Pages: 1 2 [3] 4 5 ... 8   Go Down

Author Topic: If its not megapixels what is it?  (Read 46608 times)

hjulenissen

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2051
Re: If its not megapixels what is it?
« Reply #40 on: May 18, 2011, 06:18:05 am »

There is scientific evidence with regard to the higher conductivity of certain types of Oxygen-free copper. However, also scientifically demonstrated, the difference on the signal frequencies compared with common oxigen free labled copper wire is so small, that the effect is negligeable. The effects compared with no-oxigen free labled copper wire depends on the wire in question. Cable length of course amplifies the effects.

So, there is scientific evidence, and there is a difference (albeit extremely small between oxigen free varieties).
If connectors are used, the quality of the connectors (and the shielding) can also impact the sound quality.
I took the original poster as meaning implicitly scientific evidence of audible benefits of audiophile cables. In this context it is probably less interesting that two line level cables may measure differently in the MHz range, or that two digital images show there to be measurable differences in their chrominance. What IS interesting is that people are selling and buying cables costing the same as a used car without there (to my knowledge) ever having been published peer-reviewed, relevant research indicating that this has any influence whatsoever on what the listener actually hears (fondness bias excluded).

-k
Logged

John R Smith

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1357
  • Still crazy, after all these years
Re: If its not megapixels what is it?
« Reply #41 on: May 18, 2011, 06:32:06 am »

You are forgetting that with denser sensor sampling, the combined resolution of a lens and sensor array is improved (as can be seen in its MTF curve). So the image we start with (before resampling) has a higher resolution. In addition it will have lower aliasing tendency (which in turn allows more sharpening), but that's an other issue.

But surely one will only see this higher resolution and lower aliasing if all the pixels are present? Or are you saying that this is like audio, where if you record at 96 and then downsample to 48, you get a better result than recording at 48?

John
« Last Edit: May 18, 2011, 06:34:33 am by John R Smith »
Logged
Hasselblad 500 C/M, SWC and CFV-39 DB
an

hjulenissen

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2051
Re: If its not megapixels what is it?
« Reply #42 on: May 18, 2011, 06:37:17 am »

But surely one will only see this higher resolution and lower aliasing if all the pixels are present?
If you are downsampling using anything but insane methods, all of the source pixels will affect the output - "throwing away pixels" is not an accurate description.

Talking about a simple world here, where we always have a vast number of photons, etc. The downsampling method could be worse or better*) than switching to a lower resolution sampling grid in the first place, but I believe that it will usually be better.

An idealized monochrome 10 MP camera with no AA-filter, 100% pixel fill-rate and perfect lense should give an output similar to an equally idealized 40 MP camera where every 2x2 pixel have been averaged. That is a crude downsampling, software can do better (or the same, if you like).
Quote
Or are you saying that this is like audio, where if you record at 96 and then downsample to 48, you get a better result than recording at 48?

John
Is it? Audibly? Or is it just common/good practice to capture the possibly "once-in-a-lifetime" content using converters at their higheste advertised resolution?
-h
*)Talking about PSF/MTF here, not indirect consequences of having too small pixel sites etc
« Last Edit: May 18, 2011, 06:48:45 am by hjulenissen »
Logged

John R Smith

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1357
  • Still crazy, after all these years
Re: If its not megapixels what is it?
« Reply #43 on: May 18, 2011, 06:49:28 am »

If you are downsampling using anything but insane methods, all of the source pixels will affect the output - "throwing away pixels" is not an accurate description. The downsampling method could be worse or better than switching to a lower resolution sampling grid in the first place, but I believe that it will usually be better.

Aha. I don't understand the technicalities of downsampling, but if this is true for images, perhaps it might be true for audio as well. Both are in the digital domain, after all. So, perhaps the implication might be that for a given print size the more megapixels you start with the better, because even though downsampled it will look better than a lower megapixel back at native resolution? It's a bit hard to see that this could be true - after all, imagine an image composed of a grid array just 10x8 pixels. Would a 50MP image downsampled to that look better in some way? It's interesting, isn't it - how can we find out?

