Pages: [1] 2 3   Go Down

Author Topic: Raw DR ratings based on one camera!  (Read 27199 times)

uuglypher

  • Newbie
  • *
  • Offline Offline
  • Posts: 34
Raw DR ratings based on one camera!
« on: May 17, 2016, 10:41:29 am »

I have been informed, in direct response to my question posed to DxOMark “How many individual cameras of a given make and model are tested for Dynamic range?” that the answer is “one camera”.  DxOMark tests raw capture DR at different ISOs and reports the highest DR observed, with no mention of the associated ISO.

Several observations/questions come to mind.

First and foremost is that independent testing of the amount of extra raw-accesible dynamic range (ERADR) that exists beyond the point at which the JPEG-adjusted histogram trips the overexposure warning “blinkies” or “zebras” reveals considerable quantitative differences among cameras of the same make and model. In one popular full-frame DSLR the ERADR varies from ⅔ stop to Two and two-thirds stops (And possibly more than three stops according to unsubstantiated rumor).

Obviously, there is no assurance that the camera makers attempt to standardize to any reliable degree the DR of camera's of a single model. Not being au fait with camera production and marketing I cannot presume to offer a definitive reason for this state of affairs re: unreliable DR ratings of any particular camera model. 

 However, one likely cause does come to mind.

The semiconductor industry has, since Shockley co-invented the transistor up to the present day, been plagued with the phenomenon of “process variance” , “process variability”, or “process excursion”,  the detection and limiting of which has been inadequate to prevent the obvious variance in dynamic range provided by our cameras’ light sensors. Literature from the semiconductor industry and researchers on causes of and proposed remedies for “process variance” continue, at pace, to the present.

 It appears, then,  that the camera manufacturers accept sensors guaranteed to provide a minimum required DR, then claim that there is a bit more DR “overhead” adequate to help “recover some blown highlights’ ...and the sweep the rest of the ERADR Under the rug, hoping no one will notice.

Well, those of us concerned with capturing the highest possible quality image data have noticed!

The reason we notice is that failure to fully utilize the brightest available stop of DR results in sacrifice of 50% or more of the total tonal spectrum and chromatic available to the image file (absolute numbers of lost tones depends on the bit-depth of the captured file). Failure to use the next brightest stop sacrifices 75% of tones, and failure to use the third brightest stop would leave the image file with only 12.5% of its tonal “birthright”.  Additionally, use of progressively less of the available dynamic range at any ISO results in capture of progressively more noise. The photographer ignores his camera’s extra raw-accessible dynamic range (ERADR) at considerable risk to print quality.

The camera manufacturers ought explicitly notify potential buyers and current owners of
this state of affair relative to the indeterminate dynamic range of their  camera model, and
the means by which to test their camera to determine its amount (in useable stops or parts thereof) of extra raw-accessible dynamic range (ERADR).

Those parties purporting to report accurate and useful tests on new camera models ought test an adequate number of camera's of each model.  Results on dynamic range should be reported in terms of number of sensors tested, ISO’s at which tested, and the means,  medians,  and standard deviations from the means (assuming parametric distribution of data which, though likely, must be characterized as such).

I suggest that at the moment, except for the photographers who  have independently tested their camera’s allotment of ERADR and actually use that extra DR ,  no photographers are utilizing their camera’s available dynamic range to the maximum possible extent.

It would be wise to recall and consider the late Bruce Fraser’s caution that proper exposure of raw image data for maximum image data quality  is accomplished by exposing as closely as possible to blowing highlights, without actually doing so!

Sage advice, indeed!

Dave (uuglypher)
Logged

AlterEgo

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1995
Re: Raw DR ratings based on one camera!
« Reply #1 on: May 17, 2016, 11:38:24 am »

It would be wise to recall and consider the late Bruce Fraser’s caution that proper exposure of raw image data for maximum image data quality  is accomplished by exposing as closely as possible to blowing highlights, without actually doing so!
actually it would be wise to remember that not all "highlights" are actually important and some actually might be OK to clip in 1, 2, 3, ... raw channels in return for a better S/N in more important parts of the image.
Logged

AlterEgo

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1995
Re: Raw DR ratings based on one camera!
« Reply #2 on: May 17, 2016, 11:42:55 am »

