Dxomark "measured" ISO vs Manufacturer ISO

[Allan Olesen]

This is a question to both Endos and ET2:


Are you being deliberately ambiguous to stir up some opposition and keep the discussion going?

You are discussing two different claims which do not necessarily contradict each other, and it has been going on for ever now, probably only because you want it to go on. Otherwise you would have made an effort to remove the confusion.

It is like watching this kind of conversation:

ET2: This car is red.

Endos: No, it has four wheels.

ET2: Are you stupid? It is clearly red.

Endos: No, look: Two wheels on the right side and two wheels on the left side. That is four.

ET2: No, I have measured the wavelength of the light emitted from the car. It is 690 nm. That is clearly red.

And on and on and on and...

 

[ET2]

Allan Olesen wrote:

This is a question to both Endos and ET2:

Are you being deliberately ambiguous to stir up some opposition and keep the discussion going?

You are discussing two different claims which do not necessarily contradict each other, and it has been going on for ever now, probably only because you want it to go on.

That's not true. Endos originally was claiming that we will see lower shutter speeds on A99. He based that claim on Dxomark "measured" ISOs. After it became clear that that claim isn't true (especially after X100 example in this thread), he just started debating a different topic; that is "Raws" are underexpose anyway, even though the camera behaves normally. Whatever happens under the hood is totally invisible to the user. As far as the end users are concerned, they gets 2-stop faster shutter speed at ISO 6400 (compared to ISO 1600) -- just as expected.

We need to look at DPR's "measured" ISOs that are based of a calibrated light meter.

For A99

"we simply compare the exposure for each shot to the metered light level (using a calibrated Sekonic L-358), middle gray matched. We estimate the accuracy of these results to be +/- 1/6 EV (the margin of error given in the ISO specifications). We found that measured ISO from the Sony SLT-A99 is within +/- 1/6 EV of indicated ISO across its entire ISO range."

For D600

"In our tests we found that measured ISOs from the Nikon D600 match the marked ISOs within 1/6 stop accuracy, meaning ISO 100 indicated = ISO 100 measured."

For 5D MArk III

"In our tests we found that measured ISOs from the Canon EOS 5D Mark III match the marked ISOs within 1/6 stop accuracy, meaning ISO 100 indicated = ISO 100 measured."

This should have ended this debate a long time.

 

[jonas ar]

TrojMacReady wrote:

jonas ar wrote:

TrojMacReady wrote:

Visible image exposure will not be underexposed by the suggested amounts, as also explained by DXOMark in their X100 review for example:

"On the surface this is odd, but perfectly explainable:

What looks odd according to the reference is the following observation: a selected ISO 1600 corresponds to a simple ISO 1000. This measured ISO remains constant for selected ISOs 3200 and 6400.

So far as I know Bill Claff and Marianne Oelund have measured the use of variable analog gain in D800 and determined that the gain is constant above ISO1250 or 2000 depending on the meaurements used. However, the DxO measured ISO vs manufactorer ISO is linear all the way to the highest ISO

Why does the curve not flatten?

Cheers,

Jonas

There are 2 types of digital gain as also briefly mentioned and explained by Guillermo here.

The X100 relies on digital gain as applied in the RAW converter or jpeg engine after the fact (curves). The D800 (and many other Exmor cameras AFAIK) rely on digital gain before writing the RAW file by simply mutiplying values. From what I gathered in the posts from Marianne back then, that's the type of digital gain that she detected. Sadly not one mentioned (or detected) and explained by DXOmark.

Thanks for that!


It makes one wonder even more why Nikon an others implement unnecessary highlight clipping in raw

 

[Endos]

Cudacke wrote:

Endos wrote:

Yes, but also for JPEGs, TIFFs, etc. you need a software that knows his structure, is able to decode it for show it to you. What about DNG files? They are RAW files with a well known structure.

Again. Please stop making non sense...
any kind of data including the analog one would need the reader to know its structure to understand it.

However, Image doesn't exist in RAW is because RAW is the untranslated electronic signal of recorded light intensity. It is not an image. Its data are electronic signal. It, however, can be translated to a image if you know the structure of it and what target audience to do the specific processing required to turn the data into a viewable image representation.