John
« Last Edit: May 18, 2011, 06:53:42 am by John R Smith »
Logged
Hasselblad 500 C/M, SWC and CFV-39 DB
an

Gigi

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 524
    • some work
Re: If its not megapixels what is it?
« Reply #44 on: May 18, 2011, 07:08:06 am »

As long as this discussion is ranging, can we consider the  "1% of the time" when you need a high quality back. That's always been a curious statement. I agree with it, but 1% of what? Is it of the 500 shots that the CaNikon shooter might take in a day? Or 1% of the scenes you see - how many scenes do you see in a day? Do you see differently with different gear?

Using quality equipment is not readily correlated to use patterns. The relationship is not linear. Nor is it exponential.... In fact, as one gets more choosy (assuming this is a good thing) about ones work, or the images, the need for higher quality tools becomes greater. THe usefulness of the tools is more notable. Their purposefulness becomes clearer. They even (dare we say) become more appropriate.

But this matter is far from straightforward. The use of high quality tools has a funny way of becoming insidious: it loops around a bit, as one then begins  to find images that work with the higher quality tools and are not possible with lesser equipment. That's no longer an issue of the 1% category, that's a whole other matter entirely. Imagine that you have images that you could not have taken without the fancy tool. Is the tool now to be regarded as still fancy, or has it become essential now?

To make this even more complex, how many of us have had experience of walking away from the sophisticated tool syndrome, trying a Holga (or equal) and enjoying the release of a no-worry approach?  

The point is that as we look for rationalizations or even logic for improving our toolkit, I'd like to suggest  that the relationship between the work and tools is quite complex. Its not just budget, or the new gizmo, although those can be factors. Game changers like the IQ series will impact work flow and even to some degree the products of the photographer. MFDB lets me be the 4x5 photographer I had wanted to be but didn't seem to do. Maybe its the gear, maybe the time in one's life.

One goal can be then  to find the best equipment that will extend the envelope  without breaking the bank or adding too much complexity. Each their own.
Logged
Geoff

Bart_van_der_Wolf

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 8815
Re: If its not megapixels what is it?
« Reply #45 on: May 18, 2011, 07:35:19 am »

As long as this discussion is ranging, can we consider the  "1% of the time" when you need a high quality back. That's always been a curious statement. I agree with it, but 1% of what? Is it of the 500 shots that the CaNikon shooter might take in a day? Or 1% of the scenes you see - how many scenes do you see in a day? Do you see differently with different gear?

Hi Geoffey,

I suppose that a significant part of the people who buy such high MP backs have more than a 1% of their assignments where the difference matters. Different horses ...

Cheers,
Bart
Logged
== If you do what you did, you'll get what you got. ==

hjulenissen

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2051
Re: If its not megapixels what is it?
« Reply #46 on: May 18, 2011, 07:39:44 am »

Aha. I don't understand the technicalities of downsampling, but if this is true for images, perhaps it might be true for audio as well. Both are in the digital domain, after all.
It is certainly true for audio, perhaps a little less so for photography.
Quote
So, perhaps the implication might be that for a given print size the more megapixels you start with the better, because even though downsampled it will look better than a lower megapixel back at native resolution? It's a bit hard to see that this could be true - after all, imagine an image composed of a grid array just 10x8 pixels. Would a 50MP image downsampled to that look better in some way? It's interesting, isn't it - how can we find out?
I believe that the 50MP would/could look a lot "better" when resampled to 10x8 pixels than a native 10x8 pixel sensor would on its own. Given, of course, that we did not have to worry about noise, dynamic range, storage cost/time, and all of those real-world hassles.