The reason we notice is that failure to fully utilize the brightest available stop of DR results in sacrifice

not all cameras behave properly near clipping in raw channels... as for the perennial "tonal values" if you use Sony's compressed raws you might be surprised expecting more "tonal values" as you move closer to clipping  ;D
Logged

uuglypher

  • Newbie
  • *
  • Offline Offline
  • Posts: 34
Re: Raw DR ratings based on one camera!
« Reply #3 on: May 17, 2016, 12:36:22 pm »

Hi, AlterEgo.
both of your observations are certainly valid, and both, I think, are best approached on a case-by -case basis with full awareness of the camera's full complement of DR (the purpose of my original post)
Thanks,
Dave (uuglypher)
Logged

Dave Ellis

  • Full Member
  • ***
  • Offline Offline
  • Posts: 103
Re: Raw DR ratings based on one camera!
« Reply #4 on: May 17, 2016, 04:06:06 pm »

I have been informed, in direct response to my question posed to DxOMark “How many individual cameras of a given make and model are tested for Dynamic range?” that the answer is “one camera”.  DxOMark tests raw capture DR at different ISOs and reports the highest DR observed, with no mention of the associated ISO.

Hi Dave

In fairness to DXOMark, I think we should note that they do publish curves of DR vs ISO. As for testing more than one camera sample only, I guess that is a matter for them in terms of resources they can sustain. They do make the information available to us for free.

First and foremost is that independent testing of the amount of extra raw-accesible dynamic range (ERADR) that exists beyond the point at which the JPEG-adjusted histogram trips the overexposure warning “blinkies” or “zebras” reveals considerable quantitative differences among cameras of the same make and model. In one popular full-frame DSLR the ERADR varies from ⅔ stop to Two and two-thirds stops (And possibly more than three stops according to unsubstantiated rumor).
There are a couple of points to raise her.
1. The point at which blinkies come in is dependant on the (User controlled) Picture Control settings in the camera, even when shooting raw. Settings such as Brightness and Contrast can affect the tonal distribution in the jpeg.
2. One of the first things that needs to be done when processing a raw file, either in-camera or with external software, is that a White Point (and Black Point) must be set. The raw data is just a set of numbers with no inherent fixed range from black through to white. Sometimes the raw file may contain a suggested figure for the raw processor to use, sometimes not. It seems to me that the in-camera white point settings used by manufacturers tend to be conservative, probably to allow for variations in the characteristics of cameras from one unit to another.
 
It appears, then,  that the camera manufacturers accept sensors guaranteed to provide a minimum required DR, then claim that there is a bit more DR “overhead” adequate to help “recover some blown highlights’ ...and the sweep the rest of the ERADR Under the rug, hoping no one will notice.

That's probably a bit unfair to camera manufacturers. I haven't ever seen camera manufacturers publishing DR specs for their cameras. It would be too much of a "can of worms".

I think what it boils down to is that if individuals want to get the best DR out of their camera, they need to do some simple tests themselves to get a feel for how much "highlight margin" they actually have.

Dave
Logged

AlterEgo

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1995
Re: Raw DR ratings based on one camera!
« Reply #5 on: May 17, 2016, 04:16:36 pm »

First and foremost is that independent testing of the amount of extra raw-accesible dynamic range (ERADR) that exists beyond the point at which the JPEG-adjusted histogram trips the overexposure warning “blinkies” or “zebras” reveals considerable quantitative differences among cameras of the same make and model. In one popular full-frame DSLR the ERADR varies from ⅔ stop to Two and two-thirds stops (And possibly more than three stops according to unsubstantiated rumor).
that probably was just a faulty testing (do you happen to have a link ?) - once you start using UniWB and make a proper OOC JPG parameters everything will be nice within <= 1/3 EV.
« Last Edit: May 17, 2016, 05:16:40 pm by AlterEgo »
Logged

uuglypher

  • Newbie
  • *
  • Offline Offline
  • Posts: 34
Re: Raw DR ratings based on one camera!
« Reply #6 on: May 17, 2016, 05:01:02 pm »

Hi Dave

In fairness to DXOMark, I think we should note that they do publish curves of DR vs ISO. As for testing more than one camera sample only, I guess that is a matter for them in terms of resources they can sustain. They do make the information available to us for free.
There are a couple of points to raise her.
1. The point at which blinkies come in is dependant on the (User controlled) Picture Control settings in the camera, even when shooting raw. Settings such as Brightness and Contrast can affect the tonal distribution in the jpeg.
2. One of the first things that needs to be done when processing a raw file, either in-camera or with external software, is that a White Point (and Black Point) must be set. The raw data is just a set of numbers with no inherent fixed range from black through to white. Sometimes the raw file may contain a suggested figure for the raw processor to use, sometimes not. It seems to me that the in-camera white point settings used by manufacturers tend to be conservative, probably to allow for variations in the characteristics of cameras from one unit to another.
 
