Best ISO for IQ280

[voidshatter]

Hi,

- I would say that base ISO seems to be around 35 for the 80 MP backs.

- I don't know if full frame CCD has great dynamic range, that may depend on definition of both great and dynamic range. What I may see is that MFD systems tend to underexpose and thereby protect highlights. For instance, the default film curve in Capture One makes image very bright.

- Readout noise is higher on CCD, so the darks are noisy.

- Yes 16 bit is indeed awesome, but so is 14 bits and even 12 bits.

For the kind of work I am doing my P45+ is mostly quite OK, there are exceptions, however.

Best regards
Erik

Hi, base ISO for 80MP Dalsa is 100. More info here.

If you compare the dynamic range of any CCD (regardless of format size) against a 3-year old Nikon D800E you would get disappointed by the CCD. More info here and here.

Given the inferior dynamic range of all CCDs (more info here and here), there isn't much point to have 16-bit filled with noise in the lower bits. You get better room for post-processing with a 14-bit file from a modern Sony CMOS sensor (e.g. Pentax 645Z or Nikon D810).

[sbernthal]

for people like you who have no clue about what is called native ISO. You can dial +1.5EV exposure compensation for your flash and use ISO 100 without relying on ISO 35.

So if we're getting personal, I will offer these comments:

1. Do you own a Credo/IQ digital back? (no)
2. Do you own a studio? (no)

I would also add this: if you say "dial +1.5EV exposure compensation for your flash", this tells me what you think a flash is.
From your "superstitions" you clearly don't understand why digital backs exist and what they CAN do.
I would agree that for you buying a digital back was a mistake.
Of course, in your mind, all the tens of thousands of digital back buyers are suckers, and you are the one who sees the truth for what it really is.

And over all else - why would you offer advice on hardware you don't own, when so many others own it?
I will give you that you have a lot of appreciation for yourself and your ideas.

[voidshatter]

So if we're getting personal, I will offer these comments:

1. Do you own a Credo/IQ digital back? (no)
2. Do you own a studio? (no)

I would also add this: if you say "dial +1.5EV exposure compensation for your flash", this tells me what you think a flash is.
From your "superstitions" you clearly don't understand why digital backs exist and what they CAN do.
I would agree that for you buying a digital back was a mistake.
Of course, in your mind, all the tens of thousands of digital back buyers are suckers, and you are the one who sees the truth for what it really is.

And over all else - why would you offer advice on hardware you don't own, when so many others own it?
I will give you that you have a lot of appreciation for yourself and your ideas.

I don't know where your confidence is from but it is kinda funny that a Credo user can judge my gear. 

[Wayne Fox]

Hi, base ISO for 80MP Dalsa is 100. More info here.

Not sure where in that page you are getting this (long thread, no time), but DxO mark clearly shows the maximum dynamic range of the back is achieved at ISO 35, which is also what Phase stated was the base/native ISO of the back when it was introduced.

So using ISO 35 on the IQ back yields it’s greatest dynamic range. ISO 35 is not a “false” ISO setting like using ISO 50 on most dSLR’s which have CMOS sensors that achieve their best dynamic range at around 100 ISO.  If you move to IS0 100 on the IQ180/280, you lose about 2 stops of potential dynamic range.  You can check the chart at DxO to confirm this.  This is the anchor around the back, and why the IQ250 is offering so much better results with high dynamic range imagery. (and why many of us long for a full frame 80-100mp CMOS back designed well enough to perform with tech camera lenses).

To the OP, in a studio situation using ISO 100 may work just fine, because you may have enough control that 8 or 9 stops of dynamic range is all you need (and I think DxO still rates it at almost 12 stops at that ISO although I think the ratings are a stop or so abetter than I’m comfortable with). If you move to ISO 35 it forces you to add light, which is what the sensor is wanting.( No point in “ETTR’ to make up for this, that would basically be identical to moving to ISO 35 and adjusting the lights).  But bottom line if you set the camera to ISO 100, you are basically underexposing the ideal amount of light for the sensor by about 1.3 stops, which isn’t bad, all camera ISO settings trade more sensitivity for less dynamic range. CCD backs just do it a little differently than CMOS backs.

[sbernthal]

That's exactly what I was trying to say.
Maybe you explained it better.

[voidshatter]

Not sure where in that page you are getting this (long thread, no time), but DxO mark clearly shows the maximum dynamic range of the back is achieved at ISO 35, which is also what Phase stated was the base/native ISO of the back when it was introduced.

