I remembered that a couple of years ago, I had the chance to try a H4D40 in an informal shooting. These are ASA100
I opened up Phocus and processed these files, as you will see these have very clean shadows although trying to recover highlights destroys the middle tones.
This is a Kodak sensor. Comments
(https://dl.dropboxusercontent.com/u/10120389/1.jpg)
-
Dear Hans,
Apart from lossy compression curve Sony are using local lossy compression, packing data into 7 bits with occasional 11 bit keys.
Have you tried Sony for, say, shooting jewelry? How smooth are specular highlights? Do they tend to disappear? Have you studied how sharpening affects the upper 3 stops (that includes skin), where Sony offers only about 1000 levels?
Sony compression scheme is quite good, effective, and well-known. The reason other companies are not using such schemes much is that introducing artifacts already at the stage of writing raw is not exactly a welcome innovation. Also, given the compression curve, ETTR has little sense:
(http://blog.lexa.ru/sites/blog.lexa.ru/files/images/nexc3-curve.png)
Given the choice between lossy raw and lossless raw, my preference is lossless as it takes postprocessing better and has more archive value.
-
Hi Iliah.
As for the the effect of the way Sony Raw makes groups of 16 pixels I have no idea if there are situations where that would have a visual effect. But as far as the compressed curve is concerned, I tried to explain that there are no 1000 levels in the top 3 EV. Not with Sony and not with Nikon. At least not in a single pixel. For the simple reason that shot noise limits the resolution to say 1:200 (for the highest EV of the dynamic range). Nothing you or any ADC can change to that, simply physics. And when looking at a larger group of pixels, the values will average out, resulting in more resolution, but this is also the case with Sony, this is simple noise dithering, a well understood principle.
But in the end the only thing of concern is what the image looks like. If I remember correctly you own (or is it owned?) an a900. With that body there was a choice. Do you have examples showing how cRaw degrades a particular image? Or perhaps it would be even possible for you, given your programming experience, to simply take a D800 file and treat the Raw file in the same way Sony does? Maybe the later is asking a bit much.
I really would like to see an example, a clear a/b comparison, showing how the Sony compression degrades a picture. This discussion has been going on and on (on many fora, over a number of years) and I have yet to see a clear example.
-
Dear Hans,
I'm not sure the idea of introducing artifacts without accounting for the propagation and amplification of those artifacts during the raw and colour conversion, and later during postprocessing is a good one. More than once I was to add noise to SONY upper midtones and highlights to avoid posterization. I do shoot with SONY cameras a lot, but I still prefer lossless.
And, if it was just shot noise - little to no problem. But it is not just that.
I will work on D800 example as you suggested, just very busy now.
-
Hi,
Just an observation…
Best regards
Erik
-
But I will serious consider going to the A7r for use with my Canon lenses once there is DXO raw conversion support for it.
Already seems to be supported (http://www.dxo.com/intl/photography/dxo-optics-pro/supported-equipment) for Sony, Zeiss, Tamron, Sigma lenses.
-
Hi,
Just an observation…
Best regards
Erik
Yep, IQ260 underexposed and at iso 100, not a good combo.
-
Wow! This is getting crazy. I am going to guess that Chris kind of took these as snap shots (with a tripod). No lighting or set up, quick post, if any. Lets wait until he throughs up some lit, staged and produced shots.
-
Yep, IQ260 underexposed and at iso 100, not a good combo.
Actually, I think maybe it is "underexposed" simply because it is at exposure index 100, but is a sensor of base ISO speed about 29, and the camera does not apply any additional amplification at the higher speed setting (from 200 up, there is additional amplification). However the comparison above is missing the f-stop for the IQ260 sample, so I am not sure.
-
The Sony and Canon are at F/11 2Sec ISO100
The IQ 260 file I used was at F/8 2/3 1Sec ISO100. There is also a 2Sec exposure in the DropBox folder I've linked. What's curious to me is that people keep saying the IQ file is underexposed and yet it has less hilight detail than the Sony file. That doesn't make any sense to me.
-
The Sony and Canon are at F/11 2Sec ISO100
The IQ 260 file I used was at F/8 2/3 1Sec ISO100. There is also a 2Sec exposure in the DropBox folder I've linked. What's curious to me is that people keep saying the IQ file is underexposed and yet it has less hilight detail than the Sony file. That doesn't make any sense to me.
Hi Chris,
It would make sense if the IQ file has less dynamic range than the Sony file, which it possibly does. However, Erik's observational conversion crop of the fireplace looks horrible, where my (un-posted) initial conversion tests look much better.
It may be related to the Raw converters and settings used. In addition, Capture One will boost the apparent exposure by almost a stop when a film curve response is used. I haven't had the time to do a better analysis of the Raw data yet, but I'll look into it.
Cheers,
Bart
-
Yes,
I am aware of it. What is interesting that the Sony has still good detail. On the other hand the Sony is also used at 100 ISO and Chris indicates that he uses 100 ISO normally. Chris also indicates that highlights were better on his 1s exposure than his 2s exposure. Exposure has two ends.
Anyway, I am thankful that he shares the experience. Hopefully he continues sharing his findings.
Best regards
Erik
Yep, IQ260 underexposed and at iso 100, not a good combo.
-
Here's what exposures looked like straight out of the back as a point of reference. I didn't feel that I could produce an acceptable image from the brighter exposures.
(http://christopherbarrett.net/forum_images/CameraTest_Exposures.jpg)
-
I still am not quite sure what happened after the first post.
To my unscientific eye, the Phase has amazing color separation and the Sony wipes the floor with the Canon.
Highlight recovery has always been an issue with C1 without affecting everything else, so I prefer to do localized edits to recover specific hot areas. It's pretty much the only thing where LR's editing tools have an edge over C1 in my experience.
-
Here's what exposures looked like straight out of the back as a point of reference. I didn't feel that I could produce an acceptable image from the brighter exposures.
(http://christopherbarrett.net/forum_images/CameraTest_Exposures.jpg)
Thx. I agree.
-
Hi,
I have downloaded the IQ260 and Sony A7r files, and looked in both Lightroom and Capture One.
Here is what I see:
- The IQ 260 image is sharper at the pixel level. Significantly sharper!
