Why are the tests of IQ250 against CCD-backs taking so long?

[LKaven]

Typical f-stops for cinema work are f/2.8-4.  Usually not slower than that.

[Ken R]

Ok, first off, I have worked in probably 50 film for tv commercial productions and some movies over the past 11 years. I am a member of the Cinematographers Guild (Local 600). I know several DP's and have worked side by side with them as a unit stills photographer.

In some cases DPs like to shoot wide open. Whether it is with low budget Zeiss Superspeeds (T1.2-1.5), the uber expensive Master Primes (T1.3), the Cooke S4/S5's (T2) or any other lens set of their choice. Yes, focus pullers hate this (usually the 1stAC) but it is their job and quite a few do it very well. In other cases they shoot at T5.6-T8. Just like photographers, cinematographers have their preferences.

The "cinematic look" (if there is such a look per se) can be attributed to the mostly wide aspect ratio horizontal framing, generally well lit and produced scenes and superb color correction that generally is used to create a desired look and feel and not necessarily to be color accurate.

That said (going back to the IQ250) I am looking at some of the RAW files that Doug made and they look really really good. They do look a touch more saturated and contrasty compared to the IQ260 (both at base iso) but the color looks nice.

[BernardLanguillier]

Not quite; at least not yet. It is becoming more and more clear that the difference in color handling between sensors for medium format cameras and those for smaller formats is primarily a difference in the designs of the color filter arrays, in turn dictated by different priorities: color accuracy, low light performance, handling weird light sources like fluorescent .... (I suppose it could also be related to sensor/pixel size and to differences in CFA design expertise between Kodak/Truesense and Dalsa on one side and Sony, Canon, Nikon, etc. on the other.)  The idea that differences in how the electrons in a photo-site are processed (the only substantial difference between CCD and CMOS technology) effect color handling seems to be on its way out.

If it is just a matter of designing the CFA for more accurate, robust color, and Sony if has the ability to do that, then for now, the formats larger than 36x24mm will maintain that advantage, along with the resolution advantage. (Aside: many subjects are not amenable to stitching, so "greater one-shot resolution" is still a real advantage in many situations.)

But maybe Sony and Nikon (or Canon) will at some time decide to offer "high color accuracy priority" models in 36x24mm format, just as Nikon and others have started offering the option of no AA filter.  Then the case for choosing formats like 44x33mm and 54x40mm rather than 36x24mm would shrink a bit more. (My guess is that other factors will keep the "bigger than 36x24" sector alive, but that it will continue to shrink in size.)

Yes, color appears to be a bit better, we'll see what comes out from Nikon and Canon next since the sensor of the D800 is 2.5 years old now.

Now, we were told by P45+ owners for years that the DR of their back was in a totally different leagues compared to DSLRs like the D3x... only to find out a few years later after those back became mainstream second hand... that their DR is in fact not that much better... not to say that it is worse... only did the systematic under-exposure of RAW files give the illusion of highlights recovery.

So there are always "huge differences" in favor of the latest backs but I tend to take those claims with a pinch of salt. Huge is a very relative thing and modesty isn't part of modern marketing tactics.

Cheers,
Bernard

[eronald]

Disclosure: I know nothing whatsoever about cinema capture.
I probably misused my adjectives

By "cinematic feel", I intended to describe the isolation of the subject from background which is often achieved by a combination of lighting, depth of field and camera motion.

Edmund

Ok, first off, I have worked in probably 50 film for tv commercial productions and some movies over the past 11 years. I am a member of the Cinematographers Guild (Local 600). I know several DP's and have worked side by side with them as a unit stills photographer.

In some cases DPs like to shoot wide open. Whether it is with low budget Zeiss Superspeeds (T1.2-1.5), the uber expensive Master Primes (T1.3), the Cooke S4/S5's (T2) or any other lens set of their choice. Yes, focus pullers hate this (usually the 1stAC) but it is their job and quite a few do it very well. In other cases they shoot at T5.6-T8. Just like photographers, cinematographers have their preferences.