That's probably a bit unfair to camera manufacturers. I haven't ever seen camera manufacturers publishing DR specs for their cameras. It would be too much of a "can of worms".

I think what it boils down to is that if individuals want to get the best DR out of their camera, they need to do some simple tests themselves to get a feel for how much "highlight margin" they actually have.

Dave
Logged

uuglypher

  • Newbie
  • *
  • Offline Offline
  • Posts: 34
Re: Raw DR ratings based on one camera!
« Reply #7 on: May 17, 2016, 07:19:50 pm »

Hi Dave

In fairness to DXOMark, I think we should note that they do publish curves of DR vs ISO. As for testing more than one camera sample only, I guess that is a matter for them in terms of resources they can sustain. They do make the information available to us for free.
There are a couple of points to raise her.
1. The point at which blinkies come in is dependant on the (User controlled) Picture Control settings in the camera, even when shooting raw. Settings such as Brightness and Contrast can affect the tonal distribution in the jpeg.
2. One of the first things that needs to be done when processing a raw file, either in-camera or with external software, is that a White Point (and Black Point) must be set. The raw data is just a set of numbers with no inherent fixed range from black through to white. Sometimes the raw file may contain a suggested figure for the raw processor to use, sometimes not. It seems to me that the in-camera white point settings used by manufacturers tend to be conservative, probably to allow for variations in the characteristics of cameras from one unit to another.
 
That's probably a bit unfair to camera manufacturers. I haven't ever seen camera manufacturers publishing DR specs for their cameras. It would be too much of a "can of worms".

I think what it boils down to is that if individuals want to get the best DR out of their camera, they need to do some simple tests themselves to get a feel for how much "highlight margin" they actually have.

Dave

Hi, Dave,
Thanks for your comments.
There is no doubt that DxO I'd furnished but one camera to test, but I am concerned, and I think, rightly, that the reported DR taken as gospel far too readily by eager camera purchasers.

And your caution concerning the user settings is well taken. Honestly I hadn't thought of that because mine are always  zeroed-out.  I'll be sure to include that caution in the instructions for the procedure.

The test series is begun shortly before the JPEG clipping occurs and continues, in 1/3 stop increments, for another three full stops. So far this has been adequate with only a couple of camera's found with two and 2/3 stops of ERADR.  This provides the greatest possible accuracy in determining the number of one-third stops of exposure found to be include the evidence of "JPEG clipping" and continuing until a tonally normalized raw image reveals actual clipped highlights ( also attested by the characteristic clipping "spike" in that image's histogram.
And yes, white points and black points are established for normalization of the test exposures.

You are right that the camera manufacturers do not publish DR specs...which, of course they can't , given the variation extant among camera's of the same model!

There is no question that many, if not, possibly, the great majority of photographers are unaware of the actual ERADR - the actual "overhead" of extra raw-accessible dynamic range available from their individual cameras. It would seem not unreasonable that the industry acknowledge the effects of "process variance" on the functioning of camera sensors and, in the user manuals,and/or via some other wildly accessible medium  clearly describe the process by which the owner of the new camera can reasonably and accurately determine their camera's  "DR overhead" and thereby more conveniently, should they wish, utilize as much as possible of their  available dynamic range and avoid the consequences of not doing so.

What are the chances of that happening? Zero if awareness of the problem is zero!  I'm glad you agree that it is the responsibility of the photographer, if desirous of capturing image data of highest possible quality, to learn how to test their camera for its allotment of extra raw-accessible dynamic range.