The 50MP image would give us a lot more information about how the scene "really was". This means that we can effectively delay the scene "gridding" (sampling to a pixel lattice) to a later point, perhaps sitting calm in front of [insert favorite image editor here] where we can try out different approaches and study their pros and cons side-by-side. Do you want an optical anti-aliasing filter or not? Both can be effectively be "simulated" when converting 50 MP to 10x8 pixels. Do you want to only convert the "blue" channel into a monochrome output with maximum amount of details? High-resolution sensor is the way. Do you want to simulate a native 10x8 pixel sensor? Find a suitable blurring kernel, and do NxM pixel averaging. Simulating the artifacts of Bayer at 10x8 is also possible, although more complex (and why would you?).

I do, however, believe that the importance of 50 MP over e.g. 25 MP is greatly exaggerated by many.

-h
Logged

Anders_HK

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1010
    • andersloof.com
Re: If its not megapixels what is it?
« Reply #47 on: May 18, 2011, 08:05:46 am »

Gents,

All respect but I believe you are a tad too focused on pixels. While 80MP can be a handy number, reality is that P1 generations of backs have gone 22MP, 30/39MP, 40/60MP to 80MP and Leaf generations 22MP, 28/33MP, 40/56MP to 80MP (sorry if I did not get the P1 correct). Thus we can expect there to soon be in likes of 100/120MP and after that NNN... etc...

So what? Some need the extra pixels, some do not. However the more interesting per say is what else there is? Having compared raw files from 56MP and 80MP Leaf backs to my mere 28MP;

Pixels aside, what I ultimately find most interesting with the IQ180 and Aptus/Afi-II 12 is the advancements that appear to be in image quality; there is finer gradation of color tones, broader dynamic range, improved color palette etc.

This thread with link was interesting http://www.luminous-landscape.com/forum/index.php?topic=54243.0  ;)

Regards
Anders
Logged

michael

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 5084
Re: If its not megapixels what is it?
« Reply #48 on: May 18, 2011, 08:32:26 am »

I think you missed my question. I agree with you that many, if not most, of the photographers on any short list of fine art landscape photographers for whom ultimate image quality is the end all/be all now use medium format backs, yet none of them, as far as I know, use technical cameras and lenses. This is true despite the fact that, at least in terms of theoretical resolution, as opposed to real world, what can you achieve in the field, sharpness, technical cameras and lenses are superior to medium format cameras and lenses. Why do you think that is? I read that you recently purchased an Alpa STC to use with a back. I will be curious to hear your views on whether you feel that the Alpa enhances or detracts from your ability to make compelling images.

Sorry about that.

I don't think that many use technical cameras. It comes down to a matter of personal style and experience as much as anything else. I come from a background of shooting with large format for landscape and nature and therefore am comfortable with a technical camera – it's demands and rewards. I have been frustrated over the past 5-8 years though by the limitations (mainly focusing) imposed when using MF backs on them.

It appears though that the combination of a back with Live View and Focus Mask, combined with top digital MF glass from Schneider and Rodenstock, plus the small size and light weight of a camera like the Alpa STC, works well for me.

Michael
Logged

michael

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 5084
Re: If its not megapixels what is it?
« Reply #49 on: May 18, 2011, 08:36:50 am »

Finally!! megapixels earn their place in voodoo land alongside silver cables for audiophiles..

There is no scientific evidence to show any difference between a normal el cheapo (shielded) copper cable linking your amp to speaker - but those with "golden ears" can really hear the difference between cables.. btw - if you can't 'hear' the difference you (obviously) don't have a 'golden ear'...
now a new one in photography land...something 'special' about 80 megapixel 11x8 prints  versus prints from ordinary lesser megapixel backs..if you can't see the difference ..you (obviously) don't have the 'golden eyes'...
 :D

next move is blind ABX testing of teh phenomenon - and we have photogrphicaudiophilia ...yee haaa!


Obviously you've never seen an 8X10" contact print. Obviously you've never seen an 11X17" print from an 80 Megapixel back. If you had you wouldn't be quite so snide in your comments.

It's easy to poke fun at things that we are not familiar with, isn't it?

Michael

Logged

michael

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 5084
Re: If its not megapixels what is it?
« Reply #50 on: May 18, 2011, 08:53:39 am »

Some Facts About Printing High Resolution Image Files

Assuming that we are talking about the concrete and visible rather than the theoretical, there is indeed a lot to be gained by starting with higher resolution when printing, and this has nothing to do with what we call downsampling.

