GFX 100RF ISO 80 black point subtraction

JimKasson said:

Truman Prevatt said:

JimKasson said:

Truman Prevatt said:

eliehbk said:

So which one to use to have the most dynamic range and amount of noise?

Click to expand...

I second that question. P2P shows additional about 1/3 a stops at ISO 80. On the other hand the histograms shown point to about 1/3 stop shift toward highlight end as a result of the metering. So the question seems to be realizing that and using a -1/3 EC to shift the exposure down by 1/3 of a stop to compensate from the metering and the way the histogram is calculated at ISO 80 result the in slight amount of DR shown in Photons2Photos?

Click to expand...

Which DR metric are you referring to? Bill gives you metrics that allow at least two.

JimKasson said:

As an aside on the 100RF I modulate the ISO with the front wheel, keeping it as low as possible while meeting my requirement for shutter speed and aperture. I don't avoid 80 and I often find Fuji is not quite as aggressive as I would like in protecting highlights and often use -1/3 EC at all ISO's. I actually would like Fuji to implement a second metering mode similar to that on my Z8 with a little more highlight protection. They don't, I do it myself. I've not notice any negative impact on shooting ISO 80. I might be missing something or might not have run into a scene where there would be an impact.

So while charts and histograms are nice, question arise when two different sources show charts that point to what seems to be contradictory conclusions. What does it mean using the camera in the wild?

Click to expand...

Can you be explicit and detailed about the contradictions you see? Bill and I are generally pretty much in agreement.

Click to expand...

The photonstophotos chart givens a PDR of 12.55 at ISO 80 and 12.29 at ISO 100, or a little less than 1/3 stop more DR for that sensor.

Click to expand...

So your chosen metric for DR is PDR? That's a pretty good one.

JimKasson said:

The histograms shown in one of the post shows a slight shift toward the highlight ( a little more prone to clip) at ISO80 than ISO100.

Click to expand...

That is not taken into account in photon transfer functions like the ones that Bill and I calculate.

JimKasson said:

The shift in the histogram can be eliminated by the good old Fuji -1/3 EC which can be common. is that an artifact in the testing protocol of photonstophotos?

Click to expand...

See above.

JimKasson said:

So my question is does the ISO 80 produce slightly more DR or is that an artifact in the testing protocol of photonstophotos?

Click to expand...

It does, but it does it with a (cheap, IMO) trick -- jettisoning half the read noise -- that is not reflective of a read noise difference after demosaicing and black point compensation in the raw converter.

JimKasson said:

If this shift is understood and calibrated for in the metering and calculating the exposure - is there a reason to avoid ISO 80. if there DR is the same at 100 and 80, then that would be such a reason.

Click to expand...

The FWC is higher at ISO 80, so that could be reason to use it if you are OK with what it does to the black point.

Click to expand...

SrMi said:

Similarly, SL2 and Q2 cameras have higher FWC at ISO 50 than at ISO 100 (per bclaff), but most users avoid it because the histogram and clipping behave strangely.

Click to expand...

Leica Q2 Monochrom highlights at ISO 100 — crops - the last word

This is the 15th post in a series about the Leica Q2 Monochrom. You can see all the other posts in the series by looking in the Category List drop-down menu on the right side of the page. In the last post, I discovered some truly weird behavior in the highlights of ISO 100 images

blog.kasson.com

Leica Q2 Monochrom highlights at ISO 100 — graphs - the last word

This is the 14th post in a series about the Leica Q2 Monochrom. You can see all the other posts in the series by looking in the Category List drop-down menu on the right side of the page. I've been trying to get some quantitative results in my Leica Q2M ISO 100 highlight testing, and

blog.kasson.com

Leica Q2 Monochrom highlight linearity at ISO 100 - the last word

This is the tenth post in a series about the Leica Q2 Monochrom. You can see all the other posts in the series by looking in the Category List drop-down menu on the right side of the page. When I looked at the engineering dynamic range (EDR) of the Q2 Monochrom, I stumbled onto the

blog.kasson.com

JimKasson said:

Truman Prevatt said:

JimKasson said:

Truman Prevatt said:

eliehbk said:

So which one to use to have the most dynamic range and amount of noise?

Click to expand...

I second that question. P2P shows additional about 1/3 a stops at ISO 80. On the other hand the histograms shown point to about 1/3 stop shift toward highlight end as a result of the metering. So the question seems to be realizing that and using a -1/3 EC to shift the exposure down by 1/3 of a stop to compensate from the metering and the way the histogram is calculated at ISO 80 result the in slight amount of DR shown in Photons2Photos?

Click to expand...

Which DR metric are you referring to? Bill gives you metrics that allow at least two.

JimKasson said:

As an aside on the 100RF I modulate the ISO with the front wheel, keeping it as low as possible while meeting my requirement for shutter speed and aperture. I don't avoid 80 and I often find Fuji is not quite as aggressive as I would like in protecting highlights and often use -1/3 EC at all ISO's. I actually would like Fuji to implement a second metering mode similar to that on my Z8 with a little more highlight protection. They don't, I do it myself. I've not notice any negative impact on shooting ISO 80. I might be missing something or might not have run into a scene where there would be an impact.

So while charts and histograms are nice, question arise when two different sources show charts that point to what seems to be contradictory conclusions. What does it mean using the camera in the wild?

Click to expand...

Can you be explicit and detailed about the contradictions you see? Bill and I are generally pretty much in agreement.

Click to expand...

The photonstophotos chart givens a PDR of 12.55 at ISO 80 and 12.29 at ISO 100, or a little less than 1/3 stop more DR for that sensor.