The "cinematic look" (if there is such a look per se) can be attributed to the mostly wide aspect ratio horizontal framing, generally well lit and produced scenes and superb color correction that generally is used to create a desired look and feel and not necessarily to be color accurate.

That said (going back to the IQ250) I am looking at some of the RAW files that Doug made and they look really really good. They do look a touch more saturated and contrasty compared to the IQ260 (both at base iso) but the color looks nice.

[ErikKaffehr]

Hi,

I don't think colour accuracy is the issue at hand, it is more about pleasant or adequate colour rendition possibly under set conditions.

Accuracy is something that can be measured. In my case Adobe Standard profile is much more accurate compared with Capture One's standard profile on my P45+, and the SLT 99 runs circles around both.

But, the quite good Adobe Standard profile on the P45+ is a bit ugly on some real subjects. I have built my own DCP profiles that are more to my taste, and I am now quite happy with colour rendition.

The DeltaE figures are enclosed, the full article is here: http://echophoto.dnsalias.net/ekr/index.php/photoarticles/79-p45-colour-rendition

The article is about the P45+ and not comparing with the SLT 99. I don't know which rendering I prefer.

Best regards
Erik

Not quite; at least not yet. It is becoming more and more clear that the difference in color handling between sensors for medium format cameras and those for smaller formats is primarily a difference in the designs of the color filter arrays, in turn dictated by different priorities: color accuracy, low light performance, handling weird light sources like fluorescent .... (I suppose it could also be related to sensor/pixel size and to differences in CFA design expertise between Kodak/Truesense and Dalsa on one side and Sony, Canon, Nikon, etc. on the other.)  The idea that differences in how the electrons in a photo-site are processed (the only substantial difference between CCD and CMOS technology) effect color handling seems to be on its way out.

If it is just a matter of designing the CFA for more accurate, robust color, and Sony if has the ability to do that, then for now, the formats larger than 36x24mm will maintain that advantage, along with the resolution advantage. (Aside: many subjects are not amenable to stitching, so "greater one-shot resolution" is still a real advantage in many situations.)

But maybe Sony and Nikon (or Canon) will at some time decide to offer "high color accuracy priority" models in 36x24mm format, just as Nikon and others have started offering the option of no AA filter.  Then the case for choosing formats like 44x33mm and 54x40mm rather than 36x24mm would shrink a bit more. (My guess is that other factors will keep the "bigger than 36x24" sector alive, but that it will continue to shrink in size.)

[EricWHiss]

Fine numbers for reproducing big fields of homogenous color such as the squares on the Macbeth chart have little to do with the color differences between CCD sensors and CMOS sensors, and pleasant reproduction is tangential since that all can be changed with an ICC profile or a the color tools in post.      Being able to discern a subtle color change between two small adjacent regions has been a primary advantage of the CCD sensors in MFDB, as well as more accurately capturing tonal gradation across a subject.

I've tested the d800e side by side to my Kodak and Dalsa sensored MFDB and when you zoom into areas on flowers, skin, or things like green leaves on trees - the MFDB files have lots of color gradation where the D800e will be flat and even.  The tree leaves at a certain size and below become one color of green, not the yellow, green orange in the MFDB files.  A patch of skin is not one flat color, but a huge range of colors - red, yellow, blue, purple, brown.  Just zoom in on a MFDB shot (that hasn't been to the retoucher and is taken on a model without foundation or make-up everywhere.) all these little colors are there. Now look at the D800 - its one big mush.   

I believe the illusion of depth can be enhanced with better color detail and tonality (and by larger sensors / film plane ) and that is why I think this all matters. A more believable image is what I'm after.    It's weird too to circle back to film, because film has really less accurate colors and I'm sure less detail too, but in a way a color film image can look like it has more depth - probably because of greater 'tonality'.

I won't make a comment on the IQ 250 since I haven't done my own tests, but that's what I'd be looking for - the ability to discern small color differences between adjacent areas. 