Printers don't output a one-to-one image, they use dithering. The dithering algorithms, whether in the printer's firmware or in a RIP, are very sophisticated. They produce a decent image with very low input resolution (as low as 180ppi), and the latest printers can take advantage of as much as 720ppi original files, and use the data effectively.

What this means is that a 40-50-60-80MP back is doing more than simply giving you more megapixels for big prints and cropping. It's also is of considerable value when making small prints.

Since I got back from Utah I've made some 24X36" prints from my IQ180 files and am very pleased with them. I have also made some 11X17" prints and am even more pleased, because they really do shine.

That's why the silver audio cable comment was rude, gratuitous, and know-nothing. This issue has nothing to do with "golden eyed" perception and everything to do with the simple realities of contemporary printer driver dithering.

Michael
Logged

ternst

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 427
Re: If its not megapixels what is it?
« Reply #51 on: May 18, 2011, 09:07:41 am »

Count me as another landscape photographer (going on 36 years as a pro) who uses a tech camera - my Phase camera system has been gathering dust ever since I took delivery of an Alpa with those wonderful German lenses. It has not really slowed me down, and in some cases I can shoot faster than with the Phase camera. I always find it humorous when folks debate the merits of what someone else is buying - and then get into arguments about flea farts when another opinion is expressed that is opposite of theirs - who really cares - customers certainly do not. Buy and use what you want and go out and take pictures. I do understand that for many folks this sort of debate IS their hobby though and it is fun to be on a grand stage like Michael has provided, and so I guess that is OK...
Logged

hjulenissen

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2051
Re: If its not megapixels what is it?
« Reply #52 on: May 18, 2011, 09:09:12 am »

Some Facts About Printing High Resolution Image Files

Assuming that we are talking about the concrete and visible rather than the theoretical, there is indeed a lot to be gained by starting with higher resolution when printing, and this has nothing to do with what we call downsampling.

Printers don't output a one-to-one image, they use dithering. The dithering algorithms, whether in the printer's firmware or in a RIP, are very sophisticated. They produce a decent image with very low input resolution (as low as 180ppi), and the latest printers can take advantage of as much as 720ppi original files, and use the data effectively.

What this means is that a 40-50-60-80MP back is doing more than simply giving you more megapixels for big prints and cropping. It's also is of considerable value when making small prints.

Since I got back from Utah I've made some 24X36" prints from my IQ180 files and am very pleased with them. I have also made some 11X17" prints and am even more pleased, because they really do shine.

That's why the silver audio cable comment was rude, gratuitous, and know-nothing. This issue has nothing to do with "golden eyed" perception and everything to do with the simple realities of contemporary printer driver dithering.

Michael
I agree that the comment you are referring to was not constructive.

Most of your post seems to be a theoretical explanation for why high MP counts are good, and thus you are doing the very same thing that you seem to critizise. The fact that dithering algorithms may accept 360 "dpi" as input or 720 or 1440 does in no way guarantee that the output will benefit from it in a measurable or perceivable way. The fact that you are pleased with high MP count cameras does not guarantee that it is the number of MP that make you satisfied about them.

I believe that it is commonly accepted that dithering is a method of trading resolution between the spatial domain ("dpi") and the amplitude domain ("gradations"). In general, the domains could be swapped (time for audio), but lets talk about photography. We have a high spatial resolution printer with low amplitude resolution, we want to trade it for a more balanced space/amplitude resolution. Therefore, a printer could very well be able to spit out patterns at 720dpi or some other high resolution, without necessitating (or even benefiting from) an input that whose resolution matched that. Then we get to the issue of real end-to-end resolution and human vision.