Thanks again, Dave, for your careful reading of my thoughts, and for  your thoughtful and helpful comments.
Very best regards,

Dave


Logged

uuglypher

  • Newbie
  • *
  • Offline Offline
  • Posts: 34
Re: Raw DR ratings based on one camera!
« Reply #8 on: May 17, 2016, 08:15:12 pm »

that probably was just a faulty testing (do you happen to have a link ?) - once you start using UniWB and make a proper OOC JPG parameters everything will be nice within <= 1/3 EV.

Hi, Alter Ego,
Thank for your comments. The thechniques for estimating the "extra raw-accessible dynamic range"( ERADR), "Stealth Dynamic Range", "Overhead" have evolved over the past ten to twelve years among those practicing ETTR /EBTR (expose to the right/ expose beyond the right). Some of us started after reading Bruce Fraser's Adobe White Paper in 2004
"Raw Capture, Linear Gamma, and Exposure"
Adobe White Paper, 2004
http://www.adobe.com/digitalimag/pdfs/linear_gamma.pdf

...which he concluded with::

"I’d be remiss if I didn’t mention one of the great strengths of Adobe® Camera Raw in this area, the extended recovery feature, which kicks in when you set the Exposure slider to negative values. Most raw converters give up on highlights once a single channel is driven to clipping, but Camera Raw does its best to reconstruct highlight retail from a single channel. Depending on the camera model and the color temperature settings, you may be able to recover as much as one stop of highlight detail, though one-third stop is more typical.
If you use Camera Raw, it’s worth spending some time conducting exposure tests to see just how far you can comfortably push the exposure."

In 2012 John Shaw posted:
http://www.johnshawphoto.com/ettr-to-the-far-right/

After more than a decade's experience with EBTR I posted a summary tutorial:
"Tutorial: Welcome to Raw Exposure"
http://www.uglyhedgehog.com/t-372364-1.html

I've heard several suggestions similar to yours that UniWB might contribute to more accurate determination of ERADR, but I have  to see substantive evidence making that case.

At the moment, using the technique used by a number of photographers for a decade or more, one can get a reasonably accurate (and replicable !) hand on their camera's allotment of ERADR and then be able, under appropriate conditions, to use their camera's maximally available DR to positive effect.

Thanks, Alter Ego,  for your thoughts and comments. Any further thoughts on the practical application of UniWB in furtherance of the aims of EBTR, and especially images in evidence of its utility, would be much appreciated.

Best regards,
Dave
Logged

AlterEgo

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1995
Re: Raw DR ratings based on one camera!
« Reply #9 on: May 18, 2016, 09:46:00 am »

Hi, Alter Ego,
Thank for your comments. The thechniques for estimating the "extra raw-accessible dynamic range"( ERADR), "Stealth Dynamic Range", "Overhead" have evolved over the past ten to twelve years among those practicing ETTR /EBTR (expose to the right/ expose beyond the right). Some of us started after reading Bruce Fraser's Adobe White Paper in 2004
"Raw Capture, Linear Gamma, and Exposure"
Adobe White Paper, 2004
http://www.adobe.com/digitalimag/pdfs/linear_gamma.pdf

...which he concluded with::

"I’d be remiss if I didn’t mention one of the great strengths of Adobe® Camera Raw in this area, the extended recovery feature, which kicks in when you set the Exposure slider to negative values. Most raw converters give up on highlights once a single channel is driven to clipping, but Camera Raw does its best to reconstruct highlight retail from a single channel. Depending on the camera model and the color temperature settings, you may be able to recover as much as one stop of highlight detail, though one-third stop is more typical.
If you use Camera Raw, it’s worth spending some time conducting exposure tests to see just how far you can comfortably push the exposure."

In 2012 John Shaw posted:
http://www.johnshawphoto.com/ettr-to-the-far-right/

After more than a decade's experience with EBTR I posted a summary tutorial:
"Tutorial: Welcome to Raw Exposure"
http://www.uglyhedgehog.com/t-372364-1.html

I've heard several suggestions similar to yours that UniWB might contribute to more accurate determination of ERADR, but I have  to see substantive evidence making that case.

At the moment, using the technique used by a number of photographers for a decade or more, one can get a reasonably accurate (and replicable !) hand on their camera's allotment of ERADR and then be able, under appropriate conditions, to use their camera's maximally available DR to positive effect.