- The IQ 260 image has significant noise in my processing (worst in C1 and better in LR5)
- Better sharpness on IQ260 in C1 compared to Lightroom
- Highlights probably better on IQ260
- Darks very noisy on IQ260 while Sony keeps good detail
Another observation is that IQ260 needs LCC correction which lifts shadows quite a lot. Check the remote control in the lower right corner, quite noisy. The Sony doesn't need LCC with the Canon lens used.
It seems that Sony is working well at ISO 100.
Best regards
Erik
-
Already seems to be supported (http://www.dxo.com/intl/photography/dxo-optics-pro/supported-equipment) for Sony, Zeiss, Tamron, Sigma lenses.
Yeah I saw that shortly after posting. Now I need a new excuse. I may play with some of these raw files in DXO.
-
Hello,
I'm one of those shooters at the coal face every day taking pictures and my Nikon D800E is bringing home the bacon with clean shadows and great resolution. For me CCD chip's are relics from the last decade as film was to the last century
Cheers
Simon.
-
Hello,
I'm one of those shooters at the coal face every day taking pictures and my Nikon D800E is bringing home the bacon with clean shadows and great resolution. For me CCD chip's are relics from the last decade as film was to the last century
Cheers
Simon.
Different usage scenarios, different expectations.
I have that same camera (OK, without the E) and I prefer the relics from the past for skin tonality.
-
Different usage scenarios, different expectations.
I have that same camera (OK, without the E) and I prefer the relics from the past for skin tonality.
True skin is so passé, don't you think? Like any color consultant I have profiles which you just dump over a file and they paint in pink, beige or latte in the right places. :P
Hmm, come to think of it one of the well known posters here kind of does the same, but he won't admit it. Of course this does not detract from the quality of his setups.
Edmund
-
Hi Erik,
Here is what I came up with on the IQ260.... Rather clean.
-
True skin is so passé, don't you think?
Not sure if you are being serious here. Either way some digital cameras have a nasty (in my opinion) video look to them when it comes to skin tone. Some people don't mind or notice that look whilst others like synn and myself prefect how film renders skin. The problem I have with bad skin rendition is that it makes people look like mannequins/rubber dolls, not a look I like.
Not keen on heavily airbrushed skin either, that a different sort of ick. It's more uncanny valley than attractive to me.
Like any color consultant I have profiles which you just dump over a file and they paint in pink, beige or latte in the right places. :P
Hmm, come to think of it one of the well known posters here kind of does the same, but he won't admit it. Of course this does not detract from the quality of his setups.
Are you talking about fixing issues with highlight rendition here?
-
How can it be that Erik finds such awful noise in the IQ260 image - using Phase One's own softawre no less ("worst in C1") - yet Phil can obtain a significantly superior image?
-
How can it be that Erik finds such awful noise in the IQ260 image - using Phase One's own softawre no less ("worst in C1") - yet Phil can obtain a significantly superior image?
I'm sure he used luminance noise to compensate, sharpening as well.
ACH
-
Noise reduction was up from 50 to 58 (50 is "neutral") on the previous above image..
here it is at ZERO...
-
Hi,
My images are here: http://echophoto.dnsalias.net/ekr/Articles/CBTest/
The settings used are enclosed below.
Best regards
Erik
Noise reduction was up from 50 to 58 (50 is "neutral") on the previous above image..
here it is at ZERO...
-
Looks like you are pushing the image. Perhaps that is what you are testing for...?
but the settings I have a re close to "asshot" with minor tweaks.
(what a hyphen can do?!)
-
Hi,
I simply adjusted the images to taste. My taste. I have done it for all three images. WB was on white passepartout on one of the prints on the wall. After that I adjusted windows to get some detail, adjusted blacks so I get some blacks, pushed shadows to get shadow detail. That is what I do for all my images.
Best regards
Erik
Looks like you are pushing the image. Perhaps that is what you are testing for...?
but the settings I have a re close to "asshot" with minor tweaks.
(what a hyphen can do?!)
-
So perhaps the file is less flexible when compared. That can often be important when editing. I simply tried to bring the HL details and expose the Blacks to taste.
I like the blade effect on the Christmas lights. Oh and I also like Chris's taste, a Herman Miller fan and a nice movie selection...perhaps we can start a rental program for the flix? :-P
-
Well, comparing LR5 and Capture One it was the same file.
Also, I was probably one of the first Lightroom buyers. It is a program invented for me! I was in progress of writing something similar to Lightroom, so for me it was like revelation! I have used LR since Beta 3. I own Capture 1 but we don't make friends, so you can say I have bias for LR.
Best regards
Erik
So perhaps the file is less flexible when compared. That can often be important when editing. I simply tried to bring the HL details and expose the Blacks to taste.
I like the blade effect on the Christmas lights. Oh and I also like Chris's taste, a Herman Miller fan and a nice movie selection...perhaps we can start a rental program for the flix? :-P
-
Antonio, Phil, Erik,
Thank you for the clarification.
-
Erik,
Your settings introduce a blue magenta tone into the bookcase....
-
Hi,
I just used the passepartout in the image below for WB, that is all I did (both LR and C1). As I said, I am not a C1 user. I have had it for something like half a year, and process some images with it.
I would suggest it doesn't affect noise levels. I did not do LCC correction in LR5, because I did not get it working.
Best regards
Erik
Erik,
Your settings introduce a blue magenta tone into the bookcase....
-
Actually that blue color for the bookcase is correct.
-
Actually that blue color for the bookcase is correct.
thanks for that calrification. It was isolated, so it makes sense..
It hadnet caught my attention unil I saw Eriks download image. Should have cross checked :-)
-
Actually that blue color for the bookcase is correct.
Chirs,
Next time you do these samples would you please paint the bookcase in a dark brown? It would make color checking easier.
Also it would be greatly appreciated if you added some fill light to balance the windows, a little more stylized propping and for me one or two models so I can compare skintones.
Oh yea, since you were running your RED sound would also be appreciated to add to the ambience.
I look forward to this and thanks in advance.
BC
-
Chris,
Would be easier if you just sent me the cameras per UPS - I promise I will return them promptly in 2 months :)
On second thoughts, thank you for posting a test which I would have found hard to do myself ... even if you had sent me the hardware.