Click to expand...

So your chosen metric for DR is PDR? That's a pretty good one.

JimKasson said:

The histograms shown in one of the post shows a slight shift toward the highlight ( a little more prone to clip) at ISO80 than ISO100.

Click to expand...

That is not taken into account in photon transfer functions like the ones that Bill and I calculate.

JimKasson said:

The shift in the histogram can be eliminated by the good old Fuji -1/3 EC which can be common. is that an artifact in the testing protocol of photonstophotos?

Click to expand...

See above.

JimKasson said:

So my question is does the ISO 80 produce slightly more DR or is that an artifact in the testing protocol of photonstophotos?

Click to expand...

It does, but it does it with a (cheap, IMO) trick -- jettisoning half the read noise -- that is not reflective of a read noise difference after demosaicing and black point compensation in the raw converter.

JimKasson said:

If this shift is understood and calibrated for in the metering and calculating the exposure - is there a reason to avoid ISO 80. if there DR is the same at 100 and 80, then that would be such a reason.

Click to expand...

The FWC is higher at ISO 80, so that could be reason to use it if you are OK with what it does to the black point.

Click to expand...

Truman Prevatt said:

JimKasson said:

Truman Prevatt said:

JimKasson said:

Truman Prevatt said:

eliehbk said:

So which one to use to have the most dynamic range and amount of noise?

Click to expand...

I second that question. P2P shows additional about 1/3 a stops at ISO 80. On the other hand the histograms shown point to about 1/3 stop shift toward highlight end as a result of the metering. So the question seems to be realizing that and using a -1/3 EC to shift the exposure down by 1/3 of a stop to compensate from the metering and the way the histogram is calculated at ISO 80 result the in slight amount of DR shown in Photons2Photos?

Click to expand...

Which DR metric are you referring to? Bill gives you metrics that allow at least two.

Truman Prevatt said:

As an aside on the 100RF I modulate the ISO with the front wheel, keeping it as low as possible while meeting my requirement for shutter speed and aperture. I don't avoid 80 and I often find Fuji is not quite as aggressive as I would like in protecting highlights and often use -1/3 EC at all ISO's. I actually would like Fuji to implement a second metering mode similar to that on my Z8 with a little more highlight protection. They don't, I do it myself. I've not notice any negative impact on shooting ISO 80. I might be missing something or might not have run into a scene where there would be an impact.

So while charts and histograms are nice, question arise when two different sources show charts that point to what seems to be contradictory conclusions. What does it mean using the camera in the wild?

Click to expand...

Can you be explicit and detailed about the contradictions you see? Bill and I are generally pretty much in agreement.

Click to expand...

The photonstophotos chart givens a PDR of 12.55 at ISO 80 and 12.29 at ISO 100, or a little less than 1/3 stop more DR for that sensor.

Click to expand...

So your chosen metric for DR is PDR? That's a pretty good one.

Truman Prevatt said:

The histograms shown in one of the post shows a slight shift toward the highlight ( a little more prone to clip) at ISO80 than ISO100.

Click to expand...

That is not taken into account in photon transfer functions like the ones that Bill and I calculate.

Truman Prevatt said:

The shift in the histogram can be eliminated by the good old Fuji -1/3 EC which can be common. is that an artifact in the testing protocol of photonstophotos?

Click to expand...

See above.

Truman Prevatt said:

So my question is does the ISO 80 produce slightly more DR or is that an artifact in the testing protocol of photonstophotos?

Click to expand...

It does, but it does it with a (cheap, IMO) trick -- jettisoning half the read noise -- that is not reflective of a read noise difference after demosaicing and black point compensation in the raw converter.

Truman Prevatt said:

If this shift is understood and calibrated for in the metering and calculating the exposure - is there a reason to avoid ISO 80. if there DR is the same at 100 and 80, then that would be such a reason.

Click to expand...

The FWC is higher at ISO 80, so that could be reason to use it if you are OK with what it does to the black point.

Click to expand...

Thanks, Jim. The light bulb finally went on. Let me summarize what you are saying. By eliminating half the read noise, the resulting read noise is no longer Gaussian zero mean. Therefore, there is little to be gained at ISO 80 and the slight DR increase is more or less a mirage.

Click to expand...

highdesertmesa said:

Truman Prevatt said:

JimKasson said:

Truman Prevatt said:

JimKasson said:

Truman Prevatt said:

eliehbk said:

So which one to use to have the most dynamic range and amount of noise?

Click to expand...

I second that question. P2P shows additional about 1/3 a stops at ISO 80. On the other hand the histograms shown point to about 1/3 stop shift toward highlight end as a result of the metering. So the question seems to be realizing that and using a -1/3 EC to shift the exposure down by 1/3 of a stop to compensate from the metering and the way the histogram is calculated at ISO 80 result the in slight amount of DR shown in Photons2Photos?

Click to expand...

Which DR metric are you referring to? Bill gives you metrics that allow at least two.

highdesertmesa said:

As an aside on the 100RF I modulate the ISO with the front wheel, keeping it as low as possible while meeting my requirement for shutter speed and aperture. I don't avoid 80 and I often find Fuji is not quite as aggressive as I would like in protecting highlights and often use -1/3 EC at all ISO's. I actually would like Fuji to implement a second metering mode similar to that on my Z8 with a little more highlight protection. They don't, I do it myself. I've not notice any negative impact on shooting ISO 80. I might be missing something or might not have run into a scene where there would be an impact.

