November 17, 2017, 12:44:12 PM

Author Topic: Analysis of RAW samples at Fred Miranda show weak DR  (Read 165757 times)

privatebydesign

  • Canon EF 600mm f/4L IS II
  • **********
  • Posts: 6539
  • Would you take advice from a cartoons stuffed toy?
Re: Analysis of RAW samples at Fred Miranda show weak DR
« Reply #1050 on: August 17, 2017, 02:22:56 PM »
I download the RAW files from DPReview and do my own processing to them for any comparison I am interested in.

I do this too in order to get a feeling of the raw output from different cameras, but the problem I find is that virtually all the images, whether they be from DPR or others, are either exposed for an unedited ooc jpeg, or under exposed. Often the two go together. So it is difficult for me to establish from these on-line downloadable files what the results would be if the exposures were optimal for the kind of output I had in mind. This brings us back to the point you were making about being able to compare optimally exposed and processed files for a specific camera, rather than an across the range standard that often disadvantages Canon.
The tests we do are in a controlled environment using a sphere with an evenly lit aperture that can be consistently checked so we know its the same from one test to another. This has a grate that has an equivalent of a 20 stop DR change and shots taken with this using a base ISO of 100 (for instance) are then analysed automatically i.e. no human interpretation. The process we use is identical to manufacturers.

The same light sphere can also be used to shoot resolution tests with the CIPA High resolution chart which we also do however these are not automatically referenced and do rely on human "knowledge" the same as using a projector to test lenses for a number of aberrations but again nothing different to how the majors lens manufacturers work because we have visited them and had their training as well as our own.

Finally we shoot LOTS of footage / images in varying light conditions and know the ones that will really test equipment.

I stand by the Bill Claff comment my point was his methods are consistent.

Yes I agree his methods are consistent, what I am questioning is the relevance of the results.

What real world value do esoteric numbers have from an undemosaiced RAW file? My point is if every file has to run through an algorithm to be able to see the image then what is happening is we are seeing people applying the same post processing to different files or showing differences in unseeable data, neither of those has much relevance to people interested in actual output image quality. Yes that information can give pointers to performance, but neither illustrates what we can actually expect from the files if we bought the camera.

Further, to Sporgon's point, even consistent test images are regularly found wanting, have you seen the illumination difference between the right side (brighter) and left side (darker) of the test images DPReview puts out?

I am not saying jeffa can't be more consistent, or take better quality test images, what I am saying is how you present those comparisons is critical for that comparison to have meaning and relevance. Personally the only relevant comparison I am interested in is an optimally processed image that was optimally exposed because that is what I will be working with if I buy the camera.

Optimally processed invariably, and in my experience, means different processing for each camera.
Too often we lose sight of the fact that photography is about capturing light, if we have the ability to take control of that light then we grow exponentially as photographers. More often than not the image is not about lens speed, sensor size, DR, MP's or AF, it is about the light.

canon rumors FORUM

Re: Analysis of RAW samples at Fred Miranda show weak DR
« Reply #1050 on: August 17, 2017, 02:22:56 PM »

tomscott

  • EOS 5DS R
  • ******
  • Posts: 1072
  • Photographer & Graphic Designer
    • Tom Scott | Photography
Re: Analysis of RAW samples at Fred Miranda show weak DR
« Reply #1051 on: August 17, 2017, 04:18:38 PM »
6D MKII, 5D MKIII STOLEN, 7D MKII 70D 17-55mm F2.8 16-35mm F2.8 II L 24-70mm F2.8 L 24-105mm F4 L 70-200mm F2.8 II L 100-400mm F4.5-5.6 II L 2x II 1.4X III 580EX

Aglet

  • EOS-1D X Mark II
  • *******
  • Posts: 1610
Re: Analysis of RAW samples at Fred Miranda show weak DR
« Reply #1052 on: August 17, 2017, 08:13:20 PM »
..test images are regularly found wanting, have you seen the illumination difference between the right side (brighter) and left side (darker) of the test images DPReview puts out?

I believe that's a specific feature of the test so you can see the difference of those color wheels at different illumination levels.  Hue shift, noise, etc.

privatebydesign

  • Canon EF 600mm f/4L IS II
  • **********
  • Posts: 6539
  • Would you take advice from a cartoons stuffed toy?
Re: Analysis of RAW samples at Fred Miranda show weak DR
« Reply #1053 on: August 17, 2017, 08:29:37 PM »
..test images are regularly found wanting, have you seen the illumination difference between the right side (brighter) and left side (darker) of the test images DPReview puts out?

I believe that's a specific feature of the test so you can see the difference of those color wheels at different illumination levels.  Hue shift, noise, etc.

I don't.
Too often we lose sight of the fact that photography is about capturing light, if we have the ability to take control of that light then we grow exponentially as photographers. More often than not the image is not about lens speed, sensor size, DR, MP's or AF, it is about the light.

Talys

  • EOS 7D Mark II
  • *****
  • Posts: 564
  • Canon 6DII
Re: Analysis of RAW samples at Fred Miranda show weak DR
« Reply #1054 on: August 18, 2017, 01:00:15 AM »
Just added a few images from the wedding

http://www.canonrumors.com/forum/index.php?topic=33212.msg682114;topicseen#new

Anyone who is seriously interested in purchasing a 6DII should check out the link.  Tom's photos have a good mix of real-world uses, and then it's pretty easy to judge for yourself whether 6DII is good enough for you.