Edmund
Chirs,
Next time you do these samples would you please paint the bookcase in a dark brown? It would make color checking easier.
Also it would be greatly appreciated if you added some fill light to balance the windows, a little more stylized propping and for me one or two models so I can compare skintones.
Oh yea, since you were running your RED sound would also be appreciated to add to the ambience.
I look forward to this and thanks in advance.
BC
-
To put it simply, this "stupid camera test" with stupid raw files got me convinced on the a A7r! Amazon has it for less than $2300.
Stupid test, stupid cameras, stupid buyers....eh! Whadaya gonna do?! :-)
Thanks Chris.
-
To put it simply, this "stupid camera test" with stupid raw files got me convinced on the a A7r! Amazon has it for less than $2300.
Stupid test, stupid cameras, stupid buyers....eh! Whadaya gonna do?! :-)
Thanks Chris.
I just rented that stupid camera should be here in a few days. (A7R + Canon EF lens adapter) ;D ;D
-
That's just so stupid Ken! :-) hehehehe
I was going to be even more stupid and buy it for $2200 off Amazon
Who's renting it out?
-
That's just so stupid Ken! :-) hehehehe
I was going to be even more stupid and buy it for $2200 off Amazon
Who's renting it out?
lensrentals.com
-
Oh and I also like Chris's taste, a Herman Miller fan and a nice movie selection...perhaps we can start a rental program for the flix? :-P
Gotta credit Charles & Ray Eames with the chair design too! And Chris has the recent Eames doc on DVD, so the rental program will definitely be of some value. ;D
-Dave-
-
Absolutely. Although I have seen it (Netflix has great docs), a repeat wouldn't be bad. Of course, after others. I'm not in a rush ;-)
-
And yet another stupid comparison … but sometimes it's just fun.
It doesn't deserve an own thread so I thought I'll just add it here. I hope Chris don't mind.
P45 on Contax 645, Zeiss CZ Planar 2.0/80 @ f8, ISO50, 1/60'', 39MP
versus
Sony A7R, Zeiss ZM Planar 2.0/50 @ f8, ISO50, 1/60'', 36MP
Actually I've tested some things with my 55mm lens on the Contax in the backyard. Since the tripod was in place I thought why not quickly compare the Contax/P45 and the A7R?!
This is why the position of the tripod remained the same. However, coincidentally the lenses cover almost the same height of the scene. Horizontally A7R shows a wider part of the scene (naturally). So the magnification of the scene is somewhat larger on the P45 and this is why the comparison is a bit in favour for the P45.
Also, the ZM Planar 50 is a bit sharper at f5.6 on the A7R while the Contax 2.0/80 shows about the same crispness at f5.6 and f8 on the P45.
Too, there are certainly better lenses for the A7R - but on a tech cam with digital large format lenses also the P45 would look better. So lens wise I think the test is pretty fair.
Then again: who cares :-) This isn't for science ...
Sorry - the scene is totally boring! (But it's great to check the registration of my split screen… so…)
Enough said… here it is (P45 on the left, A7R on the right in all screenshots).
Have fun!
Scene:
(http://www.tw.p1.spacequadrat.de/files/01_P45_A7R_cap.jpg)
100% crop near center:
(http://www.tw.p1.spacequadrat.de/files/02_P45_A7R_center.jpg)
100% crop from the right edge:
(http://www.tw.p1.spacequadrat.de/files/03_P45_A7R_edge.jpg)
Shadow detail … these crops come from the corner on the bottom right.
I've pushed exposure on both shots by 4 stops in C1.
100% crop, +4EV, without NR applied:
(http://www.tw.p1.spacequadrat.de/files/04a_P45_A7R_4EV.jpg)
100% crop, +4EV, with NR applied on the P45 (in addition some slight color adjustment):
(http://www.tw.p1.spacequadrat.de/files/04b_P45_A7R_4EV_P45NR.jpg)
I am really amazed how well the little A7R performs with an adapted ZM lens (not particularly in this test... but in general). And the camera will certainly have a place in my kit (mainly for stitching, though).
I am also amazed how well the 8 years old sensor of the P45 compares to the latest and greatest 35mm sensor in this comparison (however, in low light, especially tungsten, things will look worse).
-
Very impressive performance from the Sony/Zeiss combo, tho_mas. Aren't you in "shutter shock" territory with a 1/60th sec. Tv? Sure don't see it.
IMO the A7r is so close. Put a quieter shutter in there and (perhaps) do a better job of minimizing vibration, and Bob's yer Uncle.
-Dave-
-
Tho_mas
Thanks for sharing that. Like you wrote, it's interesting how well the p45 holds up to a new design but also how good the Sony is. How many generations old is the p45 now? P45, P45+, IQ1, IQ 2?
One observation, you can see the differences in the DOF between format sizes by looking at the branch in the foreground all the way left on bottom. It's much more OOF in the MF than in the 135. Did you focus on the building or at infinity?
-
Very impressive performance from the Sony/Zeiss combo, tho_mas. Aren't you in "shutter shock" territory with a 1/60th sec. Tv? Sure don't see it.
IMO the A7r is so close. Put a quieter shutter in there and (perhaps) do a better job of minimizing vibration, and Bob's yer Uncle.
I've been following the blog of one of our fellow members, Jim Kasson, for some time . He's been researching and writing on 'vibration and resolution issues, particularly regarding the D800 and now the Sony A7r. Jim appears to combine an extremely high level of technical expertise with a healthy dose of pragmatism . a welcome breath of fresh air.
His Jan-1 posting included these words:
<<
It’s a challenging business. So challenging, in fact, that we should think about whether or not the changes we can see in this very controlled testing situation aren’t unimportant in real field use. When all is said and done, the effects of the shutter shock are on the same order as the focusing errors under nearly ideal conditions.
>>
Recommended.
http://blog.kasson.com/?m=20140101
--
ps
Tho_mas - thanks for posting this !
-
Hi,
Yes, Jim has a lot of good stuff!
Best regards
Erik
I've been following the blog of one of our fellow members, Jim Kasson, for some time . He's been researching and writing on 'vibration and resolution issues, particularly regarding the D800 and now the Sony A7r. Jim appears to combine an extremely high level of technical expertise with a healthy dose of pragmatism . a welcome breath of fresh air.