So while charts and histograms are nice, question arise when two different sources show charts that point to what seems to be contradictory conclusions. What does it mean using the camera in the wild?

Click to expand...

Can you be explicit and detailed about the contradictions you see? Bill and I are generally pretty much in agreement.

Click to expand...

The photonstophotos chart givens a PDR of 12.55 at ISO 80 and 12.29 at ISO 100, or a little less than 1/3 stop more DR for that sensor.

Click to expand...

So your chosen metric for DR is PDR? That's a pretty good one.

highdesertmesa said:

The histograms shown in one of the post shows a slight shift toward the highlight ( a little more prone to clip) at ISO80 than ISO100.

Click to expand...

That is not taken into account in photon transfer functions like the ones that Bill and I calculate.

highdesertmesa said:

The shift in the histogram can be eliminated by the good old Fuji -1/3 EC which can be common. is that an artifact in the testing protocol of photonstophotos?

Click to expand...

See above.

highdesertmesa said:

So my question is does the ISO 80 produce slightly more DR or is that an artifact in the testing protocol of photonstophotos?

Click to expand...

It does, but it does it with a (cheap, IMO) trick -- jettisoning half the read noise -- that is not reflective of a read noise difference after demosaicing and black point compensation in the raw converter.

highdesertmesa said:

If this shift is understood and calibrated for in the metering and calculating the exposure - is there a reason to avoid ISO 80. if there DR is the same at 100 and 80, then that would be such a reason.

Click to expand...

The FWC is higher at ISO 80, so that could be reason to use it if you are OK with what it does to the black point.

Click to expand...

Thanks, Jim. The light bulb finally went on. Let me summarize what you are saying. By eliminating half the read noise, the resulting read noise is no longer Gaussian zero mean. Therefore, there is little to be gained at ISO 80 and the slight DR increase is more or less a mirage.

Click to expand...

It's still native ISO 80, it just has the same black point subtraction you could do in post to eliminate read noise. The only downside is 1) you like read noise and are mad they removed it and/or 2) you don't know they did this and can't establish a meaningful relationship between the meter and true RawDigger exposure when bouncing back and forth between ISO 80 and other ISOs.

If you don't mind the black point subtraction and know how to meter at ISO 80, I don't see why one wouldn't want to use ISO 80.

Personally – I now like ISO 80 because the in-camera metering histogram and blinking highlight warnings are very close to matching the RAW files. At ISO 100 and higher, I have to intentionally overexpose a little (and guesstimate) in order to give the exposure enough light to saturate the RAW without clipping.

Click to expand...

highdesertmesa said:

Truman Prevatt said:

JimKasson said:

Truman Prevatt said:

JimKasson said:

Truman Prevatt said:

eliehbk said:

So which one to use to have the most dynamic range and amount of noise?

Click to expand...

I second that question. P2P shows additional about 1/3 a stops at ISO 80. On the other hand the histograms shown point to about 1/3 stop shift toward highlight end as a result of the metering. So the question seems to be realizing that and using a -1/3 EC to shift the exposure down by 1/3 of a stop to compensate from the metering and the way the histogram is calculated at ISO 80 result the in slight amount of DR shown in Photons2Photos?

Click to expand...

Which DR metric are you referring to? Bill gives you metrics that allow at least two.

highdesertmesa said:

As an aside on the 100RF I modulate the ISO with the front wheel, keeping it as low as possible while meeting my requirement for shutter speed and aperture. I don't avoid 80 and I often find Fuji is not quite as aggressive as I would like in protecting highlights and often use -1/3 EC at all ISO's. I actually would like Fuji to implement a second metering mode similar to that on my Z8 with a little more highlight protection. They don't, I do it myself. I've not notice any negative impact on shooting ISO 80. I might be missing something or might not have run into a scene where there would be an impact.

So while charts and histograms are nice, question arise when two different sources show charts that point to what seems to be contradictory conclusions. What does it mean using the camera in the wild?

Click to expand...

Can you be explicit and detailed about the contradictions you see? Bill and I are generally pretty much in agreement.

Click to expand...

The photonstophotos chart givens a PDR of 12.55 at ISO 80 and 12.29 at ISO 100, or a little less than 1/3 stop more DR for that sensor.

Click to expand...

So your chosen metric for DR is PDR? That's a pretty good one.

highdesertmesa said:

The histograms shown in one of the post shows a slight shift toward the highlight ( a little more prone to clip) at ISO80 than ISO100.

Click to expand...

That is not taken into account in photon transfer functions like the ones that Bill and I calculate.

highdesertmesa said:

The shift in the histogram can be eliminated by the good old Fuji -1/3 EC which can be common. is that an artifact in the testing protocol of photonstophotos?

Click to expand...

See above.

highdesertmesa said:

So my question is does the ISO 80 produce slightly more DR or is that an artifact in the testing protocol of photonstophotos?

Click to expand...

It does, but it does it with a (cheap, IMO) trick -- jettisoning half the read noise -- that is not reflective of a read noise difference after demosaicing and black point compensation in the raw converter.

highdesertmesa said:

If this shift is understood and calibrated for in the metering and calculating the exposure - is there a reason to avoid ISO 80. if there DR is the same at 100 and 80, then that would be such a reason.

Click to expand...

The FWC is higher at ISO 80, so that could be reason to use it if you are OK with what it does to the black point.

Click to expand...

Thanks, Jim. The light bulb finally went on. Let me summarize what you are saying. By eliminating half the read noise, the resulting read noise is no longer Gaussian zero mean. Therefore, there is little to be gained at ISO 80 and the slight DR increase is more or less a mirage.