Thanks, Alter Ego,  for your thoughts and comments. Any further thoughts on the practical application of UniWB in furtherance of the aims of EBTR, and especially images in evidence of its utility, would be much appreciated.

Best regards,
Dave

I am using the following settings since Aug 2015 for Sony A7R2 for OOC JPG to have proper zebra/blinkies indication for clipping in raw :

WB = UniWB
DRO/Auto HDR = D-Range Opt./DRO Lv5
Creative Style = Neutral (contrast -3, saturation -3, sharpness +3)
Color Space = sRGB
Zebra = 100+

or the just minimal set :

WB = UniWB
Creative Style = Neutral (contrast -3, saturation -3, ...)
Zebra = 100+

which were based on the settings that I was using for Sony A7 since Feb 2014, except Creative Style for A7 in my case was Portrait (contrast -3, saturation -3, ...)

I find that this gives me <= 1/3 EV precision comparing blinkies with what rawdigger shows in raw... UniWB was/is the key element... that simple
Logged

uuglypher

  • Newbie
  • *
  • Offline Offline
  • Posts: 34
Re: Raw DR ratings based on one camera!
« Reply #10 on: May 18, 2016, 02:47:18 pm »

I am using the following settings since Aug 2015 for Sony A7R2 for OOC JPG to have proper zebra/blinkies indication for clipping in raw :

WB = UniWB
DRO/Auto HDR = D-Range Opt./DRO Lv5
Creative Style = Neutral (contrast -3, saturation -3, sharpness +3)
Color Space = sRGB
Zebra = 100+

or the just minimal set :

WB = UniWB
Creative Style = Neutral (contrast -3, saturation -3, ...)
Zebra = 100+

which were based on the settings that I was using for Sony A7 since Feb 2014, except Creative Style for A7 in my case was Portrait (contrast -3, saturation -3, ...)

I find that this gives me <= 1/3 EV precision comparing blinkies with what rawdigger shows in raw... UniWB was/is the key
element... that simple

Thanks very much, A.E., for your information. I am, of course, curious as to the amount of additional raw-accessible EV.s you've documented in your A7R2.
Thanks again,
Dave
Logged

AlterEgo

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1995
Re: Raw DR ratings based on one camera!
« Reply #11 on: May 18, 2016, 03:05:44 pm »

I am, of course, curious as to the amount of additional raw-accessible EV.s you've documented in your A7R2.

as I noted <= 1/3 EV comparing camera indication (which is based on OOC JPG) vs what rawdigger shows in raw (true raw channels details)...  UniWB makes that quite immune to scene illumination light spectrum, that's is the whole point of it.
Logged

Bart_van_der_Wolf

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 8913
Re: Raw DR ratings based on one camera!
« Reply #12 on: May 18, 2016, 03:27:05 pm »

as I noted <= 1/3 EV comparing camera indication (which is based on OOC JPG) vs what rawdigger shows in raw (true raw channels details)...  UniWB makes that quite immune to scene illumination light spectrum, that's is the whole point of it.

Hi,

I get the impression that Dave is referring to the 'recovered' from '1 or 2 channel clipped' extrapolated 'data' in the highlights.

Of course that wouldn't be counted as real DR, but as extrapolated range, and the amount that e.g. LR achieves is highly dependent on the WhiteBalance as well.

For example, a clipped Red channel of a Red tulip will be hard to 'recover' from the Green channel which is much lower and from a different color, and Blue is probably virtually non-existent. So extrapolating (desaturated) Red from Green and Blue would cause ugly (high)lights, and the luminance detail is probably quite wrong for Red. A colorbalance that reduces Green and Blue makes matters even worse.

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

AlterEgo

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1995
Re: Raw DR ratings based on one camera!
« Reply #13 on: May 18, 2016, 03:55:10 pm »

I get the impression that Dave is referring to the 'recovered' from '1 or 2 channel clipped' extrapolated 'data' in the highlights.

but that will be also 1) subject dep. (as noted already) 2) raw converter & camera profile dep. and then 3) what is the criteria of "being recovered" exactly ? too many variables...

Logged

uuglypher

  • Newbie
  • *
  • Offline Offline
  • Posts: 34
Re: Raw DR ratings based on one camera!
« Reply #14 on: May 18, 2016, 04:57:50 pm »

Hi,

I get the impression that Dave is referring to the 'recovered' from '1 or 2 channel clipped' extrapolated 'data' in the highlights.