JPEG on the other hand is a specific image format that is designed to store a human viewable image data.

If you don't know the difference between the two, do your reading first.

Seems that you have reading compression issues. Can you please enlighten me and tell me what is false in what I wrote?

 

[Endos]

Cudacke wrote:

Endos wrote:

Cudacke wrote:

DXO clearly says it is not cheating because because before you reach the final out put image stage, the "Manufacture ISO" is not even exist yet in the capture.

I never said there is any cheating.

You did as following and now you lied.

I never said that. Quote where I said that please.

Please stop and go study to get your knowledge right.

I suggest you the same.

 

[Endos]

ET2 wrote:

Allan Olesen wrote:

This is a question to both Endos and ET2:

Are you being deliberately ambiguous to stir up some opposition and keep the discussion going?

You are discussing two different claims which do not necessarily contradict each other, and it has been going on for ever now, probably only because you want it to go on.

That's not true. Endos originally was claiming that we will see lower shutter speeds on A99.

That's not true. I never said that. Can you please quote where I said this?

 

[Endos]

Allan Olesen wrote:

This is a question to both Endos and ET2:

Are you being deliberately ambiguous to stir up some opposition and keep the discussion going?

You are discussing two different claims which do not necessarily contradict each other, and it has been going on for ever now, probably only because you want it to go on. Otherwise you would have made an effort to remove the confusion.

It is like watching this kind of conversation:

ET2: This car is red.

Endos: No, it has four wheels.

ET2: Are you stupid? It is clearly red.

Endos: No, look: Two wheels on the right side and two wheels on the left side. That is four.

ET2: No, I have measured the wavelength of the light emitted from the car. It is 690 nm. That is clearly red.

And on and on and on and...

You are right. The problem with ET2 is that he thinks I said something I never said.

 

[mike_2008]

I thank the OP for the thread, I read the whole thing and learned something new about iso in dxomark.

I don't agree at all that ET2 is being ambiguous, I think he is being as forthright as its possible to be on a complex and confusing subject. Endos on the other hand appears to enjoy moving the goalposts (perhaps even unintentionally) by being consistently unclear about whether he is discussing final image exposures or digital numbers in a raw file.

It comes down to this. DXOMARK iso numbers don't mean anything at all with regards to the output image exposure for a given shutter speed, iso, aperture and luminosity. You can take any digital numbers and scale them to what you want in raw conversion or another stage. Indeed, the whole concept of 'iso' in digital photography is actually rather inappropriate, it would be simpler to speak of gain, which can be analogue or digital.

Anyway, what matters to us as photographers is whether the output images have an appropriate and similar exposure across different cameras. There should be a standard output for a given iso (gain), luminosity, aperture and exposure time. Interestingly there are different scales which are in the same standard, see here:

Film speed - Wikipedia

en.wikipedia.org

One of these is the "Standard Output Sensitivity" technique, which I guess is that used in our sRGB DSLRs.

"the Standard Output Sensitivity (SOS) technique, also new in the 2006 version of the standard, effectively specifies that the average level in the sRGB image must be 18% gray plus or minus 1/3 stop when the exposure is controlled by an automatic exposure control system calibrated per ISO 2721 and set to the EI with no exposure compensation. Because the output level is measured in the sRGB output from the camera, it is only applicable to sRGB images—typically JPEG—and not to output files in raw image format. It is not applicable when multi-zone metering is used."

Note carefully that the SOS iso standard refers only to data in sRGB output, not raw files. This is why endos has been tying himself in knots.

A final comment, my congratulations to TrojMacReady for humiliating the irritating poster 'just having fun' who's only object here appears to be to lie and insult sony users. Why the mods permit this is beyond me. The specific point I refer to is jhf's repeated assertion that it was 'obvious' in the D600-A99 comparison at iso3200 that there was nearly a full stop difference in final image exposures, which Troj dispelled by simply showing that in reality there was only approximately 10% difference in final image exposure. Obviously it was too much to expect jhf to admit his egregious mistake and apologise. Thanks to troj for showing who's wearing clothes, and who isn't!