Click to expand...

It's still native ISO 80, it just has the same black point subtraction you could do in post to eliminate read noise.

Click to expand...

highdesertmesa said:

The only downside is 1) you like read noise and are mad they removed it

Click to expand...

highdesertmesa said:

and/or 2) you don't know they did this and can't establish a meaningful relationship between the meter and true RawDigger exposure when bouncing back and forth between ISO 80 and other ISOs.

Click to expand...

highdesertmesa said:

If you don't mind the black point subtraction and know how to meter at ISO 80, I don't see why one wouldn't want to use ISO 80.

Personally – I now like ISO 80 because the in-camera metering histogram and blinking highlight warnings are very close to matching the RAW files.

Click to expand...

highdesertmesa said:

At ISO 100 and higher, I have to intentionally overexpose a little (and guesstimate) in order to give the exposure enough light to saturate the RAW without clipping.

Click to expand...

JimKasson said:

highdesertmesa said:

Truman Prevatt said:

JimKasson said:

Truman Prevatt said:

JimKasson said:

Truman Prevatt said:

eliehbk said:

So which one to use to have the most dynamic range and amount of noise?

Click to expand...

I second that question. P2P shows additional about 1/3 a stops at ISO 80. On the other hand the histograms shown point to about 1/3 stop shift toward highlight end as a result of the metering. So the question seems to be realizing that and using a -1/3 EC to shift the exposure down by 1/3 of a stop to compensate from the metering and the way the histogram is calculated at ISO 80 result the in slight amount of DR shown in Photons2Photos?

Click to expand...

Which DR metric are you referring to? Bill gives you metrics that allow at least two.

JimKasson said:

As an aside on the 100RF I modulate the ISO with the front wheel, keeping it as low as possible while meeting my requirement for shutter speed and aperture. I don't avoid 80 and I often find Fuji is not quite as aggressive as I would like in protecting highlights and often use -1/3 EC at all ISO's. I actually would like Fuji to implement a second metering mode similar to that on my Z8 with a little more highlight protection. They don't, I do it myself. I've not notice any negative impact on shooting ISO 80. I might be missing something or might not have run into a scene where there would be an impact.

So while charts and histograms are nice, question arise when two different sources show charts that point to what seems to be contradictory conclusions. What does it mean using the camera in the wild?

Click to expand...

Can you be explicit and detailed about the contradictions you see? Bill and I are generally pretty much in agreement.

Click to expand...

The photonstophotos chart givens a PDR of 12.55 at ISO 80 and 12.29 at ISO 100, or a little less than 1/3 stop more DR for that sensor.

Click to expand...

So your chosen metric for DR is PDR? That's a pretty good one.

JimKasson said:

The histograms shown in one of the post shows a slight shift toward the highlight ( a little more prone to clip) at ISO80 than ISO100.

Click to expand...

That is not taken into account in photon transfer functions like the ones that Bill and I calculate.

JimKasson said:

The shift in the histogram can be eliminated by the good old Fuji -1/3 EC which can be common. is that an artifact in the testing protocol of photonstophotos?

Click to expand...

See above.

JimKasson said:

So my question is does the ISO 80 produce slightly more DR or is that an artifact in the testing protocol of photonstophotos?

Click to expand...

It does, but it does it with a (cheap, IMO) trick -- jettisoning half the read noise -- that is not reflective of a read noise difference after demosaicing and black point compensation in the raw converter.

JimKasson said:

If this shift is understood and calibrated for in the metering and calculating the exposure - is there a reason to avoid ISO 80. if there DR is the same at 100 and 80, then that would be such a reason.

Click to expand...

The FWC is higher at ISO 80, so that could be reason to use it if you are OK with what it does to the black point.

Click to expand...

Thanks, Jim. The light bulb finally went on. Let me summarize what you are saying. By eliminating half the read noise, the resulting read noise is no longer Gaussian zero mean. Therefore, there is little to be gained at ISO 80 and the slight DR increase is more or less a mirage.

Click to expand...

It's still native ISO 80, it just has the same black point subtraction you could do in post to eliminate read noise.

Click to expand...

It doesn't have the FWC you'd expect. And if they get it wrong, as so many camera manufacturers do when they do in-camera black point subtraction, and they get it wrong in the wrong direction, the errors are irrecoverable. They also are doing the black point subtraction at the precision of the raw file. In post you can do it with greater precision.

Click to expand...

JimKasson said:

JimKasson said:

The only downside is 1) you like read noise and are mad they removed it

Click to expand...

This is sophistic. There are good reasons not to subtract the black point in camera. If subtracting it in camera is a good idea, why do they only so it for ISO 80?

Click to expand...

JimKasson said:

JimKasson said:

and/or 2) you don't know they did this and can't establish a meaningful relationship between the meter and true RawDigger exposure when bouncing back and forth between ISO 80 and other ISOs.

Click to expand...

There is that pesky FWC thingie.

JimKasson said:

If you don't mind the black point subtraction and know how to meter at ISO 80, I don't see why one wouldn't want to use ISO 80.

Personally – I now like ISO 80 because the in-camera metering histogram and blinking highlight warnings are very close to matching the RAW files.

Click to expand...

If the metering is doen right, why should it vary with ISO setting?

Click to expand...

JimKasson said:

JimKasson said:

At ISO 100 and higher, I have to intentionally overexpose a little (and guesstimate) in order to give the exposure enough light to saturate the RAW without clipping.

Click to expand...

--
https://blog.kasson.com

Click to expand...