So using ISO 35 on the IQ back yields it’s greatest dynamic range. ISO 35 is not a “false” ISO setting like using ISO 50 on most dSLR’s which have CMOS sensors that achieve their best dynamic range at around 100 ISO.  If you move to IS0 100 on the IQ180/280, you lose about 2 stops of potential dynamic range.  You can check the chart at DxO to confirm this.  This is the anchor around the back, and why the IQ250 is offering so much better results with high dynamic range imagery. (and why many of us long for a full frame 80-100mp CMOS back designed well enough to perform with tech camera lenses).

To the OP, in a studio situation using ISO 100 may work just fine, because you may have enough control that 8 or 9 stops of dynamic range is all you need (and I think DxO still rates it at almost 12 stops at that ISO although I think the ratings are a stop or so abetter than I’m comfortable with). If you move to ISO 35 it forces you to add light, which is what the sensor is wanting.( No point in “ETTR’ to make up for this, that would basically be identical to moving to ISO 35 and adjusting the lights).  But bottom line if you set the camera to ISO 100, you are basically underexposing the ideal amount of light for the sensor by about 1.3 stops, which isn’t bad, all camera ISO settings trade more sensitivity for less dynamic range. CCD backs just do it a little differently than CMOS backs.

DxOMark only measures ISO 100 for the IQ180. They didn't even bother putting the data collected for ISO 35 because they knew that ISO 35 is identical as ISO 100. Neither did sensorgen.info because they also know that ISO 35 is just extended ISO offering you no benefit other than forcing you do ETTR of +1.5EV exposure compensation.

You gain the same dynamic range between ISO 35 and ISO 100 (they both give you the best dynamic range out of the IQ180). Anything above ISO 100 would cause the IQ180 lose dynamic range.

[ErikKaffehr]

Hi,

I vould rather say that DxO-mark measures base ISO to be 29 when the baxck is set to ISO 35. So at ISO 35 the camera is exposing very close to nominal ISO. AFAIK as I recall at 100 ISO the back will underexpose 1.5 stops or so. That may give a margin of 1.5EV in the highlights, unless thrown away in the generated raw file.

So, I would say best ISO is 35.

As a side aspect, the two factors involved are quantum efficiency (QE)  and full well capacity (FWC). Low QE leads to low base ISO, as it takes many photons to fill the well, due to low QE. On the other hand large FWC also leads to low base ISO.

The impression I have is that the new Sony sensors have quite large FWC compared to pixel size, so their base ISO is also quite low.

Best regards
Erik

DxOMark only measures ISO 100 for the IQ180. They didn't even bother putting the data collected for ISO 35 because they knew that ISO 35 is identical as ISO 100. Neither did sensorgen.info because they also know that ISO 35 is just extended ISO offering you no benefit other than forcing you do ETTR of +1.5EV exposure compensation.

You gain the same dynamic range between ISO 35 and ISO 100 (they both give you the best dynamic range out of the IQ180). Anything above ISO 100 would cause the IQ180 lose dynamic range.

[voidshatter]

Hi,

I vould rather say that DxO-mark measures base ISO to be 29 when the baxck is set to ISO 35. So at ISO 35 the camera is exposing very close to nominal ISO. AFAIK as I recall at 100 ISO the back will underexpose 1.5 stops or so. That may give a margin of 1.5EV in the highlights, unless thrown away in the generated raw file.

So, I would say best ISO is 35.

As a side aspect, the two factors involved are quantum efficiency (QE)  and full well capacity (FWC). Low QE leads to low base ISO, as it takes many photons to fill the well, due to low QE. On the other hand large FWC also leads to low base ISO.

The impression I have is that the new Sony sensors have quite large FWC compared to pixel size, so their base ISO is also quite low.

Best regards
Erik

Hi,

Dxo also measures ISO 29 when the IQ180 is shooting at ISO 100. There is no difference between 35 and 100. The IQ180 is essentially not up to the standard when advertising ISO. It's Phase One's clever way to wrap up for marketing purposes to make users feel good. (You buy something that doesn't perform to meet ISO standard but you are fooled by the extended ISO trick.)

[BJL]

Dxo also measures ISO 29 when the IQ180 is shooting at ISO 100. There is no difference between 35 and 100.

Which means that the base ISO speed (the highlight saturation-based _minimum_ recommended exposure index according to the relevant ISO standard ISO12232, where it is called SSat) is 29, and there is no difference in the in-camera amplification between the 35 and 100 settings for Exposure Index.

