Another Stupid Camera Test: IQ 260 / A7r / 5d2 / Epic

[Hans van Driest]

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.

[eronald]

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

[ACH DIGITAL]

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

[Iliah]

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:


Given the choice between lossy raw and lossless raw, my preference is lossless as it takes postprocessing better and has more archive value.

[Hans van Driest]

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. 

[Iliah]

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.

[ErikKaffehr]

Hi,

Just an observation…

Best regards
Erik

[Isaac]

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 for Sony, Zeiss, Tamron, Sigma lenses.

[Ken R]

Hi,

Just an observation…

Best regards
Erik

Yep, IQ260 underexposed and at iso 100, not a good combo.

[JoeKitchen]

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. 

[BJL]

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.

[Chris Barrett]

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.

[Bart_van_der_Wolf]

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

[ErikKaffehr]

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.

[Chris Barrett]

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. 

[synn]

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.

[Ken R]

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. 

Thx. I agree.

[ErikKaffehr]

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

[AlanG]

Already seems to be supported 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.

[HarperPhotos]

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.