SrMi said:

highdesertmesa said:

Truman Prevatt said:

JimKasson said:

Truman Prevatt said:

JimKasson said:

Truman Prevatt said:

eliehbk said:

So which one to use to have the most dynamic range and amount of noise?

Click to expand...

I second that question. P2P shows additional about 1/3 a stops at ISO 80. On the other hand the histograms shown point to about 1/3 stop shift toward highlight end as a result of the metering. So the question seems to be realizing that and using a -1/3 EC to shift the exposure down by 1/3 of a stop to compensate from the metering and the way the histogram is calculated at ISO 80 result the in slight amount of DR shown in Photons2Photos?

Click to expand...

Which DR metric are you referring to? Bill gives you metrics that allow at least two.

SrMi said:

As an aside on the 100RF I modulate the ISO with the front wheel, keeping it as low as possible while meeting my requirement for shutter speed and aperture. I don't avoid 80 and I often find Fuji is not quite as aggressive as I would like in protecting highlights and often use -1/3 EC at all ISO's. I actually would like Fuji to implement a second metering mode similar to that on my Z8 with a little more highlight protection. They don't, I do it myself. I've not notice any negative impact on shooting ISO 80. I might be missing something or might not have run into a scene where there would be an impact.

So while charts and histograms are nice, question arise when two different sources show charts that point to what seems to be contradictory conclusions. What does it mean using the camera in the wild?

Click to expand...

Can you be explicit and detailed about the contradictions you see? Bill and I are generally pretty much in agreement.

Click to expand...

The photonstophotos chart givens a PDR of 12.55 at ISO 80 and 12.29 at ISO 100, or a little less than 1/3 stop more DR for that sensor.

Click to expand...

So your chosen metric for DR is PDR? That's a pretty good one.

SrMi said:

The histograms shown in one of the post shows a slight shift toward the highlight ( a little more prone to clip) at ISO80 than ISO100.

Click to expand...

That is not taken into account in photon transfer functions like the ones that Bill and I calculate.

SrMi said:

The shift in the histogram can be eliminated by the good old Fuji -1/3 EC which can be common. is that an artifact in the testing protocol of photonstophotos?

Click to expand...

See above.

SrMi said:

So my question is does the ISO 80 produce slightly more DR or is that an artifact in the testing protocol of photonstophotos?

Click to expand...

It does, but it does it with a (cheap, IMO) trick -- jettisoning half the read noise -- that is not reflective of a read noise difference after demosaicing and black point compensation in the raw converter.

SrMi said:

If this shift is understood and calibrated for in the metering and calculating the exposure - is there a reason to avoid ISO 80. if there DR is the same at 100 and 80, then that would be such a reason.

Click to expand...

The FWC is higher at ISO 80, so that could be reason to use it if you are OK with what it does to the black point.

Click to expand...

Thanks, Jim. The light bulb finally went on. Let me summarize what you are saying. By eliminating half the read noise, the resulting read noise is no longer Gaussian zero mean. Therefore, there is little to be gained at ISO 80 and the slight DR increase is more or less a mirage.

Click to expand...

It's still native ISO 80, it just has the same black point subtraction you could do in post to eliminate read noise. The only downside is 1) you like read noise and are mad they removed it and/or 2) you don't know they did this and can't establish a meaningful relationship between the meter and true RawDigger exposure when bouncing back and forth between ISO 80 and other ISOs.

If you don't mind the black point subtraction and know how to meter at ISO 80, I don't see why one wouldn't want to use ISO 80.

Personally – I now like ISO 80 because the in-camera metering histogram and blinking highlight warnings are very close to matching the RAW files. At ISO 100 and higher, I have to intentionally overexpose a little (and guesstimate) in order to give the exposure enough light to saturate the RAW without clipping.

Click to expand...

The problem is that the relation of JPEG histogram to raw histogram is different at ISO80 than and other ISOs. Automatic metering also gives you different results.

Click to expand...

highdesertmesa said:

SrMi said:

highdesertmesa said:

Truman Prevatt said:

JimKasson said:

Truman Prevatt said:

JimKasson said:

Truman Prevatt said:

eliehbk said:

So which one to use to have the most dynamic range and amount of noise?

Click to expand...

I second that question. P2P shows additional about 1/3 a stops at ISO 80. On the other hand the histograms shown point to about 1/3 stop shift toward highlight end as a result of the metering. So the question seems to be realizing that and using a -1/3 EC to shift the exposure down by 1/3 of a stop to compensate from the metering and the way the histogram is calculated at ISO 80 result the in slight amount of DR shown in Photons2Photos?

Click to expand...

Which DR metric are you referring to? Bill gives you metrics that allow at least two.

highdesertmesa said:

As an aside on the 100RF I modulate the ISO with the front wheel, keeping it as low as possible while meeting my requirement for shutter speed and aperture. I don't avoid 80 and I often find Fuji is not quite as aggressive as I would like in protecting highlights and often use -1/3 EC at all ISO's. I actually would like Fuji to implement a second metering mode similar to that on my Z8 with a little more highlight protection. They don't, I do it myself. I've not notice any negative impact on shooting ISO 80. I might be missing something or might not have run into a scene where there would be an impact.

So while charts and histograms are nice, question arise when two different sources show charts that point to what seems to be contradictory conclusions. What does it mean using the camera in the wild?

Click to expand...

Can you be explicit and detailed about the contradictions you see? Bill and I are generally pretty much in agreement.

Click to expand...

The photonstophotos chart givens a PDR of 12.55 at ISO 80 and 12.29 at ISO 100, or a little less than 1/3 stop more DR for that sensor.