[torger]

I'm a bit suspicious about the tonal gradient argument, haven't really seen it myself but I can't say that I've made a thorough comparison either. It's a thing than could be quite easily measured though, both on processed JPEGs and directly in the raw file. It would be an interesting test.

It could be the case that what we actually see is that the slightly noisier CCD provides a more pleasing tonal structure rather than a more accurate. It could also be the case that a higher full well capacity on the larger CCD pixels provides a real advantage in the midtones thanks to a lower shot noise.

When it comes to small differences it's extremely easy to trick oneself into seeing things we want to see, if we want CCD to be better we'll see that it is better, and the opposite if we want CMOS to be better. Therefore it can be good to sanity check with measuring and/or do blind testing, if one really is interested in an objective evaluation.

I'm suspecting that it's about noise characteristics rather than actual tonal gradients, that the CCD has a smoother noise and the CMOS has blotchier noise. If so it can be quite hard to show in a measurement. I don't think CMOS have to have blotchy noise characteristics though, but it's been quite common.

Here's a snap from an old side-by-side comparison of 3 stop pushed color patches Aptus 75 top left, D7000 (sony exmor, very similar pixels to D800) top right and Canon 5Dmk2 bottom. The small bottom left patch is showing the original brightness before pushing.



Note the smoother noise of the Aptus 75 and the blotchy noise in the D7000 blue patch, I also think the red patch is a bit blotchier. Unfortunately that test did not include any object where one could side-by-side compare a gradient object, a fine gradient from one color to another would be a nice test patch.

[synn]

Fine numbers for reproducing big fields of homogenous color such as the squares on the Macbeth chart have little to do with the color differences between CCD sensors and CMOS sensors, and pleasant reproduction is tangential since that all can be changed with an ICC profile or a the color tools in post.      Being able to discern a subtle color change between two small adjacent regions has been a primary advantage of the CCD sensors in MFDB, as well as more accurately capturing tonal gradation across a subject.

I've tested the d800e side by side to my Kodak and Dalsa sensored MFDB and when you zoom into areas on flowers, skin, or things like green leaves on trees - the MFDB files have lots of color gradation where the D800e will be flat and even.  The tree leaves at a certain size and below become one color of green, not the yellow, green orange in the MFDB files.  A patch of skin is not one flat color, but a huge range of colors - red, yellow, blue, purple, brown.  Just zoom in on a MFDB shot (that hasn't been to the retoucher and is taken on a model without foundation or make-up everywhere.) all these little colors are there. Now look at the D800 - its one big mush.   

I believe the illusion of depth can be enhanced with better color detail and tonality (and by larger sensors / film plane ) and that is why I think this all matters. A more believable image is what I'm after.    It's weird too to circle back to film, because film has really less accurate colors and I'm sure less detail too, but in a way a color film image can look like it has more depth - probably because of greater 'tonality'.

I won't make a comment on the IQ 250 since I haven't done my own tests, but that's what I'd be looking for - the ability to discern small color differences between adjacent areas. 

This is absolutely true.

The main difference I see by applying the IQ250 profile on a D800 file is that the overall color cast is reduced and the file looks less "Global". The Credo 40 easily has better separation between subtle tonal differences.

[LKaven]

The advantages in mid-tone response of an MFDB seems more of a function of the number of photons being collected overall. 

[Paul2660]

I'm a bit suspicious about the tonal gradient argument, haven't really seen it myself but I can't say that I've made a thorough comparison either. It's a thing than could be quite easily measured though, both on processed JPEGs and directly in the raw file. It would be an interesting test.

It could be the case that what we actually see is that the slightly noisier CCD provides a more pleasing tonal structure rather than a more accurate. It could also be the case that a higher full well capacity on the larger CCD pixels provides a real advantage in the midtones thanks to a lower shot noise.

