Best ISO for IQ280

[voidshatter]

If the counter attached to the pixel reaches 100% of what it can count, then you do have too many photons, and then this area has no detail.
More photons is always best, but only if you calibrate the counter (ISO) not to multiply so much it reaches 100%.

For the same aperture, the same shutter speed and the same light, ISO 100 of IQ280 will not give you more area of blown highlight when compared against ISO 35. ISO 200 however, will cause more area of blown highlight.

[voidshatter]

You always get better shadow recovery at lower ISOs.  Of course Sbernthal's observation about highlights and shooting at ISO 100 is very valid too.  

I need to see RAW files as hard evidence (for the case of ISO 35 vs ISO 100). Please supply RAW files of the same aperture, the same shutter speed and the same light, for ISO 35 and ISO 100.

[JoeKitchen]

I need to see RAW files as hard evidence (for the case of ISO 35 vs ISO 100). Please supply RAW files of the same aperture, the same shutter speed and the same light, for ISO 35 and ISO 100.

No, because I would not shoot like that.  It does not make sense.  

I would properly expose the image for the given ISO.  If you properly expose an ISO 35 image and compare it to a properly exposed ISO 100 image, the ISO 35 image gives you better shadows.  

Now of course if you under expose an ISO 35 image and push it, it will look very very similar to an properly exposed ISO 100, but there is no reason to shoot like that unless you do not have the proper equipment.  

By the way, have you considered graded neutral density filters?  I ask because Joe Cornish shoots images like the one posted, with as much of the image properly exposed as possible through the lens.  His images, at 100%, are very nice.  Nice micro contrast, nice colors.  Sure there are limits to those filters, but I would rather have most of the image properly exposed and only work on areas that got missed, instead of have the majority of the images not properly exposed and work on all of it. 

[voidshatter]

No, because I would not shoot like that.  It does not make sense.  

Then it makes no sense for you to shoot a Nikon D810 at anything above ISO 32 when you can shoot at ISO 32.

I would properly expose the image for the given ISO.  If you properly expose an ISO 35 image and compare it to a properly exposed ISO 100 image, the ISO 35 image gives you better shadows.  

Now of course if you under expose an ISO 35 image and push it, it will look very very similar to an properly exposed ISO 100, but there is no reason to shoot like that unless you do not have the proper equipment.  

You can shoot at ISO 35 for any "proper exposure" you want. I can shoot at ISO 100 for a brighter look of playback and then pull down exposure in post-processing and still achieve as good image quality as yours. You don't have advantage of image quality over me.

By the way, have you considered graded neutral density filters?  I ask because Joe Cornish shoots images like the one posted, with as much of the image properly exposed as possible through the lens.  His images, at 100%, are very nice.  Nice micro contrast, nice colors.  Sure there are limits to those filters, but I would rather have most of the image properly exposed and only work on areas that got missed, instead of have the majority of the images not properly exposed and work on all of it. 

Of course I use ND grad and I don't need a lecture on that. Indeed ND grad is vulnerable to non-straightline skyline scenes. For example, if the sun is lower than a building, then your ND grad will cut the building making the building look dark and you would still need to push the shadow out of the building in post-processing. Indeed ND grad will make any cloud above the sun to be underexposed, relying on the dynamic range of the sensor in post-processing.

Example of a scene where an ND grad would fail: https://500px.com/photo/92579427/

[JoeKitchen]

Then it makes no sense for you to shoot a Nikon D810 at anything above ISO 32 when you can shoot at ISO 32.

You can shoot at ISO 35 for any "proper exposure" you want. I can shoot at ISO 100 for a brighter look of playback and then pull down exposure in post-processing and still achieve as good image quality as yours. You don't have advantage of image quality over me.

Of course I use ND grad and I don't need a lecture on that. Indeed ND grad is vulnerable to non-straightline skyline scenes. For example, if the sun is lower than a building, then your ND grad will cut the building making the building look dark and you would still need to push the shadow out of the building in post-processing. Indeed ND grad will make any cloud above the sun to be underexposed, relying on the dynamic range of the sensor in post-processing.

Example of a scene where an ND grad would fail: https://500px.com/photo/92579427/

1.  Yes, that is how I would work, if that ISO was available on the given camera.  

2.  What if we both had proper exposures for the given iso, so the image looked as close as we wanted it to without any post work, but we needed to push the shadows.  Yours would look worse then mine.  

3.  Good example, however you are already under-exposing almost all of the image already.  Instead, why not properly expose most of the image and push the top part of the building to match the rest?  Easier work-flow in post?  (I could imagine it would be difficult to match the top of the building to the bottom.)

[voidshatter]

1.  Yes, that is how I would work, if that ISO was available on the given camera.  

Then do you think Nikon is silly to advertise ISO 32 as extended ISO for the D810?

2.  What if we both had proper exposures for the given iso, so the image looked as close as we wanted it to without any post work, but we needed to push the shadows.  Yours would look worse then mine.  

Your definition of "proper exposure" is based on the playback inside the camera - you think it is proper when you see a good histogram in the playback inside the camera/digital back. My definition of "proper exposure" is to check with RawDigger and verify that I have made ETTR as much as possible without blowing out the highlight details I want - I can shoot at ISO 100 for an overexposed look for the playback. My way will give best image quality out of a single RAW file (regardless of ISO 35 or ISO 100). Your way is not reliable and may waste dynamic range, because the histogram inside the camera/digital back is not reliable, and you may be wasting highlight recoverability.

3.  Good example, however you are already under-exposing almost all of the image already.  Instead, why not properly expose most of the image and push the top part of the building to match the rest?  Easier work-flow in post?  (I could imagine it would be difficult to match the top of the building to the bottom.)