This if anything means that the ISO Exposure Index setting of 35 is closest to the saturation-based measurement SSat that DXO measures (and annoyingly insists on suggesting is the one and only true "ISO" speed measurement, in flat contradiction to what the ISO standard itself says!).  I do not see that EI=35 setting as a marketing trick, but it does mean than when working at that setting and exposing "on-meter", you might have to be more careful about not blowing highlights, because there are only about three stops between metered average light levels and blown highlights, and this is the bare minimum headroom suggested by the ISO standard, not an ideal as some people seem to believe.

P. S. That Ssat [= "ISO" as measured and reported by DXO] is roughly speaking the exposure index such that if metering on a subject of 18% reflectivity, a part of the scene of 100% reflectivity and the same illumination would come in one stop below saturation.

[voidshatter]

Which means that the base ISO speed (the highlight saturation-based _minimum_ recommended exposure index according to the relevant ISO standard ISO12232, where it is called SSat) is 29, and there is no difference in the in-camera amplification between the 35 and 100 settings for Exposure Index.

This if anything means that the ISO Exposure Index setting of 35 is closest to the saturation-based measurement SSat that DXO measures (and annoyingly insists on suggesting is the one and only true "ISO" speed measurement, in flat contradiction to what the ISO standard itself says!).  I do not see that EI=35 setting as a marketing trick, but it does mean than when working at that setting and exposing "on-meter", you might have to be more careful about not blowing highlights, because there are only about three stops between metered average light levels and blown highlights, and this is the bare minimum headroom suggested by the ISO standard, not an ideal as some people seem to believe.

It is indeed a trick to hide the fact that the low light performance is very poor. If Phase One is honest to have the manufacturer's ISO match the measured ISO (so that the plot is close to the nominal line) then everyone will be able to see that noise is catastrophic for anything above ISO 200 (instead of what people now see for ISO 800, severely not up to the ISO standard).

[JoeKitchen]

It is indeed a trick to hide the fact that the low light performance is very poor. If Phase One is honest to have the manufacturer's ISO match the measured ISO (so that the plot is close to the nominal line) then everyone will be able to see that noise is catastrophic for anything above ISO 200 (instead of what people now see for ISO 800, severely not up to the ISO standard).

Really?  You just needed to work another bashing of Phase One into this conversation. 

And by the way, why anyone would under expose there image at base ISO to only push it in post to match an ISO 100 image is beyond me.

[sbernthal]

why anyone would under expose there image at base ISO to only push it in post to match an ISO 100 image is beyond me.

My point about pushing is not that you would do that intentionally, but yhat just in case you might need to push once you are in post, the ISO 35 has more data.

[voidshatter]

My point about pushing is not that you would do that intentionally, but yhat just in case you might need to push once you are in post, the ISO 35 has more data.

I agree that shadow recoverability is important, however given the same aperture, the same shutter speed and the same light, ISO 35 has no more data than ISO 100, because ISO 35 is just extended ISO, and ISO 100 is a base native ISO that is far away from the ISO standard. ISO 200 however, will be worse than ISO 100.

[ErikKaffehr]

I would say that base ISO as measured by DxO is 29 ISO, pretty close to 35, BTW.

Best regards
Erik

I agree that shadow recoverability is important, however given the same aperture, the same shutter speed and the same light, ISO 35 has no more data than ISO 100, because ISO 35 is just extended ISO, and ISO 100 is a base native ISO that is far away from the ISO standard. ISO 200 however, will be worse than ISO 100.

[JoeKitchen]

My point about pushing is not that you would do that intentionally, but yhat just in case you might need to push once you are in post, the ISO 35 has more data.

I agree with what you had to say.  

In my mind, capturing a perfect exposure at ISO 100, and then reducing the ISO to 35 without changing the light or exposure to capture another image, only to push that ISO 35 exposure in post to match the ISO 100 and compare, makes no sense.  

No one would actually shoot like that.  They would pick the ISO settings that suit their needs, given the perimeters and subject, and exposure correctly.  

So the test is kind of pointless for photographers who are concentrating on image making, not camera testing.  

[BJL]

It is indeed a trick to hide the fact that the low light performance is very poor.

I guess you mean that the old minimum exposure index setting of 100 was a trick, whereas the new minimum of 35 is more informative.