Click to expand...

So your chosen metric for DR is PDR? That's a pretty good one.

highdesertmesa said:

The histograms shown in one of the post shows a slight shift toward the highlight ( a little more prone to clip) at ISO80 than ISO100.

Click to expand...

That is not taken into account in photon transfer functions like the ones that Bill and I calculate.

highdesertmesa said:

The shift in the histogram can be eliminated by the good old Fuji -1/3 EC which can be common. is that an artifact in the testing protocol of photonstophotos?

Click to expand...

See above.

highdesertmesa said:

So my question is does the ISO 80 produce slightly more DR or is that an artifact in the testing protocol of photonstophotos?

Click to expand...

It does, but it does it with a (cheap, IMO) trick -- jettisoning half the read noise -- that is not reflective of a read noise difference after demosaicing and black point compensation in the raw converter.

highdesertmesa said:

If this shift is understood and calibrated for in the metering and calculating the exposure - is there a reason to avoid ISO 80. if there DR is the same at 100 and 80, then that would be such a reason.

Click to expand...

The FWC is higher at ISO 80, so that could be reason to use it if you are OK with what it does to the black point.

Click to expand...

Thanks, Jim. The light bulb finally went on. Let me summarize what you are saying. By eliminating half the read noise, the resulting read noise is no longer Gaussian zero mean. Therefore, there is little to be gained at ISO 80 and the slight DR increase is more or less a mirage.

Click to expand...

It's still native ISO 80, it just has the same black point subtraction you could do in post to eliminate read noise. The only downside is 1) you like read noise and are mad they removed it and/or 2) you don't know they did this and can't establish a meaningful relationship between the meter and true RawDigger exposure when bouncing back and forth between ISO 80 and other ISOs.

If you don't mind the black point subtraction and know how to meter at ISO 80, I don't see why one wouldn't want to use ISO 80.

Personally – I now like ISO 80 because the in-camera metering histogram and blinking highlight warnings are very close to matching the RAW files. At ISO 100 and higher, I have to intentionally overexpose a little (and guesstimate) in order to give the exposure enough light to saturate the RAW without clipping.

Click to expand...

The problem is that the relation of JPEG histogram to raw histogram is different at ISO80 than and other ISOs. Automatic metering also gives you different results.

Click to expand...

Yes I was acknowledging that and saying if one is aware of it, it can be an advantage because the in-camera metering at ISO 80 is close to what I would use for ETTR. I just have to remember to revert to my usual more aggressive ETTR at ISO 100 and up.

Click to expand...

SrMi said:

highdesertmesa said:

SrMi said:

highdesertmesa said:

Truman Prevatt said:

JimKasson said:

Truman Prevatt said:

JimKasson said:

Truman Prevatt said:

eliehbk said:

So which one to use to have the most dynamic range and amount of noise?

Click to expand...

I second that question. P2P shows additional about 1/3 a stops at ISO 80. On the other hand the histograms shown point to about 1/3 stop shift toward highlight end as a result of the metering. So the question seems to be realizing that and using a -1/3 EC to shift the exposure down by 1/3 of a stop to compensate from the metering and the way the histogram is calculated at ISO 80 result the in slight amount of DR shown in Photons2Photos?

Click to expand...

Which DR metric are you referring to? Bill gives you metrics that allow at least two.

SrMi said:

As an aside on the 100RF I modulate the ISO with the front wheel, keeping it as low as possible while meeting my requirement for shutter speed and aperture. I don't avoid 80 and I often find Fuji is not quite as aggressive as I would like in protecting highlights and often use -1/3 EC at all ISO's. I actually would like Fuji to implement a second metering mode similar to that on my Z8 with a little more highlight protection. They don't, I do it myself. I've not notice any negative impact on shooting ISO 80. I might be missing something or might not have run into a scene where there would be an impact.

So while charts and histograms are nice, question arise when two different sources show charts that point to what seems to be contradictory conclusions. What does it mean using the camera in the wild?

Click to expand...

Can you be explicit and detailed about the contradictions you see? Bill and I are generally pretty much in agreement.

Click to expand...

The photonstophotos chart givens a PDR of 12.55 at ISO 80 and 12.29 at ISO 100, or a little less than 1/3 stop more DR for that sensor.

Click to expand...

So your chosen metric for DR is PDR? That's a pretty good one.

SrMi said:

The histograms shown in one of the post shows a slight shift toward the highlight ( a little more prone to clip) at ISO80 than ISO100.

Click to expand...

That is not taken into account in photon transfer functions like the ones that Bill and I calculate.

SrMi said:

The shift in the histogram can be eliminated by the good old Fuji -1/3 EC which can be common. is that an artifact in the testing protocol of photonstophotos?

Click to expand...

See above.

SrMi said:

So my question is does the ISO 80 produce slightly more DR or is that an artifact in the testing protocol of photonstophotos?

Click to expand...

It does, but it does it with a (cheap, IMO) trick -- jettisoning half the read noise -- that is not reflective of a read noise difference after demosaicing and black point compensation in the raw converter.

SrMi said:

If this shift is understood and calibrated for in the metering and calculating the exposure - is there a reason to avoid ISO 80. if there DR is the same at 100 and 80, then that would be such a reason.

Click to expand...

The FWC is higher at ISO 80, so that could be reason to use it if you are OK with what it does to the black point.

Click to expand...

Thanks, Jim. The light bulb finally went on. Let me summarize what you are saying. By eliminating half the read noise, the resulting read noise is no longer Gaussian zero mean. Therefore, there is little to be gained at ISO 80 and the slight DR increase is more or less a mirage.