His Jan-1 posting included these words:
<<
It’s a challenging business. So challenging, in fact, that we should think about whether or not the changes we can see in this very controlled testing situation aren’t unimportant in real field use. When all is said and done, the effects of the shutter shock are on the same order as the focusing errors under nearly ideal conditions.
>>
Recommended.
http://blog.kasson.com/?m=20140101
--
ps
Tho_mas - thanks for posting this !
-
http://blog.kasson.com/?m=20140101
Yeah, I've been following Jim's blog but hadn't read that last one. Hmmm...given that I take photos mainly handheld (aside from the Pentax, which for me is a specialized tool and does lend itself to tripod use) and have no intention of changing that aspect of my approach due to the capabilities of any particular sensor, I think I can rule out "shutter shock" as a real-world issue with the A7r.
Another point: I have my eye on the evolving lineup of 4k displays. When I find one that meets my price/performance sweetspot I'll go for it. I'm not that interested in such a display for editing but rather presentation. It would replace my printer (and prints) for all but a small number of extra-wide panos. I can see switching an A7r into 16:9 format and leaving it there. The Sony offers enough resolution that I could do a 4:1 downsample of those photos and still nearly fill a 4k frame. Pretty cool!
Time to ruminate some more. ;)
-Dave-
-
Hi,
Myself shoot mostly on tripod. Not so much for stability but for composition and relaxed work. So I use tripod whenever practical.
What I would comment is more on the 4K displays. I do some printing, but I am essentially limited by wall space. What I do often with my pictures is making slide shows. Here I can also mix in motion. I find it a great way of presenting images. The only thing is that present day HD (1080P) doesn't have the quality I feel I need. So I really looking forward to 4K projectors coming down in price.
Best regards
Erik
Yeah, I've been following Jim's blog but hadn't read that last one. Hmmm...given that I take photos mainly handheld (aside from the Pentax, which for me is a specialized tool and does lend itself to tripod use) and have no intention of changing that aspect of my approach due to the capabilities of any particular sensor, I think I can rule out "shutter shock" as a real-world issue with the A7r.
Another point: I have my eye on the evolving lineup of 4k displays. When I find one that meets my price/performance sweetspot I'll go for it. I'm not that interested in such a display for editing but rather presentation. It would replace my printer (and prints) for all but a small number of extra-wide panos. I can see switching an A7r into 16:9 format and leaving it there. The Sony offers enough resolution that I could do a 4:1 downsample of those photos and still nearly fill a 4k frame. Pretty cool!
Time to ruminate some more. ;)
-Dave-
-
Hi,
I would suggest that the effect you see relates to longer focal length of Planar 80 compared to the Zeiss lens. The Planar would need stopping down one stop more to give equivalent DoF.
Best regards
Erik
Tho_mas
Thanks for sharing that. Like you wrote, it's interesting how well the p45 holds up to a new design but also how good the Sony is. How many generations old is the p45 now? P45, P45+, IQ1, IQ 2?
One observation, you can see the differences in the DOF between format sizes by looking at the branch in the foreground all the way left on bottom. It's much more OOF in the MF than in the 135. Did you focus on the building or at infinity?
-
Aren't you in "shutter shock" territory with a 1/60th sec. Tv? Sure don't see it.
With my kit I am not in shutter shock territory at all.
I've tested shutter shock with the ZM 2.0/50 carefully.
Kit:
For the time being I've mounted a Novoflex QPL1 (without rubber!) on the A7R.
Head is an Arca Swiss D4 (geared version).
Tripods tested are 2 wood tripods (large with single extension and center column, mid sized with 2 extensions and a flat mounting plate) and a small FLM carbon with 3 extensions and center column.
I could not detect any sign of shutter shock at any shutter speed on one of these tripods shooting in landscape or portrait orientation at close or far distance (everything carefully tightened and locked down).
I do see shutter shock with an old Linhof ballhead when all connections are damped with rubber & cork.
Me personally I take from this is that shutter shock is not a problem at all when you use decent plates, tripod head and trripod. At least not in conjunction with the small, light ZM Planar 50mm ( right now I only have this lens for the A7R).
Hope this helps....
-
Did you focus on the building or at infinity?
I've focused on the roof exactly in the center of the image. It's around 50 meters distance. So for the 50mm lens its infinity (focused using live view) and for the 80mm lens it's short before infinity (focused on custom shimmed split screen).
-
What I would comment is more on the 4K displays. I do some printing, but I am essentially limited by wall space. What I do often with my pictures is making side shows. Here I can also mix in motion. I find it a great way of presenting images. The only thing is that present day HD (1080P) doesn't have the quality I feel I need. So I really looking forward to 4K projectors coming down in price.
IMO that covers it nicely. Besides zooming & panning with stills I've also been shooting some video. I enjoy it! People are just used to looking at screens. I'm used to looking at screens. :) Might as well embrace that and go with it. A large HD screen can look coarse with still images, at least when you have a nice print to compare it to, but multiply the pixels 4x and I think we'll be in good shape.
-Dave-
-
With my kit I am not in shutter shock territory at all.
I've tested shutter shock with the ZM 2.0/50 carefully.
Me personally I take from this is that shutter shock is not a problem at all when you use decent plates, tripod head and tripod. At least not in conjunction with the small, light ZM Planar 50mm (right now I only have this lens for the A7R).
Hope this helps....
Yes, thanks. I own the ZM 50mm Planar too...as good as it gets without moving into exotic design territory à la the Leica 50/2 ASPH or the 55mm Otus. I've got the sturdy tripod gear (made for large format use) should I need/want it. A concern about the A7r I can, I suspect, dismiss.
-Dave-
-
Hi,
I don't know. Lloyd initially discovered the problem on the Leica M. The reduction in sharpness is subtle. It obviously depends on factors. The effect is probably much less than mirror induced vibration on a DSLR. It may also be that once you notice the problem it is much more obvious than before.
My recent cameras have either electronic first curtain or central shutter, so I don't have observed this problem.
Quite obviously, Sony is using an old sensor for the A7r, probably the one used in the D800/D800E but without the OLP filter. So the old sensor doesn't have new features like electronic first curtain and on sensor phase detection AF. In all probability a new camera is in the works, with a sensor using smaller pixels and new features.