Of course that wouldn't be counted as real DR, but as extrapolated range, and the amount that e.g. LR achieves is highly dependent on the WhiteBalance as well.

For example, a clipped Red channel of a Red tulip will be hard to 'recover' from the Green channel which is much lower and from a different color, and Blue is probably virtually non-existent. So extrapolating (desaturated) Red from Green and Blue would cause ugly (high)lights, and the luminance detail is probably quite wrong for Red. A colorbalance that reduces Green and Blue makes matters even worse.

Cheers,
Bart

Hi, Bart,
Thanks for your comment.
Actually, what I'm looking for with the EBTR technique is real, grade-A, certified extra raw-accessible dynamic range that exists "beyond the right" side of the JPEG-adjusted histogram frame (and its clipping warning) that the manufacturer's algorithms have dictated as the limit of the DR available for JPEG file exposure.  Now I'm not talking about when we get to the outer ragged edge where A.C.R. is doing its damnedest to rescue image data from two or even from one remnant color channel. Before y'get to that point there is invariably some useful FULL-fledged DR..maybe only 1/3 Stop,but maybe as much as two and 2/3 stops or more...(the majority have at least one full stop) that is just sitting there .... unrecognized, unheralded, and un-used.

The important point is that it's DR able to be used, and ought be recognized as a component of the DR accessible for raw data capture. And as such, we need to recognize that DR that is NOT used by the brightest stop of the chosen exposure, exacts a price for not being used...for being wasted...in terms of forfeited tonal spectra (brightness and chromatic) ...quantitatively dictated by the file's bit-depth.

The fact that the makers of light sensors haven't worked out a way to predictably  limit their delivered dynamic range is likely a major reason that raw histograms are not routinely provided in our cameras.  At least that seems reasonable to me. If there's a good argument to the contrary, I'd love to hear it.

So, for the time being, it seems (to me and to some like-minded folk...) altogether meet and proper that photographers hoping to capture the highest quality image data make every effort to determine their camera's allotment of ERADR...and then, when appropriate...to use it!

Best regards,
Dave
Logged

uuglypher

  • Newbie
  • *
  • Offline Offline
  • Posts: 34
Re: Raw DR ratings based on one camera!
« Reply #15 on: May 19, 2016, 03:34:45 pm »

actually it would be wise to remember that not all "highlights" are actually important and some actually might be OK to clip in 1, 2, 3, ... raw channels in return for a better S/N in more important parts of the image.
Hi, A.E.
I had meant to respond to this comment earlier, and simply forgot. I agree whole-heartedly with your point that some blown highlights are not, in fact crucial to some images, and the extra exposure that clipped them also provides additional exposure to the darker regions of the image file to assure, depending on the bit-depth, an improved tonal spectrum and consequently better-defined shadow detail.
Point well-made!

Best regards,
Dave
Logged

bjanes

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3387
Re: Raw DR ratings based on one camera!
« Reply #16 on: May 19, 2016, 05:54:56 pm »

I have been informed, in direct response to my question posed to DxOMark “How many individual cameras of a given make and model are tested for Dynamic range?” that the answer is “one camera”.  DxOMark tests raw capture DR at different ISOs and reports the highest DR observed, with no mention of the associated ISO.

There is quite a bit sample variation between lenses, but I have seen little evidence of sample variation with sensors. Those who have determined sensor characteristics using standard methods such as described by Roger Clark here, Bill Claff, and others show very similar results for the same camera. Indeed, cameras with closely related Sony sensors show consistent results. Testing one camera is likely sufficient.

Does anyone have information about variation among sensors?

Regards,

Bill
Logged

uuglypher

  • Newbie
  • *
  • Offline Offline
  • Posts: 34
Re: Raw DR ratings based on one camera!
« Reply #17 on: May 19, 2016, 09:10:30 pm »

There is quite a bit sample variation between lenses, but I have seen little evidence of sample variation with sensors. Those who have determined sensor characteristics using standard methods such as described by Roger Clark here, Bill Claff, and others show very similar results for the same camera. Indeed, cameras with closely related Sony sensors show consistent results. Testing one camera is likely sufficient.

Does anyone have information about variation among sensors?