 

[Endos]

mike_2008 wrote:

I thank the OP for the thread, I read the whole thing and learned something new about iso in dxomark.

I don't agree at all that ET2 is being ambiguous, I think he is being as forthright as its possible to be on a complex and confusing subject. Endos on the other hand appears to enjoy moving the goalposts (perhaps even unintentionally) by being consistently unclear about whether he is discussing final image exposures or digital numbers in a raw file.

I think I was clear, but as English is my third language maybe I wasn't. I'm talking all the time about RAW files as they come from the camera, before any RAW converter applies tonal curves, amplification gains, and so on, as they are tested by DxO .

 

[EinsteinsGhost]

Endos wrote:

Cudacke wrote:

Endos wrote:

Yes, but also for JPEGs, TIFFs, etc. you need a software that knows his structure, is able to decode it for show it to you. What about DNG files? They are RAW files with a well known structure.

Again. Please stop making non sense...
any kind of data including the analog one would need the reader to know its structure to understand it.

However, Image doesn't exist in RAW is because RAW is the untranslated electronic signal of recorded light intensity. It is not an image. Its data are electronic signal. It, however, can be translated to a image if you know the structure of it and what target audience to do the specific processing required to turn the data into a viewable image representation.

JPEG on the other hand is a specific image format that is designed to store a human viewable image data.

If you don't know the difference between the two, do your reading first.

Seems that you have reading compression issues. Can you please enlighten me and tell me what is false in what I wrote?

--

Regards,
Juan
http://500px.com/endosphoto
http://www.facebook.com/endosphotography

Where would you like me to begin?

 

[TrojMacReady]

Endos wrote:

TrojMacReady wrote:

Endos wrote:

TrojMacReady wrote:

Endos wrote:

TrojMacReady wrote:

Endos wrote:

EinsteinsGhost wrote:

Endos wrote:

Cudacke wrote:

Endos wrote:

ET2 wrote:

Endos wrote:

ET2 wrote:

Endos wrote:

Yes but with underexposed results. Exactly the same if you shoot with that speeds at ISO 1600.

I don't think anyone has been so thoroughly refuted on a topic in these forums, as you have.

There is at least one, you.

Your initial claim was that A99 is "cheating" with ISOs and would need slower shutter speeds than A900.

That's right. To get the same "exposed" RAW with an A99 you need longer exposures.

Both DXO and other people had explained this, please read and learn the knowledge before you start talking nonsense.

News

www.dxomark.com

http://theonlinephotographer.typepad.com/the_online_photographer/2012/09/why-iso-isnt-iso.html
Numerical

Are you kidding? Both links says nothing that disagree with what I'm saying here.

Please try to read and learn the knowledge (from the same authors):

http://www.dxomark.com/index.php/Publications/DxOMark-Insights/Pushed-ISO-Let-s-make-it-clear

http://theonlinephotographer.typepad.com/the_online_photographer/2012/10/raw-is-not-raw.html

 

[Cudacke]

Endos wrote:

Cudacke wrote:

Endos wrote:

Yes, but also for JPEGs, TIFFs, etc. you need a software that knows his structure, is able to decode it for show it to you. What about DNG files? They are RAW files with a well known structure.

Again. Please stop making non sense...
any kind of data including the analog one would need the reader to know its structure to understand it.

However, Image doesn't exist in RAW is because RAW is the untranslated electronic signal of recorded light intensity. It is not an image. Its data are electronic signal. It, however, can be translated to a image if you know the structure of it and what target audience to do the specific processing required to turn the data into a viewable image representation.

JPEG on the other hand is a specific image format that is designed to store a human viewable image data.

If you don't know the difference between the two, do your reading first.

Seems that you have reading compression issues. Can you please enlighten me and tell me what is false in what I wrote?

--

Regards,
Juan
http://500px.com/endosphoto
http://www.facebook.com/endosphotography

I did, but you didn't read. Sorry I cannot help on your reading problem. It is your job to read. I can only point it out you didn't do.