I believe that it is possible to make sensible tests that gives near conclusive results. One suggestion is to downsample/upsample a 80MP image, print both and view them side-by-side at a reasonable distance/size. How much can you downsample before the differences are visible? If you can downsample/upsample those 80 MP to 25 MP without being able to spot the difference for a given (hopefully relevant to you) setup, would not that indicate that an ideal 25 MP would be "enough" for that kind of setup? Dpi or MP or Nyquist or whatever does not really count in this context, what we are able to see and what we like is what counts. Those technical terms may help us analyzing perceptual results, but I think that in the end, what will be used for humans should be judged by humans.

-h
« Last Edit: May 18, 2011, 09:13:40 am by hjulenissen »
Logged

gazwas

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 517
Re: If its not megapixels what is it?
« Reply #53 on: May 18, 2011, 09:11:36 am »

That's why the silver audio cable comment was rude, gratuitous, and know-nothing. This issue has nothing to do with "golden eyed" perception and everything to do with the simple realities of contemporary printer driver dithering.

While the silver audio cable comment was a little abrupt I do feel it has some truth.

I see a big difference between the files from my P65 and 1DsIII. I hear a big difference between the sound of my Michell turntable and iPod. However I do confess to have "Golden" eyes and ears because I'm looking/hearing for the differences and know what to look for.

Do my clients see a difference in my images or my wife hear a difference in the sound of my hi-fi..........?

I personally think not but I don't care because I do.
« Last Edit: May 18, 2011, 09:13:12 am by gazwas »
Logged
trying to think of something meaningful........ Err?

BillOConnor

  • Jr. Member
  • **
  • Offline Offline
  • Posts: 98
Re: If its not megapixels what is it?
« Reply #54 on: May 18, 2011, 09:44:06 am »

What is so different now than in the film days is there is no weight, size, susceptibility to wind, et al penalty for shooting with an 80 megapixel back over a 40. It seems to me, the conversation is a flashback to the old 4x5 vs. 8x10 arguments of yore except for the HUGE fact that, from 10 feet away, you can't tell if a photographer is using a 20 mp back or an 80. I'd be surprised if the 80 even weighed significantly more.
As for the cost difference, only the back costs more, everything else, camera(s), lenses (with a caveat here and there), tripod, cases, computer, printer, remain constant, so that the cost is not double to go from 40 to 80 mp, but more like 30 to 50% more. In the old days, that was not as bad as buying a Linhof Technika to replace your Wista Field. 
Logged

BernardLanguillier

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 13542
    • http://www.flickr.com/photos/bernardlanguillier/sets/
Re: If its not megapixels what is it?
« Reply #55 on: May 18, 2011, 09:56:16 am »

I use multishot backs for static scenes and repro, but I think you are vastly overestimating the percentage of work that can be stitched or captured with multishot for most photographers. Probably for those that shoot scenes that can be stitched, a scan back is a good idea too and a lot less money with higher DR.  Maybe most of your work is done with stitches, and for you its working fine but I think you go too far to suggest that people who consider these new backs have missed considering that technique.  I see that statement as both myopic and arrogant.   

I don't think you have read my comment correctly.

I am saying that people looking at the best possible image quality should stitch with their IQ180... for those images where this is an option.

Cheers,
Bernard

ErikKaffehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 11288
    • Echophoto
Re: If its not megapixels what is it?
« Reply #56 on: May 18, 2011, 10:24:30 am »

Hi,

Not an argument, but modern printers have very good native resolution, like 720 PPI on Epson's top models. So you don't need to print really large to see some benefits. The question is if the benefits are visible.

The other side of the coin is that it is very hard to produce comparable images, specially outside the "lab".

Best regards
Erik

Best regards
Erik

It strikes me, from the limited amount of testing that I have done with my own MF DB versus prints made from DSLR cameras with a much lesser megapixel count, that -

If we have two prints at say 16x20 from two MF DBs, one say 40MP and the other 80MP printed at the same resolution say 360 ppi of the same subject and framing using the same lens at the same distance -

Then both pictures as prints have the same number of pixels in each dimension.

If we arrange the image size so that the 40MP back is printing at or close to its native resolution (no upsampling or downsampling) there will be no loss of quality.