I am in favour of reducing pixel size within reasonable bounds. It may be diminishing returns, but I am pretty sure that smaller pixels will sharpen better, interpolate better and have less aliasing.
Best regards
Erik
Yes, thanks. I own the ZM 50mm Planar too...as good as it gets without moving into exotic design territory à la the Leica 50/2 ASPH or the 55mm Otus. I've got the sturdy tripod gear (made for large format use) should I need/want it. A concern about the A7r I can, I suspect, dismiss.
-Dave-
-
The Sony ARW2 format is non-linearly compressed, like gamma compression in jpegs and tiffs. Current dcraw release has a bug which causes only 12 bit to be extracted, it's easily patched to extract the full 14 bits though. I think rawdigger can use both dcraw and rawspeed decoding, I think rawspeed decoding should be correct.
I am somewhat surprised that A7r exif says 14 bit but RawDigger shows 12 bit data (going up to 1:4096) .
If we assume true 14 bit we would have 16384 values, but the significance of the LSBs (least significant bits) at high numbers would be very low. If we assume that FWC corresponds to 16384 the standard of deviation would be 128. It is quite obvious the last few bits only represent noise. On the low end more bits are used but the last two bits (or so) are still fairly stochastic, are they not?
-
Hi Anders,
I found out, Raw digger calls it the ARW-fix.
Best regards
Erik
The Sony ARW2 format is non-linearly compressed, like gamma compression in jpegs and tiffs. Current dcraw release has a bug which causes only 12 bit to be extracted, it's easily patched to extract the full 14 bits though. I think rawdigger can use both dcraw and rawspeed decoding, I think rawspeed decoding should be correct.
-
I am in favour of reducing pixel size within reasonable bounds. It may be diminishing returns, but I am pretty sure that smaller pixels will sharpen better, interpolate better and have less aliasing.
Best regards
Erik
Erik, I'm getting confused now ::)
I read a lot recently and not only on this forum about how good the "old" backs were, no, still are, with their "fat pixels".
Do we want/need/prefer smaller pixels or larger pixels, or are the trade offs such that perfection in either camp means we accept the compromises of each approach?
-
I think the appreciation of the "fat pixel" backs comes from several things.
Most (not all) of the fat pixel backs have kodak sensors which had a special saturated look, a bit like color reversal film, a look which many fall in love with. This look could be recreated with smaller pixels too, but later chips (dalsa) has chosen another more neautral approach to color which many prefer too.
I think there's also a bit of disappointment from users that most improvements has been in increased resolution, while many actually would like to see something like a 40 megapixel 6x6 sensor rather than a 80 megapixel 645 which we now have. The increased resolution has put higher requirements and cost on optics to resolve that (and feed the sensor with close-to-perpedicular light), and many would prefer optics that concentrate more on look and less on resolving power.
Life was easier in the fat pixel days, almost any lens worked well and delivered razor-sharp images in pixel peeping which is a special kind of satisfaction :-)
Concerning noise and full-well capacity etc I don't think those older fat pixel where any better in any pratical way, a modern small pixel is better at delivering a clean signal than an old fat pixel.
I read a lot recently and not only on this forum about how good the "old" backs were, no, still are, with their "fat pixels".
Do we want/need/prefer smaller pixels or larger pixels, or are the trade offs such that perfection in either camp means we accept the compromises of each approach?
-
Hi,
Sorry for not responding, I had problem with my internet connection.
One of the problems with fat pixels is that lens resolves more information than the sensor. This will lead to sensor generating spurious resolution or as I would call it "fake detail". The most well know is color moiré, but there is also monochrome moiré, color aliasing and jaggies.
I recently shot an aperture series with my P45+ and a Planar 80/2.8 lens at apertures at f/5.6 to f/22. All images up to f/11 showed a lot of artefacts, which suddenly went away at f/16. See enclosed pictures. The P45 has 6.8 my pixels.
I also the same subject with a 16-80/3.5-4.5 zoom lens on my Sony Alpha 77 at 80 mm and same distance, using f/8. The Sony Alpha 77 has 24 MP at 3.9 my, and with a full frame 645 back that would correspond to 143 MP. The Alpha 77 may be OLP filtered or not, I don't actually know.
Reducing pixel size reduces DR. Halving the pixel diameter would reduce DR by one EV, if all other factors were constant. Midtone and highlight noise level would be very little affected as it is mostly dependent on the number of incident photons captured and that is mostly a function of the sensor area.
Halving pixel size would not quadruple the information content, however, as MTF would be reduced at smaller pixel sizes.
The third enclosed image shows that a 3.9 my sensor would get much better detail.
Best regards
Erik
Erik, I'm getting confused now ::)
I read a lot recently and not only on this forum about how good the "old" backs were, no, still are, with their "fat pixels".
Do we want/need/prefer smaller pixels or larger pixels, or are the trade offs such that perfection in either camp means we accept the compromises of each approach?
-
Thank you both.
Torger, was/is the colour look a result of the filtration used? Surely only from the "profile" used to de mosaic the data? If so the same "look" could come from the Dalsa series?
The question is was that "look" which is attributed to the fat pixels not a result of pixel size but of the software?
Erik,
Indeed the detail "captured" will improve as the pixel size decreases, but has the dynamic range not been maintained or indeed improved from the fat pixel backs as the "read" technology and associated sensitivity and amplification have improved over time?
Sorry if this has strayed from the thread title but as a fat pixel owner, venerable P20, I always find the latest backs/cameras interesting to see if the "improvements" are real at the sizes I use as an amateur. Your patience is appreciated.
-
Hi,
The short version:
Sensors have gotten better. Making pixels smaller reduces the DR per pixel but the increased number of pixels helps a lot.
The long version, see below:
DR is a bit tricky. The technical definition is FWC/readout noise, with WFC is being Full Well Capacity, the number of electrons a pixel can hold. Readout noise is the noise in reading out the electrons.
For medium size pixels like (6 my) the FWC may be around 60000 electrons FWC is by and large proportional to pixel surface, even if I think some progress have been made. The major factor in DR has probably been the reduction in readout noise. Modern CMOS may have say 3 electron charges in readout noise while CCD is more like 12-15 electron charges.