However, I note that base ISO speed has little to do with low light performance; it measures only highlight handing ability before saturation, not noise levels.  For example, if one sensor has a larger well capacity than another but the same quantum efficiency, dark noise, read noise etc. it will have a lower base ISO speed SSat but the same low light sensitivity and noise characteristics, as measured for example by the rarely used ISO noise-based standards "S40" and "S10": see Wikipedia on ISO noise-based speed measures  (I link to Wikipedia because ISO charges for access to its standards documents, but CIPA might give free access to related information.)

One has to be aware that the base ISO or SSat has little connection to the ISO measure of film speed -- if anything, ISO film speed is closer to those ISO SNR based measures.

[sbernthal]

I agree with what you had to say.  

In my mind, capturing a perfect exposure at ISO 100, and then reducing the ISO to 35 without changing the light or exposure to capture another image, only to push that ISO 35 exposure in post to match the ISO 100 and compare, makes no sense.  

No one would actually shoot like that.  They would pick the ISO settings that suit their needs, given the perimeters and subject, and exposure correctly.  

So the test is kind of pointless for photographers who are concentrating on image making, not camera testing.  

I am not able to follow his logic and he is not able to follow mine.
He is insisting on the "same light different ISO" test, which makes no sense to me.
It must be my Credo ownership that limits my intellect.
As he said - I have "no clue about what is called native ISO" - must be my 20 years as computer engineer followed by my 10 years as a professional photographer that hinder my technical understanding abilities.

ISO 35 has more information for pushing shadows - that is the main thing I need to know when deciding on my ISO, since I can make the light to whatever I want.
Note that by my tests, it is not correct that in general ISO 35 has more dynamic range - for recovering highlights it is worse than ISO 100.
Both make perfect sense, if you consider the number of photons hitting the sensor for each ISO.
BTW now I remember that when I got my camera I also did a long exposure test and ISO 35 was cleaner.

[RobertJ]

Well this thread turned into a shit storm.  I have learned nothing.

[Ken R]

Hi, base ISO for 80MP Dalsa is 100. More info here.

If you compare the dynamic range of any CCD (regardless of format size) against a 3-year old Nikon D800E you would get disappointed by the CCD. More info here and here.

Given the inferior dynamic range of all CCDs (more info here and here), there isn't much point to have 16-bit filled with noise in the lower bits. You get better room for post-processing with a 14-bit file from a modern Sony CMOS sensor (e.g. Pentax 645Z or Nikon D810).

Man, you do not give up. Gotta give you mad points for being persistent. You will not be happy until "the truth" about PhaseOne is known to all and PhaseOne is gone and buried forever (and maybe you won't be happy until they give some of their profits back to those who bought into the system). You make them sound like they are scammers. That anything but the exmor is junk. Meanwhile amazing photographs and prints are being made with lots of different gear worldwide right now.

Here some real world samples of a snap I mad while scouting a location a while back. I usd a IQ160 on a H1 with 80mm H lens handheld at 1/160 sec f2.8 iso 200. Processed in Lightroom which is not ideal, with just some quick and basic adjustments.

First is the whole frame then a 100% crop.

Looks good enough to me. Iso 200 is very very far from "worhtless" on the newer dalsa ccd sensors.

Could I have gotten the same result by using iso 50 and the same exposure and just increasing exposure in post? Of course, but I did not want to have a dark preview image in the rear lcd of the back.

[chrismuc]

I made a simple test with my IQ180: 3 exposures

- at ISO 35
- at ISO 35 1.5 stops underexposed (= exposed like ISO 100)
- at ISO 100

If I open the files in ACR and add 1,5 stops exposure to the second picture, I see the following:

- the three pictures look very similar, but
- the second and third are more grainy in medium and dark areas

so my conclusion would be that the IQ180 is

- ISO 35 native* (not ISO 100 like Void suggests)
- up to ISO 100 in reality also ISO 35, just underexposed and already compensated by Phase RAW firmware (which means the picture looks correct exposed if opened in ACR or C1, just raw digger and similar raw programs can detect), but has reduced DR compared to shot at ISO 35
- above ISO 100 basically the same like below, means the real sensitivity is 1,5 stops lower than read on the back and internally compensated (which of course reduces the DR)

End of the story: Shoot ISO 35 wherever possible to get the 13.5 steps DXO-DR (acc. my experience up to 8s exposure time, then degration of file quality), then maybe still stay at ISO 35, underexpose and push in RAW conversion .

* I also read the enclosed DXO measurement of ISO sensitivity that way