When it comes to small differences it's extremely easy to trick oneself into seeing things we want to see, if we want CCD to be better we'll see that it is better, and the opposite if we want CMOS to be better. Therefore it can be good to sanity check with measuring and/or do blind testing, if one really is interested in an objective evaluation.

I'm suspecting that it's about noise characteristics rather than actual tonal gradients, that the CCD has a smoother noise and the CMOS has blotchier noise. If so it can be quite hard to show in a measurement. I don't think CMOS have to have blotchy noise characteristics though, but it's been quite common.

Here's a snap from an old side-by-side comparison of 3 stop pushed color patches Aptus 75 top left, D7000 (sony exmor, very similar pixels to D800) top right and Canon 5Dmk2 bottom. The small bottom left patch is showing the original brightness before pushing.



Note the smoother noise of the Aptus 75 and the blotchy noise in the D7000 blue patch, I also think the red patch is a bit blotchier. Unfortunately that test did not include any object where one could side-by-side compare a gradient object, a fine gradient from one color to another would be a nice test patch.

Good test/comparison.  The noise characteristics in your Canon 5D MK2 are so familiar to me.  The blue box is covered with red noise and if you zoom in (best done from an ipad) you can see some vertical banding in the Canon also.  I found this in all canon files until the 6D, which has a much more even noise, almost like film grain.   The 7100 shot also is not surprising to me, as all of the 24MP crop sensors I have used/demo'd show pretty harsh noise somewhere.  The Sony Nex-7 was even worse than the 7100 here. 

I agree in your test the Aptus is cleaner in the way the noise is displayed. 

To me one of the miss-impressions of the the Phase P45+ was that it had such great DR.  I never found that with mine.  And the DxO results for the P45+ were around 77.  Highlights were always a problem for me with the P45+ in outdoor work.  However the range of the P65+/160 is much better at base ISO of 50.  The P45+ most times was a exposure bracket camera for me, no different than the Canon's I was using at the time.  If you exposed for the highlights, you would tend to underexpose the shadows and when attempting to pull them up later in post, the details were not there and what details were there tended to have a strange look to them.  And there tended to be a considerable loss of color/sat in these same shadows.  So I regularly bracketed with the P45+ and got very good results.  This was all at base ISO 50 on the P45+.  When Capture One 7.x came out with the new processing engine, I did go back and rework some old P45+ files and did see some improvements in my problem shadows, but not to the troublesome highlights.  Only the Linear curve seemed help on those and I did not prefer the look of that output.

With the 160/260 Phase CCD I do tend to see more room in both Highlight recovery and shadows.   

Paul C.

[bjanes]

I've tested the d800e side by side to my Kodak and Dalsa sensored MFDB and when you zoom into areas on flowers, skin, or things like green leaves on trees - the MFDB files have lots of color gradation where the D800e will be flat and even.  The tree leaves at a certain size and below become one color of green, not the yellow, green orange in the MFDB files.  A patch of skin is not one flat color, but a huge range of colors - red, yellow, blue, purple, brown.  Just zoom in on a MFDB shot (that hasn't been to the retoucher and is taken on a model without foundation or make-up everywhere.) all these little colors are there. Now look at the D800 - its one big mush.   

Why don't you post some images demonstrating such massive differences? Talk is cheap, but seeing is believing.

Bill

[synn]

Why don't you post some images demonstrating such massive differences? Talk is cheap, but seeing is believing.

Bill

I have posted several, but people seem to be more into graphs than images.

[Theodoros]

Why don't you post some images demonstrating such massive differences? Talk is cheap, but seeing is believing.

Bill

He has, his (correct) observations where in coincidence with mine, his D800 (as mine) "crack" when one "digs" into shadows and developed colour casts by exploding the "blue" part of "deep grey" (obviously the red/green frequencies aren't there) as LL information. Now, …that's not (usable) DR extension …is it?

[Ken R]

I have posted several, but people seem to be more into graphs than images.