Let's assume we have a P45+ with 6.8my pixels, and let it compare to a 3.4 my pixel sensor using modern CMOS-technology. Let's assume 65000 electron charges (EC) per pixel for the P45+ and 16 EC in readout noise. So we have 65000/16=4062, converting to EV we get 11.99 EV. DxO-Mark measures 11.75 EV at 50 ISO (in 'screen, mode).
Now, let's take a CMOS sensor with half of that diameter, and assume FWC 65000/4 = 16250, if we assume 4EC in readout noise we would get a DR of 4062 that is 11.98, nearly the same as the "fat pixel CCD".
Now the small pixel CMOS has four pixels instead of one. Would be print both at the same size and assuming a sensor of the same size, the combined FWC would be 65000, read noise adds in quadrature, so we would have sqrt (4 * 4 * 4) = 8
So DR would be 12.98 EV, the smaller CMOS pixels would have a 1 EV advantage over the large pixel CCD. Would we do the same math with a CCD instead, FWC would still be at 65000 but readout noise would be 32. So DR would reduce to 10.98 EV.
Newer generations of CCD sensors have less readout noise than the older ones, but they are still in the low tens. Regarding CMOS, some sensors have external analog-digital converters (ADC), the off chip ADCs tend to be noisy, so these cameras have CCD like readout noise. Many modern CMOS sensors have thousands of on chip converters that work in parallell, these are much less noisy.
What complicates the issues is that DR I describe is a technical definition, and represents a level of noise where the signal is barely perceptible. In photographs we want a better Signal Noise ratio (SNR), on the other hand lower end of the DR scale is normally what I would call deep shadows where we would have little detail anyway.
Midtones and highlights are more affected by "shot noise", the natural variation of incident photons, which is independent of the sensor technology and depends only on exposure, sensor surface and quantum efficiency QE). QE is the percentage of the incoming photons that are detected.
Hope this helps! Here is a recommended paper, describing it more deeply: http://theory.uchicago.edu/~ejm/pix/20d/tests/noise/index.html
Best regards
Erik
Thank you both.
Torger, was/is the colour look a result of the filtration used? Surely only from the "profile" used to de mosaic the data? If so the same "look" could come from the Dalsa series?
The question is was that "look" which is attributed to the fat pixels not a result of pixel size but of the software?
Erik,
Indeed the detail "captured" will improve as the pixel size decreases, but has the dynamic range not been maintained or indeed improved from the fat pixel backs as the "read" technology and associated sensitivity and amplification have improved over time?
Sorry if this has strayed from the thread title but as a fat pixel owner, venerable P20, I always find the latest backs/cameras interesting to see if the "improvements" are real at the sizes I use as an amateur. Your patience is appreciated.
-
Erik, I like the long version. ;D
Large v small pixels is a simplification, well explained. It was, when laid out like this, amazing they did so well with CCD, but the advances in chip design and production have enabled so much more "efficiency" in signal handling. The other problem, dragging this back OT, in comparisons is the "processing" done before the RAW. This Sony in particular uses a number of "tricks" on the output as our erudite analysers have shown but it is a game changer, more than the D3 then D800 were, they were landmarks perhaps to stretch analogies.
Not mentioned because, although I know we have a number of video/mixed media users, we tend to discuss the still potential is the video which I see, from Philip Bloom, could, with a bit of tweaking from Sony be outstanding rather than pretty good.
Thanks for the link, i have seen it before. I start really well then it sort of slides away from me for a while then suddenly I'm back with some understanding. I visited the Ilford factory in the UK earlier this year and that was so many worlds away, a different planet.
-
So much tweaking, so little ground truth ...
Edmund
Erik, I like the long version. ;D
Large v small pixels is a simplification, well explained. It was, when laid out like this, amazing they did so well with CCD, but the advances in chip design and production have enabled so much more "efficiency" in signal handling. The other problem, dragging this back OT, in comparisons is the "processing" done before the RAW. This Sony in particular uses a number of "tricks" on the output as our erudite analysers have shown but it is a game changer, more than the D3 then D800 were, they were landmarks perhaps to stretch analogies.
Not mentioned because, although I know we have a number of video/mixed media users, we tend to discuss the still potential is the video which I see, from Philip Bloom, could, with a bit of tweaking from Sony be outstanding rather than pretty good.
Thanks for the link, i have seen it before. I start really well then it sort of slides away from me for a while then suddenly I'm back with some understanding. I visited the Ilford factory in the UK earlier this year and that was so many worlds away, a different planet.
-
So much tweaking, so little ground truth ...
Edmund
I suppose if Sony made a full, frank and open disclosure all the reverse engineering Depts. in Nikon, Panasonic, Canon etc etc would be redundant, we don't want that do we?
-
Torger, was/is the colour look a result of the filtration used? Surely only from the "profile" used to de mosaic the data? If so the same "look" could come from the Dalsa series?
The question is was that "look" which is attributed to the fat pixels not a result of pixel size but of the software?
I need to speculate a bit, because I don't know for sure. But it seems likely that it is a combination of the design of color filters on the sensor itself and the profiles used for the cameras in the raw converter.
The Dalsa series don't have the same color filters on the sensor so the exact same look cannot be reproduced, but you can probably come quite close.
I'm almost 100% sure that the look has nothing to do with the pixel size though, ie you could design a new sensor with smaller pixel size but with the same color filters and profile in the software and get the same look. New sensors just don't use the same color filters as they did back then.
One theory often mentioned is that the fat pixel has larger full well capacity and thus captures more photons which would somehow affect the look in a positive way. However one need to look at the total amount of captured photons and also read noise, and then modern sensors are substantially better, so I don't think that theory holds up.
-
More orthogonal filters, sensor considerably outresolved by the lens, - you can see how well these systems perform by the incessant complaints about Moiré :)
Edmund
I need to speculate a bit, because I don't know for sure. But it seems likely that it is a combination of the design of color filters on the sensor itself and the profiles used for the cameras in the raw converter.
The Dalsa series don't have the same color filters on the sensor so the exact same look cannot be reproduced, but you can probably come quite close.
I'm almost 100% sure that the look has nothing to do with the pixel size though, ie you could design a new sensor with smaller pixel size but with the same color filters and profile in the software and get the same look. New sensors just don't use the same color filters as they did back then.