Click to expand...

It's still native ISO 80, it just has the same black point subtraction you could do in post to eliminate read noise. The only downside is 1) you like read noise and are mad they removed it and/or 2) you don't know they did this and can't establish a meaningful relationship between the meter and true RawDigger exposure when bouncing back and forth between ISO 80 and other ISOs.

If you don't mind the black point subtraction and know how to meter at ISO 80, I don't see why one wouldn't want to use ISO 80.

Personally – I now like ISO 80 because the in-camera metering histogram and blinking highlight warnings are very close to matching the RAW files. At ISO 100 and higher, I have to intentionally overexpose a little (and guesstimate) in order to give the exposure enough light to saturate the RAW without clipping.

Click to expand...

The problem is that the relation of JPEG histogram to raw histogram is different at ISO80 than and other ISOs. Automatic metering also gives you different results.

Click to expand...

Yes I was acknowledging that and saying if one is aware of it, it can be an advantage because the in-camera metering at ISO 80 is close to what I would use for ETTR. I just have to remember to revert to my usual more aggressive ETTR at ISO 100 and up.

Click to expand...

I do not see that as an advantage. Both ISO are off from real raw histogram, but at different distances, which, btw, also varies by the scene.

Click to expand...

highdesertmesa said:

SrMi said:

highdesertmesa said:

SrMi said:

highdesertmesa said:

Truman Prevatt said:

JimKasson said:

Truman Prevatt said:

JimKasson said:

Truman Prevatt said:

eliehbk said:

So which one to use to have the most dynamic range and amount of noise?

Click to expand...

I second that question. P2P shows additional about 1/3 a stops at ISO 80. On the other hand the histograms shown point to about 1/3 stop shift toward highlight end as a result of the metering. So the question seems to be realizing that and using a -1/3 EC to shift the exposure down by 1/3 of a stop to compensate from the metering and the way the histogram is calculated at ISO 80 result the in slight amount of DR shown in Photons2Photos?

Click to expand...

Which DR metric are you referring to? Bill gives you metrics that allow at least two.

highdesertmesa said:

As an aside on the 100RF I modulate the ISO with the front wheel, keeping it as low as possible while meeting my requirement for shutter speed and aperture. I don't avoid 80 and I often find Fuji is not quite as aggressive as I would like in protecting highlights and often use -1/3 EC at all ISO's. I actually would like Fuji to implement a second metering mode similar to that on my Z8 with a little more highlight protection. They don't, I do it myself. I've not notice any negative impact on shooting ISO 80. I might be missing something or might not have run into a scene where there would be an impact.

So while charts and histograms are nice, question arise when two different sources show charts that point to what seems to be contradictory conclusions. What does it mean using the camera in the wild?

Click to expand...

Can you be explicit and detailed about the contradictions you see? Bill and I are generally pretty much in agreement.

Click to expand...

The photonstophotos chart givens a PDR of 12.55 at ISO 80 and 12.29 at ISO 100, or a little less than 1/3 stop more DR for that sensor.

Click to expand...

So your chosen metric for DR is PDR? That's a pretty good one.

highdesertmesa said:

The histograms shown in one of the post shows a slight shift toward the highlight ( a little more prone to clip) at ISO80 than ISO100.

Click to expand...

That is not taken into account in photon transfer functions like the ones that Bill and I calculate.

highdesertmesa said:

The shift in the histogram can be eliminated by the good old Fuji -1/3 EC which can be common. is that an artifact in the testing protocol of photonstophotos?

Click to expand...

See above.

highdesertmesa said:

So my question is does the ISO 80 produce slightly more DR or is that an artifact in the testing protocol of photonstophotos?

Click to expand...

It does, but it does it with a (cheap, IMO) trick -- jettisoning half the read noise -- that is not reflective of a read noise difference after demosaicing and black point compensation in the raw converter.

highdesertmesa said:

If this shift is understood and calibrated for in the metering and calculating the exposure - is there a reason to avoid ISO 80. if there DR is the same at 100 and 80, then that would be such a reason.

Click to expand...

The FWC is higher at ISO 80, so that could be reason to use it if you are OK with what it does to the black point.

Click to expand...

Thanks, Jim. The light bulb finally went on. Let me summarize what you are saying. By eliminating half the read noise, the resulting read noise is no longer Gaussian zero mean. Therefore, there is little to be gained at ISO 80 and the slight DR increase is more or less a mirage.

Click to expand...

It's still native ISO 80, it just has the same black point subtraction you could do in post to eliminate read noise. The only downside is 1) you like read noise and are mad they removed it and/or 2) you don't know they did this and can't establish a meaningful relationship between the meter and true RawDigger exposure when bouncing back and forth between ISO 80 and other ISOs.

If you don't mind the black point subtraction and know how to meter at ISO 80, I don't see why one wouldn't want to use ISO 80.

Personally – I now like ISO 80 because the in-camera metering histogram and blinking highlight warnings are very close to matching the RAW files. At ISO 100 and higher, I have to intentionally overexpose a little (and guesstimate) in order to give the exposure enough light to saturate the RAW without clipping.

Click to expand...

The problem is that the relation of JPEG histogram to raw histogram is different at ISO80 than and other ISOs. Automatic metering also gives you different results.

Click to expand...

Yes I was acknowledging that and saying if one is aware of it, it can be an advantage because the in-camera metering at ISO 80 is close to what I would use for ETTR. I just have to remember to revert to my usual more aggressive ETTR at ISO 100 and up.

Click to expand...