Exactly. I won't even bother. For a LOT of people that post on these and other forums the differences in quality (AND the differences in Characteristics!) between the DSLRs and the MFDBs are not big enough for them to the large cost differences between the systems. (even if it was I bet most will not pony up the $ to buy into the systems) Frankly, since we have already established that choosing a camera system is a very personal choice it is just impossible to convince anyone that the MFDB expense is worth it. Obviously for those that actually OWN MFDBs we have already answered that. And not from reading stuff on the internet but from actually working with the systems ourselves.

There is no question that the 50-60MP and obviously the 80MP digital backs offer a VERY significant resolution increase over ANY Dslr. It is not even close. It is most evident when using tech camera wide angles. Whether you need that, want that, are willing to pay for it it is another matter. Is it worth it? Only you can answer that. But, I am very glad other tools are available for us photographers.

The latest Digital backs (Id say from the P40+ and P65+ sensor till today) offer not only great resolution potential but also really good dynamic range comparable to the best DSLRs. It is splitting hairs in regards to DR no matter how you analyze it.

And my experience with the IQ160 is that it has a significantly more color differentiation than any DSLR I have used (including the D800E). That is VERY evident in some landscape scenes that I have photographed. It is quite obvious with foliage.

Also, never mind that Digital Backs fit in several camera systems and that they work great when tethered. Of course, if that does not matter to you because you never use it it does not make it irrelevant for those who do.

[torger]

I think the argument is actually about if IQ250 is going to suck or not because it's a CMOS

[Bart_van_der_Wolf]

I think the argument is actually about if IQ250 is going to suck or not because it's a CMOS

Indeed. According to some (not me), the IQ250 has to suck because it is CMOS ...

Cheers,
Bart

[EricWHiss]

Why don't you post some images demonstrating such massive differences? Talk is cheap, but seeing is believing.

Bill

It's easy to post a request for someone to serve you RAW files and also conclusions isn't it?   I've posted what to look for and  I'm sure those that are curious enough will spend time to make their own tests and conclusions.  I imagine imaging professionals, pro photographers, digital back designers and maybe those that sell digital cameras will not be satisfied with my tests and will chose to run their own anyhow.

There are a plethora of RAW files available for download from all the different cameras (except the IQ 250). You and anyone who chooses to spend maybe five or ten minutes to have a look can do so.

The only other thing I can add to this is that last Summer I spoke to Norman Koren of Imatest about how to better quantifiably test for these differences.  He suggested that the new Imatest 'Dead Leaves' test may be helpful in illustrating the differences.

[EricWHiss]

Actually, I thought of more thing to add to this while we all wait for more IQ 250 RAW files to be available:  Multishot backs can deliver even more better color tonality and separation that single shot backs. While not connected to the IQ 250,  I am excited that Hasselblad is going to offer their CMOS sensor equipped back with mutlishot capability. 

When the pixel wars are over the next frontier might be on color - maybe we'll see larger Foven type sensors or other sensors with 4 different color filters instead of just three?

[torger]

An IQ260 achromatic with a filter wheel should be the ultimate multi-shot back, right? For reproduction it could be feasible. Filter wheels have been used before.

Concerning CCD vs CMOS etc raw files shot side by side on the same subject would be needed for a proper comparison. Differences is probably going to be so small that you'd have to make A/B testing to draw any reliable conclusions. It would be great to see that with IQ260 vs IQ250 and throw in a D800 or A7r for fun.

[Jim Kasson]

An IQ260 achromatic with a filter wheel should be the ultimate multi-shot back, right? For reproduction it could be feasible. Filter wheels have been used before.

Or, if you're going to go to sequential capture, you can use a scanning back today. The IQ and freedom from false color is pretty amazing. Your subject needs to be motionless, but that's true for multishot and filter wheel work as well. You've got a kind of live view that you can use for focusing. You can find the sharpest f-stop in seconds. You have a huge assortment of lenses to choose from. You've got swings and tilts. However, it's harder to get bright continuous lights than bright strobes.

While expensive, scanning backs are quite a bit cheaper than the equivalent MF backs.

Jim