 

[Cudacke]

Endos wrote:

Cudacke wrote:

Endos wrote:

Cudacke wrote:

DXO clearly says it is not cheating because because before you reach the final out put image stage, the "Manufacture ISO" is not even exist yet in the capture.

I never said there is any cheating.

You did as following and now you lied.

I never said that. Quote where I said that please.

Please stop and go study to get your knowledge right.

I suggest you the same.

 

[Cudacke]

Allan Olesen wrote:

This is a question to both Endos and ET2:

Are you being deliberately ambiguous to stir up some opposition and keep the discussion going?

You are discussing two different claims which do not necessarily contradict each other, and it has been going on for ever now, probably only because you want it to go on. Otherwise you would have made an effort to remove the confusion.

It is like watching this kind of conversation:

ET2: This car is red.

Endos: No, it has four wheels.

ET2: Are you stupid? It is clearly red.

Endos: No, look: Two wheels on the right side and two wheels on the left side. That is four.

ET2: No, I have measured the wavelength of the light emitted from the car. It is 690 nm. That is clearly red.

And on and on and on and...

Haha well put!

 

[Cudacke]

Endos wrote:

mike_2008 wrote:

I thank the OP for the thread, I read the whole thing and learned something new about iso in dxomark.

I don't agree at all that ET2 is being ambiguous, I think he is being as forthright as its possible to be on a complex and confusing subject. Endos on the other hand appears to enjoy moving the goalposts (perhaps even unintentionally) by being consistently unclear about whether he is discussing final image exposures or digital numbers in a raw file.

I think I was clear, but as English is my third language maybe I wasn't. I'm talking all the time about RAW files as they come from the camera, before any RAW converter applies tonal curves, amplification gains, and so on, as they are tested by DxO .

 

[Endos]

TrojMacReady wrote:

Endos wrote:

TrojMacReady wrote:

Endos wrote:

TrojMacReady wrote:

Endos wrote:

TrojMacReady wrote:

Endos wrote:

EinsteinsGhost wrote:

Endos wrote:

Cudacke wrote:

Endos wrote:

ET2 wrote:

Endos wrote:

ET2 wrote:

Endos wrote:

Yes but with underexposed results. Exactly the same if you shoot with that speeds at ISO 1600.

I don't think anyone has been so thoroughly refuted on a topic in these forums, as you have.

There is at least one, you.

Your initial claim was that A99 is "cheating" with ISOs and would need slower shutter speeds than A900.

That's right. To get the same "exposed" RAW with an A99 you need longer exposures.

Both DXO and other people had explained this, please read and learn the knowledge before you start talking nonsense.

News

www.dxomark.com

http://theonlinephotographer.typepad.com/the_online_photographer/2012/09/why-iso-isnt-iso.html
Numerical

Are you kidding? Both links says nothing that disagree with what I'm saying here.

Please try to read and learn the knowledge (from the same authors):

http://www.dxomark.com/index.php/Publications/DxOMark-Insights/Pushed-ISO-Let-s-make-it-clear


http://theonlinephotographer.typepad.com/the_online_photographer/2012/10/raw-is-not-raw.html

 

[Endos]

Cudacke wrote:

Endos wrote:

Cudacke wrote:

Endos wrote:

Yes, but also for JPEGs, TIFFs, etc. you need a software that knows his structure, is able to decode it for show it to you. What about DNG files? They are RAW files with a well known structure.

Again. Please stop making non sense...
any kind of data including the analog one would need the reader to know its structure to understand it.

However, Image doesn't exist in RAW is because RAW is the untranslated electronic signal of recorded light intensity. It is not an image. Its data are electronic signal. It, however, can be translated to a image if you know the structure of it and what target audience to do the specific processing required to turn the data into a viewable image representation.

JPEG on the other hand is a specific image format that is designed to store a human viewable image data.

If you don't know the difference between the two, do your reading first.

Seems that you have reading compression issues. Can you please enlighten me and tell me what is false in what I wrote?

--

Regards,
Juan
http://500px.com/endosphoto
http://www.facebook.com/endosphotography

I did, but you didn't read. Sorry I cannot help on your reading problem. It is your job to read. I can only point it out you didn't do.