One theory often mentioned is that the fat pixel has larger full well capacity and thus captures more photons which would somehow affect the look in a positive way. However one need to look at the total amount of captured photons and also read noise, and then modern sensors are substantially better, so I don't think that theory holds up.
-
Hi,
Perception may play a role. Also when looking at an image at actual pixels, the larger pixels will mostly have an advantage. But quantity has a quality of it's own.
Best regards
Erik
More orthogonal filters, sensor considerably outresolved by the lens, - you can see how well these systems perform by the incessant complaints about Moiré :)
Edmund
-
was a bit miffed with the raws Chris so kindly posted up as there were areas of colour that didn't look right to me especially the yellow by the fireplace where mixed lighting hasn't helped the smooth tones but they looked very posterised and gritty in adobe raw. tried the new capture one and i'm now interested in this camera again.
not a scientific test but i gave the shadows a boost and roughly tried to match them by adding some saturation in AR. C1 has done a much better job of handling this area.
-
Hi,
I took a good look at that are in both LR 5.3 and Capture 1. Regarding the LR 5.3 image I have seen a similar effect and in my case it came from masking feature in the sharpening group of keys. Increasing masking will reduce/eliminate the artefacts.
Explanation: masking essentially hides sharpening in low contrast areas. So what you probably see is the bounds of masked and unmasked areas.
Best regards
Erik
was a bit miffed with the raws Chris so kindly posted up as there were areas of colour that didn't look right to me especially the yellow by the fireplace where mixed lighting hasn't helped the smooth tones but they looked very posterised and gritty in adobe raw. tried the new capture one and i'm now interested in this camera again.
not a scientific test but i gave the shadows a boost and roughly tried to match them by adding some saturation in AR. C1 has done a much better job of handling this area.
-
I thought it was something to do with the compressed raw not the adobe defaults that encourage posterisation and gritty shadows. ???
I didn't use Lightroom but adobe raw (I presume the processing is the same though) and just used the defaults apart from a bit of saturation in AR and a hefty lift of the shadows in both to help bring out any noise.
Sharpening was at default on AR with standard look on C1
The C1 tiffs were much better looking than the adobe IMHO.
-
Hi,
I don't own the A7r and it is not on my shopping list, so I have now experience with it on my own. I have looked at the Sony data compression and I would say there is nothing that would cause artefacts.
I processed the images in both Capture 1 and Lightroom 5.3, I have much more experience with LR. Screen dumps of the results are enclosed.
Best regards
Erik
I thought it was something to do with the compressed raw not the adobe defaults that encourage posterisation and gritty shadows. ???
I didn't use Lightroom but adobe raw (I presume the processing is the same though) and just used the defaults apart from a bit of saturation in AR and a hefty lift of the shadows in both to help bring out any noise.
Sharpening was at default on AR with standard look on C1
The C1 tiffs were much better looking than the adobe IMHO.
-
with your processing they are very close (LR has more sharpening) i guess your familiarity with LR got the best out of the file.
i dont use LR or AR only C1. they were both at default except for a shadow lift to bring out any noise and a small increase in saturation for AR to match to C1. i guess the A7r profile in C1 gets you most of the way there but adobe needs tweaking to remove the posterisation/artefacts.
good to see raw compression has not adversely affected the images though.
-
Hi,
It is also the other way around, my lack of skill with C1 matters a lot.
Best regards
Erik
with your processing they are very close (LR has more sharpening) i guess your familiarity with LR got the best out of the file.
i dont use LR or AR only C1. they were both at default except for a shadow lift to bring out any noise and a small increase in saturation for AR to match to C1. i guess the A7r profile in C1 gets you most of the way there but adobe needs tweaking to remove the posterisation/artefacts.
good to see raw compression has not adversely affected the images though.
-
Hi,
It is also the other way around, my lack of skill with C1 matters a lot.
Best regards
Erik
Erik, it's a surprise to me that you don't use C1P1at least with your P45+… You should really give it another go… it makes (especially with your back) a lot of difference, MO is that this way you compare a "crippled" back which makes it very unfair for it.
-
Hi,
I had used it for nine months and I decided that LR 5 works better with me. Capture One is history for me!
I had rechecked C1 vs LR5 on some recent images and I still don't feel C1 is better. May be LR5 takes a bit more work, but I find it works better for me.
Best regards
Erik
Erik, it's a surprise to me that you don't use C1P1at least with your P45+… You should really give it another go… it makes (especially with your back) a lot of difference, MO is that this way you compare a "crippled" back which makes it very unfair for it.
-
I had rechecked C1 vs LR5 on some recent images and I still don't feel C1 is better. May be LR5 takes a bit more work, but I find it works better for me.
If you had just written that you preferred LR5, that'd be like okay makes sense, everyone is different. But if you write you think LR5 is better than C1 and you have a phase back, I really am surprised as I have posted to the other thread. There are some things LR does well, some very well, but its not the equal when all is considered. I still use LR and sometimes I'll render a RAW in C1 and then continue working in LR or Photoshop.
-
Hi,
The short version:
Sensors have gotten better. Making pixels smaller reduces the DR per pixel but the increased number of pixels helps a lot.
The long version, see below:
DR is a bit tricky. The technical definition is FWC/readout noise, with WFC is being Full Well Capacity, the number of electrons a pixel can hold. Readout noise is the noise in reading out the electrons.
For medium size pixels like (6 my) the FWC may be around 60000 electrons FWC is by and large proportional to pixel surface, even if I think some progress have been made. The major factor in DR has probably been the reduction in readout noise. Modern CMOS may have say 3 electron charges in readout noise while CCD is more like 12-15 electron charges.
Let's assume we have a P45+ with 6.8my pixels, and let it compare to a 3.4 my pixel sensor using modern CMOS-technology. Let's assume 65000 electron charges (EC) per pixel for the P45+ and 16 EC in readout noise. So we have 65000/16=4062, converting to EV we get 11.99 EV. DxO-Mark measures 11.75 EV at 50 ISO (in 'screen, mode).
Now, let's take a CMOS sensor with half of that diameter, and assume FWC 65000/4 = 16250, if we assume 4EC in readout noise we would get a DR of 4062 that is 11.98, nearly the same as the "fat pixel CCD".