But the 80MP back will be downsampling to fit the image to the paper and will be losing its quality potential, as pixels are being discarded.

Therefore the apparent quality of the 80 MP image will be greatly dependent on the software downsampling algorithm, and we could expect different subjective responses to the two prints depending the printing application in use.

It is most unlikely that any image quality advantage seen from the 80MP back could be attributed to its having more megapixels, because the printing process has thrown lots of them away and the two prints are exactly the same in this regard. So the 80MP cannot be resolving more detail in the print, for example.

If there is a difference, it will surely be due to other factors, such as the manufacturer of the sensor, Bayer array design, photosite design and density, ADCs, and the firmware. All of these could contribute to smoother, more subtle colour reproduction, for example. An interesting test would be to repeat the printing exercise in monochrome, using a straight grayscale conversion, and see if there is still a detectable difference then. If not, then it is the colour handling that is the key, not zillions of pixels.

I certainly don’t dispute that Michael and others are seeing a difference in same-size small prints, but to me it seems unlikely that this has much to do with megapixel count, if the 80MP is being downsampled. As soon as the 40MP has to be upsampled, though, for large prints, then of course the 80MP should run away with the victory.

John

Logged
Erik Kaffehr
 

cunim

  • Full Member
  • ***
  • Offline Offline
  • Posts: 130
Re: If its not megapixels what is it?
« Reply #57 on: May 18, 2011, 10:43:33 am »

Some of us seem to be struggling with interactions between organic detectors and electronic signal sources.  Fascinating topic.  While it is true that very few people can hear the difference between amplifiers or preamplifiers with equivalent specs, it has been shown that experienced listeners can discriminate.  They can even distinguish (better than chance) one make/model of amplifier from another.  Clearly, there is more to the neurophysiology of music perception than the gear's simplest physical specifications can predict.

I would suggest that vision is similar.  Simple specs such as MP and SNR do not fully predict our responses.  These simple specs work to a point for everyone, but individual differences soon become decisive.  Most people are perfectly happy with web images downsampled from DSLR, but the trained perceiver is not.  He is capable of getting much more out of images - e.g "how well does this chip design respond to a noise reduction algorithm?"

In creating a market for subtly improved equipment one places initial samples in the hands of highly experienced individuals who can be relied upon to see the benefit right away.  They will reassure the rest of us that there is something to be gained there, once we train ourselves (and our clients) to see it.  Isn't it wonderful that we never stop learning?
Logged

John R Smith

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1357
  • Still crazy, after all these years
Re: If its not megapixels what is it?
« Reply #58 on: May 18, 2011, 12:56:01 pm »

Printing from LR to my Epson R2400, I can send a 10x8 print from my 39MP back at either 360 or 720 ppi, and the printer at highest output settings can deal with either without further resampling. But the difference between the prints is remarkably hard to see with the naked eye, even as close as you can focus. It is, however, clearly visible with an 8x loupe. In practice, I would not be able to tell between the two at normal viewing distance, which makes me wonder about what it is that Michael is seeing in his prints.

John
« Last Edit: May 18, 2011, 02:26:51 pm by John R Smith »
Logged
Hasselblad 500 C/M, SWC and CFV-39 DB
an

feppe

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2906
  • Oh this shows up in here!
    • Harri Jahkola Photography
Re: If its not megapixels what is it?
« Reply #59 on: May 18, 2011, 01:38:18 pm »

Printing from LR to my Epson R2400, I can send a 10x8 print from my 39MP back at either 360 or 720 ppi, and the printer at highestt output settings can deal with either without further resampling. But the difference between the prints is remarkably hard to see with the naked eye, even as close as you can focus. It is, however, clearly visible with an 8x loupe. In parctice, I would not be able to tell between the two at normal viewing distance, which makes me wonder about what it is that Michael is seeing in his prints.

Most likely placebo effect.

If anyone disagrees, I'll gladly change my POV if you setup a double-blind study - those are even rarer in photography circles than in audiophile circles.
Pages: 1 2 [3] 4 5 ... 8   Go Up