Well, when you find anything I wrote that is not true, feel free to quote them here. I will be happy to read your conclusions.

 

[EinsteinsGhost]

Endos wrote:

Cudacke wrote:

I did, but you didn't read. Sorry I cannot help on your reading problem. It is your job to read. I can only point it out you didn't do.

Well, when you find anything I wrote that is not true, feel free to quote them here. I will be happy to read your conclusions.

It is rather difficult to figure out where to start. If you don't mind, I chose this gem of yours:

Endos wrote:

ET2 wrote:
Let's see X100's F-Stop and Shutter speeds in imaging-resources studio shots
ISO 1600 - F8 - 1/250

ISO 3200- F8 - 1/500

ISO 6400 - F8 - 1 / 1000

So much for the "theory" that dxomark "measured" ISO mean anything in real life. They don't. Fuji X100 behaves perfectly. You get exactly one stop faster shutter speeds from ISO 1600 to 3200 and again from ISO 3200 to 6400.

Yes but with underexposed results. Exactly the same if you shoot with that speeds at ISO 1600.

Now two questions:

1-Why would image at ISO 6400 above be underexposed compared to ISO 1600? Is it because, per your belief, the real ISO is 1000?

2-A hypothetical question: If DXO measured ISO is higher than manufacturer ISO, is it a good thing, or bad? Either way, explain your choice.

 

[Endos]

Cudacke wrote:

Endos wrote:

mike_2008 wrote:

I thank the OP for the thread, I read the whole thing and learned something new about iso in dxomark.

I don't agree at all that ET2 is being ambiguous, I think he is being as forthright as its possible to be on a complex and confusing subject. Endos on the other hand appears to enjoy moving the goalposts (perhaps even unintentionally) by being consistently unclear about whether he is discussing final image exposures or digital numbers in a raw file.

I think I was clear, but as English is my third language maybe I wasn't. I'm talking all the time about RAW files as they come from the camera, before any RAW converter applies tonal curves, amplification gains, and so on, as they are tested by DxO .

 

[Endos]

EinsteinsGhost wrote:

Endos wrote:

Cudacke wrote:

I did, but you didn't read. Sorry I cannot help on your reading problem. It is your job to read. I can only point it out you didn't do.

Well, when you find anything I wrote that is not true, feel free to quote them here. I will be happy to read your conclusions.

It is rather difficult to figure out where to start. If you don't mind, I chose this gem of yours:

Endos wrote:

ET2 wrote:
Let's see X100's F-Stop and Shutter speeds in imaging-resources studio shots
ISO 1600 - F8 - 1/250

ISO 3200- F8 - 1/500

ISO 6400 - F8 - 1 / 1000

So much for the "theory" that dxomark "measured" ISO mean anything in real life. They don't. Fuji X100 behaves perfectly. You get exactly one stop faster shutter speeds from ISO 1600 to 3200 and again from ISO 3200 to 6400.

Yes but with underexposed results. Exactly the same if you shoot with that speeds at ISO 1600.

Now two questions:

1-Why would image at ISO 6400 above be underexposed compared to ISO 1600? Is it because, per your belief, the real ISO is 1000?

The Fuji X100 sensor doesn't support ISO6400. His maximum ISO is 1600 (measured ISO 1000). Then in reality when you choose ISO 6400 - F8 - 1 / 1000 the camera uses ISO 1600 - F8 - 1 / 1000 The result is obviously 2 stops underexposed:

ISO 1600

ISO 6400. 2 stops underexposed compared with the ISO 1600 image.

You can download the RAW files from imaging-resource and with any RAW analyzer test it for yourself (don't use a RAW converter, they push the exposure to compensate).

2-A hypothetical question: If DXO measured ISO is higher than manufacturer ISO, is it a good thing, or bad? Either way, explain your choice.

That question has nothing to do with the discussion, but well, I think it's a bad thing because having clipped highlights with the "default exposure" is not a good idea.


--
Regards,
Juan

500px

500px.com

Log into Facebook

Log into Facebook to start sharing and connecting with your friends, family, and people you know.

www.facebook.com