Plus, lovely photos.

Mikehit

  • Canon EF 300mm f/2.8L IS II
  • ********
  • Posts: 2149
Re: Analysis of RAW samples at Fred Miranda show weak DR
« Reply #1055 on: August 18, 2017, 02:54:03 AM »
Just added a few images from the wedding

http://www.canonrumors.com/forum/index.php?topic=33212.msg682114;topicseen#new

Anyone who is seriously interested in purchasing a 6DII should check out the link.  Tom's photos have a good mix of real-world uses, and then it's pretty easy to judge for yourself whether 6DII is good enough for you.

Plus, lovely photos.

To be fair, the key thing is not the quality of pictures from a skilled practitioner because I think the 6D/5D3 may well have got pretty close. Rather, it is how it measures up against the other options, and I put a lot of store in Tom's comments when he says he got images with the 6D2 that he would not have got with the 5D3. Next to the old trope of 'the best camera is the one in your hand' is 'do you know how to use the damned thing' and sometimes if the haptics of a camera make it more likely to get the shot then it is a better camera.

Talys

  • EOS 7D Mark II
  • *****
  • Posts: 564
  • Canon 6DII
Re: Analysis of RAW samples at Fred Miranda show weak DR
« Reply #1056 on: August 18, 2017, 03:18:56 AM »
Just added a few images from the wedding

http://www.canonrumors.com/forum/index.php?topic=33212.msg682114;topicseen#new

Anyone who is seriously interested in purchasing a 6DII should check out the link.  Tom's photos have a good mix of real-world uses, and then it's pretty easy to judge for yourself whether 6DII is good enough for you.

Plus, lovely photos.

To be fair, the key thing is not the quality of pictures from a skilled practitioner because I think the 6D/5D3 may well have got pretty close. Rather, it is how it measures up against the other options, and I put a lot of store in Tom's comments when he says he got images with the 6D2 that he would not have got with the 5D3. Next to the old trope of 'the best camera is the one in your hand' is 'do you know how to use the damned thing' and sometimes if the haptics of a camera make it more likely to get the shot then it is a better camera.

Indeed -- I couldn't agree more that he would have gotten great shots out of 6D/5D3 (for that matter, a 5DII), too, and his comments are very helpful. 

There's enough variety that it's a good indication that for a professional purpose (a) the 6DII isn't deficient and (b) the low light performance is pretty great.  Whether the latter is attributed to less actual noise, or noise that AI then processes in a more pleasing way is kind of immaterial to me.  I think that a reasonable person would conclude that this is a viable $2,000 option for taking those sorts of photos. 

What do most people who buy a $2,000 6DII for want out of their camera?  Frankly, I have no idea how that breaks down.  I bought my 70-200 L 2.8 II for 70% of retail from a guy who had it for 8 months that couldn't have been in more pristine condition, because, and it had, in his words, seen the light of day just once -- at the camera store.  And he would have died owning it, probably never having taken it out of the zippered pouch, had he not been moving overseas and wanted to get rid of easy-to-sell stuff.  So, go figure.  I'm sure there are people on the far opposite side of that spectrum too, and all sorts of skill levels.

But I think, for the vast majority of people who would consider the 6DII, the quality of the output and the usability are pretty good. 

canon rumors FORUM

Re: Analysis of RAW samples at Fred Miranda show weak DR
« Reply #1056 on: August 18, 2017, 03:18:56 AM »

Adelino

  • EOS Rebel T7i
  • ****
  • Posts: 107

tomscott

  • EOS 5DS R
  • ******
  • Posts: 1072
  • Photographer & Graphic Designer
    • Tom Scott | Photography
6D MKII, 5D MKIII STOLEN, 7D MKII 70D 17-55mm F2.8 16-35mm F2.8 II L 24-70mm F2.8 L 24-105mm F4 L 70-200mm F2.8 II L 100-400mm F4.5-5.6 II L 2x II 1.4X III 580EX

Keith_Reeder

  • EOS 7D Mark II
  • *****
  • Posts: 498
Re: Analysis of RAW samples at Fred Miranda show weak DR
« Reply #1059 on: August 20, 2017, 02:57:18 PM »
I stand by the Bill Claff comment my point was his methods are consistent.

And what's the value of consistent testing of inconsistent subject matter, exactly?

Mikehit

  • Canon EF 300mm f/2.8L IS II
  • ********
  • Posts: 2149
Re: Analysis of RAW samples at Fred Miranda show weak DR
« Reply #1060 on: August 20, 2017, 02:59:27 PM »
I stand by the Bill Claff comment my point was his methods are consistent.

And what's the value of consistent testing of inconsistent subject matter, exactly?

As I understand it, each frame (maybe this I unique to Canon, I don't know) has a small border where no image is recorded but is used as a noise reference point to generate the image, and Bill takes data from that area which would mean that image content is irrelevant to how he is taking his measurements.

I wait to be corrected but if so it may inspire someone to give an idiot's guise as to how he does it.

canon rumors FORUM

Re: Analysis of RAW samples at Fred Miranda show weak DR
« Reply #1060 on: August 20, 2017, 02:59:27 PM »