I do not see that as an advantage. Both ISO are off from real raw histogram, but at different distances, which, btw, also varies by the scene.

Click to expand...

Highlight clipping warning consistency should not be affected by scene changes at the same ISO. It’s not evaluating the entire scene like a metering mode, it’s finding any out of range highlight.

Or are you saying the amount of variance in the metering between 80 and 100 is influenced by something else, say large portions of the image being very dark — and therefore you’re saying the metering at 80 is not just different than 100, it’s also not consistent and therefore unpredictable relative to the RAW exposure in RawDigger?

Click to expand...

SrMi said:

highdesertmesa said:

SrMi said:

highdesertmesa said:

SrMi said:

highdesertmesa said:

Truman Prevatt said:

JimKasson said:

Truman Prevatt said:

JimKasson said:

Truman Prevatt said:

eliehbk said:

So which one to use to have the most dynamic range and amount of noise?

Click to expand...

I second that question. P2P shows additional about 1/3 a stops at ISO 80. On the other hand the histograms shown point to about 1/3 stop shift toward highlight end as a result of the metering. So the question seems to be realizing that and using a -1/3 EC to shift the exposure down by 1/3 of a stop to compensate from the metering and the way the histogram is calculated at ISO 80 result the in slight amount of DR shown in Photons2Photos?

Click to expand...

Which DR metric are you referring to? Bill gives you metrics that allow at least two.

SrMi said:

As an aside on the 100RF I modulate the ISO with the front wheel, keeping it as low as possible while meeting my requirement for shutter speed and aperture. I don't avoid 80 and I often find Fuji is not quite as aggressive as I would like in protecting highlights and often use -1/3 EC at all ISO's. I actually would like Fuji to implement a second metering mode similar to that on my Z8 with a little more highlight protection. They don't, I do it myself. I've not notice any negative impact on shooting ISO 80. I might be missing something or might not have run into a scene where there would be an impact.

So while charts and histograms are nice, question arise when two different sources show charts that point to what seems to be contradictory conclusions. What does it mean using the camera in the wild?

Click to expand...

Can you be explicit and detailed about the contradictions you see? Bill and I are generally pretty much in agreement.

Click to expand...

The photonstophotos chart givens a PDR of 12.55 at ISO 80 and 12.29 at ISO 100, or a little less than 1/3 stop more DR for that sensor.

Click to expand...

So your chosen metric for DR is PDR? That's a pretty good one.

SrMi said:

The histograms shown in one of the post shows a slight shift toward the highlight ( a little more prone to clip) at ISO80 than ISO100.

Click to expand...

That is not taken into account in photon transfer functions like the ones that Bill and I calculate.

SrMi said:

The shift in the histogram can be eliminated by the good old Fuji -1/3 EC which can be common. is that an artifact in the testing protocol of photonstophotos?

Click to expand...

See above.

SrMi said:

So my question is does the ISO 80 produce slightly more DR or is that an artifact in the testing protocol of photonstophotos?

Click to expand...

It does, but it does it with a (cheap, IMO) trick -- jettisoning half the read noise -- that is not reflective of a read noise difference after demosaicing and black point compensation in the raw converter.

SrMi said:

If this shift is understood and calibrated for in the metering and calculating the exposure - is there a reason to avoid ISO 80. if there DR is the same at 100 and 80, then that would be such a reason.

Click to expand...

The FWC is higher at ISO 80, so that could be reason to use it if you are OK with what it does to the black point.

Click to expand...

Thanks, Jim. The light bulb finally went on. Let me summarize what you are saying. By eliminating half the read noise, the resulting read noise is no longer Gaussian zero mean. Therefore, there is little to be gained at ISO 80 and the slight DR increase is more or less a mirage.

Click to expand...

It's still native ISO 80, it just has the same black point subtraction you could do in post to eliminate read noise. The only downside is 1) you like read noise and are mad they removed it and/or 2) you don't know they did this and can't establish a meaningful relationship between the meter and true RawDigger exposure when bouncing back and forth between ISO 80 and other ISOs.

If you don't mind the black point subtraction and know how to meter at ISO 80, I don't see why one wouldn't want to use ISO 80.

Personally – I now like ISO 80 because the in-camera metering histogram and blinking highlight warnings are very close to matching the RAW files. At ISO 100 and higher, I have to intentionally overexpose a little (and guesstimate) in order to give the exposure enough light to saturate the RAW without clipping.

Click to expand...

The problem is that the relation of JPEG histogram to raw histogram is different at ISO80 than and other ISOs. Automatic metering also gives you different results.

Click to expand...

Yes I was acknowledging that and saying if one is aware of it, it can be an advantage because the in-camera metering at ISO 80 is close to what I would use for ETTR. I just have to remember to revert to my usual more aggressive ETTR at ISO 100 and up.

Click to expand...

I do not see that as an advantage. Both ISO are off from real raw histogram, but at different distances, which, btw, also varies by the scene.

Click to expand...

Highlight clipping warning consistency should not be affected by scene changes at the same ISO. It’s not evaluating the entire scene like a metering mode, it’s finding any out of range highlight.

Or are you saying the amount of variance in the metering between 80 and 100 is influenced by something else, say large portions of the image being very dark — and therefore you’re saying the metering at 80 is not just different than 100, it’s also not consistent and therefore unpredictable relative to the RAW exposure in RawDigger?

Click to expand...

It depends on colors and how they are lighted, IIRC.

Click to expand...

JimKasson said:

Here's what I got with the GFX 100 II.

Click to expand...

Astro-photographie.fr

astro-photographie.fr