Regards,

Bill

Hi, Bill,
Those of us who have been practicing first, ETTR, and then EBTR since reading Bruce Fraser's Adobe White Paper in 2004 and taking to heart his urging to experiment to discover the amount of dynamic range hat exists beyond the JPEG clipping warnings have been pleased to find that essentially every digital camera we've tested has provided from 1/3 stop to Two and two-thirds stops ( and, according to reports of others three full stops and more) beyond the "clipping" of JPEGS.

On occasion some of us have had the chance to test more than one camera of the same model and been quite surprised to find that the amounts of extra raw-accessible dynamic range (ERADR) to vary significantly among individual cameras/sensors of the same model.  It seemed highly unlikely that such a supposed precision semiconductor device should actually have such a range of "production" - deliverance of such a range of dynamic ranges. But then, after some study on the history of "process variance" of silicon-based semiconductor devices, the likely cause seems obvious.   That is the basis upon which we recommend that each individual camera be tested, rather than rely upon a dynamic range determined by testing a single camera of a given model. Our reliance in the commonly used testing procedure is reinforced by the remarkable replicability of the results of repeated tests on the same individual cameras at a range of ISOs.  A number of "doubting Thomases" -and I'd have to count myself as one at the start -  have become adherents of the EBTR technique after actually giving it a sincerely fair trial.

And by the bye, Roger Clark's graphed data on dynamic range (with which I am well familiar) does not, a far as I can find, duplicate any camera models.

My best recommendation is for anyone with doubts about the efficacy of EBTR and about the supposed consistent performance of sensors in a single model of camera to actually run the tests themselves, and, until they do, remain skeptical about the supposed "invariance of performance" of our cameras light sensors.

Thanks for your serious attention to this topic, Bill, and for your thoughtful comments.

Best regards,
Dave
Logged

Guillermo Luijk

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2005
    • http://www.guillermoluijk.com
Re: Raw DR ratings based on one camera!
« Reply #18 on: May 20, 2016, 12:49:15 pm »

I have been informed, in direct response to my question posed to DxOMark “How many individual cameras of a given make and model are tested for Dynamic range?” that the answer is “one camera”.

DxOMark tests all cameras (Fuji X-Trans aside), that includes a lot of cameras with very similiar if not equal photocells (e.g. Canon 7DII and Canon 5DS), and the results have been very consistent so far. So the "one camera" database they have developed can itself be used to conclude that sensor variance is not wide enough to make any more tests necessary.

Regards

uuglypher

  • Newbie
  • *
  • Offline Offline
  • Posts: 34
Re: Raw DR ratings based on one camera!
« Reply #19 on: May 20, 2016, 03:20:31 pm »

DxOMark tests all cameras (Fuji X-Trans aside), that includes a lot of cameras with very similiar if not equal photocells (e.g. Canon 7DII and Canon 5DS), and the results have been very consistent so far. So the "one camera" database they have developed can itself be used to conclude that sensor variance is not wide enough to make any more tests necessary.

Regards

Hi, Guillermo,
The sentiments you expressed are exactly the reason that we were surprised to find variance in the extra raw-accessible dynamic range (ERADR) beyond the clipping point for the JPEG-adjusted histogram frame among camera's of the same model.
That repeated tests for ERADR at different ISOs on a variety of camera models have yielded consistent results (to 1/3 stop) has reinforced reliance in the test procedure. We are not, in the manner that we test, determining the actual DR of any tested camera, simply the "extra raw-accessible dynamic range" which is then added to the ETTR exposure in an effort to utilize the maximal available dynamic range of each of our cameras. The fact remains that different cameras of the same model have yielded different numbers of 1/3 stops of ERADR. On that basis has rested the recommendation that each camera be tested for its individual allotment of ERADR. Perhaps you can explain how  different amounts of determined ERADR among cameras of the same model might not, in fact, conflict with the concept of the accuracy of a single  total DR determination.  I've been in contact with DxO Mark but have not received helpful information in this regard nor been able to engage in any substantive discussion on the matter.

It would seem most informative if someone without a "dog in this fight" would undertake similar tests for ERADR  in several cameras of each of the same models and report those results.
I appreciate your interest and would appreciate any further thoughts you may have.

Best regards,
Dave
Logged
Pages: [1] 2 3   Go Up