Now the small pixel CMOS has four pixels instead of one. Would be print both at the same size and assuming a sensor of the same size, the combined FWC would be 65000, read noise adds in quadrature, so we would have sqrt (4 * 4 * 4) = 8
So DR would be 12.98 EV, the smaller CMOS pixels would have a 1 EV advantage over the large pixel CCD. Would we do the same math with a CCD instead, FWC would still be at 65000 but readout noise would be 32. So DR would reduce to 10.98 EV.
Newer generations of CCD sensors have less readout noise than the older ones, but they are still in the low tens. Regarding CMOS, some sensors have external analog-digital converters (ADC), the off chip ADCs tend to be noisy, so these cameras have CCD like readout noise. Many modern CMOS sensors have thousands of on chip converters that work in parallell, these are much less noisy.
What complicates the issues is that DR I describe is a technical definition, and represents a level of noise where the signal is barely perceptible. In photographs we want a better Signal Noise ratio (SNR), on the other hand lower end of the DR scale is normally what I would call deep shadows where we would have little detail anyway.
Midtones and highlights are more affected by "shot noise", the natural variation of incident photons, which is independent of the sensor technology and depends only on exposure, sensor surface and quantum efficiency QE). QE is the percentage of the incoming photons that are detected.
Hope this helps! Here is a recommended paper, describing it more deeply: http://theory.uchicago.edu/~ejm/pix/20d/tests/noise/index.html
Best regards
Erik
When doubling the number of pixels, the dynamic range on a pixel level is indeed reduced. but when looking at the whole 'picture', the dynamic range stays exactly the same (theoretical), since the noise is averaged over more pixels. averaging over double the number of pixels reduces the noise (assuming it really is noise), by 3dB, just like halving the size of the pixel (surface) resulted in an increase of 3dB.
-
The thing that surprised me the most is that the 645/35mmSchneider combo doesn't outperform the 35mm/24mmCanon lens one, in terms of distortion.
If anything the walls look more bowed in the former.
I'm very close to spending an awful lot of money on a Silvestri/Rodenstock/phase set up in an effort to make interiors look more natural.
I'm having second thoughts.
-
The thing that surprised me the most is that the 645/35mmSchneider combo doesn't outperform the 35mm/24mmCanon lens one, in terms of distortion.
If anything the walls look more bowed in the former.
I'm very close to spending an awful lot of money on a Silvestri/Rodenstock/phase set up in an effort to make interiors look more natural.
I'm having second thoughts.
The 35XL has less distortion than 24mm Canon (the Canon is not too bad though). I'd guess that the 35XL shot is done without distortion correction, and the 24mm Canon is done with. From measurements/data sheets: the 35XL has about 0.5% barrel distortion in its 90mm image circle (I'd say only 75mm of it is high quality though in terms of sharpness), and the TS-E24 has about 0.9% (which is not bad, not sure if it's entirely comparable though as the TS-E24 number is only for the 44mm image, I'd guess the true number is more like 1.5%). Due to the simple symmetric design of the 35XL it's not surprising that it's lower than for the heavily retrofocus TS-E24. In fact I would guess that the 35XL is the lens currently on the market with the least distortion of comparable focal length and image circle.
If you're worried about distortion note that the retrofocus Rodenstock lenses can have some, take a look at the data sheets to get an idea, the 23mm reaches 2%. Retrofocus design has many advantages but one of the disadvantages is that it introduces distortion.
(I own both the 35XL and the TS-E 24 lenses so I could make a test for you if you want to, I'd suspect that the TS-E24 distortion would detoriate more in shifted positions than the 35XL. I'm very pleased by the low of distortion in my 35XL, but I generally shoot landscapes so I'm not looking at straight lines as often as an interior photographer... attached a few examples of the 35XL with straight lines I happened to find in a quick browse, some with substantial shift, none of them is distortion corrected. The bridge might look a bit distorted anyway, because the bridge itself is distorted :) )
-
http://www.alpa.ch/en/products/tools/alpa-tools/alpa-lens-corrector-product.html
-
An unfortunate thing with lens correctors is that shift numbers are not stored in the exif data. This is actually also the case of DSLR shift lenses. So you need to write them down manually. One thing I like with my symmetrical lenses is that they have so low distortion I can ignore to correct it, so I don't need to keep track of which movements I had. But for extremely critical architecture work maybe even the SK35XL's 0.5% is too much.
-
An unfortunate thing with lens correctors is that shift numbers are not stored in the exif data.
With Leaf backs you can add the shift values in the EXIF. Not sure about the IQ backs ...
But you can also just shoot a reference image unshifted to recalculate the shifts later on.
-
With Leaf backs you can add the shift values in the EXIF. Not sure about the IQ backs ...
But you can also just shoot a reference image unshifted to recalculate the shifts later on.
I have a Leaf Aptus myself, yes you can enter stuff but I find this to be more cumbersome than to shoot the LCC shot, I'm not too good at using the touch screen.
If back makers really cared they could make a number of things in the GUI to make life easier for tech cam users. Say if you could pre-configure your back with what lenses you have and movement range you have, and had android/iphone style scrollboxes for selecting focal length (which you'd only do when changing lens) and aperture, shift X shift Y tilt and swing, instead of having to type all those details on a on-screen keyboard.
-
Without any distortion correction, the 35XL holds up pretty well.
Mies at IIT Crown Hall.
-
I have a Leaf Aptus myself, yes you can enter stuff but I find this to be more cumbersome than to shoot the LCC shot, I'm not too good at using the touch screen.
If back makers really cared they could make a number of things in the GUI to make life easier for tech cam users. Say if you could pre-configure your back with what lenses you have and movement range you have, and had android/iphone style scrollboxes for selecting focal length (which you'd only do when changing lens) and aperture, shift X shift Y tilt and swing, instead of having to type all those details on a on-screen keyboard.
I don't use a Leaf back... but I can imagine that editing the EXIFs is somewhat cumbersome (but at least they offer a solution... which is essentially positiv).
For some reason I can always remember the shift values applied for a certain capture (but I don't shoot high volumes ... and only very few captures of a respective motif). But, again, shooting a reference image unshifted to recalculate the shifts later in post is not a big deal.