It will not be easy task. This is long exposure and the light condition changes fast during the sunset process. If you shoot most of the frames with blown sky then there is no way to recover that in post-processing for the long exposure look. If you shoot a frame for the building before you start the long exposure process then the color temperature will be higher; if you shoot after then the color temperature will be lower. The light and shadow will also easily look abnormal when you compose two things taken at very different times.

[JoeKitchen]

Then do you think Nikon is silly to advertise ISO 32 as extended ISO?

Your definition of "proper exposure" is based on the playback inside the camera - you think it is proper when you see a good histogram in the playback inside the camera/digital back. My definition of "proper exposure" is to check with RawDigger and verify that I have made ETTR as much as possible without blowing out the highlight details I want - I can shoot at ISO 100 for an overexposed look for the playback. My way will give best image quality out of a single RAW file (regardless of ISO 35 or ISO 100). Your way is not reliable and may waste dynamic range, because the histogram inside the camera/digital back is not reliable, and you may be wasting highlight recoverability.

It will not be easy task. This is long exposure and the light condition changes fast during the sunset process. If you shoot a frame for the building before you start the long exposure process then the color temperature will be higher; if you shoot after then the color temperature will be lower. The light and shadow will also easily look abnormal when you compose two things taken at very different times.

1.  Don't know, don't use the camera.  I was not aware that that ISO was offered.  

2.  My definition of proper exposure is not based on the play back, but what I feel is the closest look I can get to what I want without any post work.  I almost always shoot tethered, so I do not really on the playback.  With that said, I see what you are saying now.  You are worried about the highlights, whereas I am worried about the shadows, especially with an interior.  Operating at ISO 100 would be better in your situation, the lowest ISO possible is best in my situation.  

3.  I meant within a single capture, mask out the under-exposed area of the building and brighten it to match the rest.  Not sure if this would be possible with a building, but with small areas of an almost smooth horizon, it may be plausible.  (I realize layering an image from seconds before or after at that time of day is not always possible.)  

And after thinking about it, if I were to use a ND Grad filter to get the sky properly exposed, I would need to worry about shadow details.  So, lower ISO would be best.  

[voidshatter]

1.  Don't know, don't use the camera.  I was not aware that that ISO was offered. 

Then you could advise Nikon and Canon to deliberately hide the fact that ISO 32 or ISO 50 are extended ISO - this makes users feel good, and ISO 32 of Nikon D810 beats ISO 35 of IQ280 in terms of advertised low ISO number as well as dynamic range.

2.  My definition of proper exposure is not based on the play back, but what I feel is the closest look I can get to what I want without any post work.  I almost always shoot tethered, so I do not really on the playback. 

If you shoot tethered, try to ETTR as much as possible without blowing out highlight by means of verification with RawDigger instead of Capture One (regardless of ISO 35 or ISO 100), then pulling down the exposure in post-processing, would give you better shadow, when compared against your feel.

3.  I meant within a single capture, mask out the under-exposed area of the building and brighten it to match the rest.  Not sure if this would be possible with a building, but with small areas of an almost smooth horizon, it may be plausible.  (I realize layering an image from seconds before or after at that time of day is not always possible.) 

Yes if you shoot an LCC with the ND grad on, then apply LCC correction in Capture One, then apply highlight recover globally on the image, you can essentially achieve the effect you are talking about. However the signal-to-noise ratio on that building will still suffer. This is also true for any cloud above the sun when you use ND grad to cover the sun. Joe Cornish does not deliberately shoot against the sun for long exposure.

[voidshatter]

And after thinking about it, if I were to use a ND Grad filter to get the sky properly exposed, I would need to worry about shadow details.

If you use an ND grad to cover the sun, then any cloud above the sun will be underexposed.

So, lower ISO would be best.  

Again, I need to see evidence by means of RAW files. ISO 35 indeed is best but not better than ISO 100.

[voidshatter]

 With that said, I see what you are saying now.  You are worried about the highlights, whereas I am worried about the shadows, especially with an interior.  Operating at ISO 100 would be better in your situation, the lowest ISO possible is best in my situation.  

No. I'm not saying ISO 100 is better in my situation. I'm saying ISO 100 is as good as ISO 35 when the user knows how to ETTR as much as possible to retain highlight details - this way gives best shadow recoverability. ISO 35 is best, but not better than ISO 100. ISO 100 is also best, but not better or worse than ISO 35.

[Paul2660]

This is from the DxO Mark scores, on the IQ180.  They do rate the Phase iso's of 35, 50 and  100 as all at iso 29. 

The problem as I see it is that photographers tend to use the same logic for iso increases on CCD and CMOS.  The issue on whether the iso on the Phase One CCD is really an increase? or is it just a metadata setting for C1 to use later on.  Unlike a CMOS iso increase which as I understand it, is an actual increase in gain on the chip. 

There have been several comparisons of IQ180 iso 35 and iso 200 shots.  I tried to find the one done a few years ago, but was not able.  But the net, was that the image looked the same between a under exposed iso 35 by 2.5 stops  and a correctly exposed.  The two images when viewed side by side look pretty much the same noise levels etc.  (Note I may have this backwards, i.e. the iso 35 was even exposed but the iso 200 was under exposed) 

I have put this question forward several times on various forums but it's never really been answered very clearly.  From reading various posts by users (as I have never found anything written officially by Phase One) is that you really don't change the gain on a CCD by pushing the iso but I am not qualified enough to know the answer. 

I my use, (P45+, IQ160, IQ260, and IQ180), I feel that you can "push" the CCD maybe one stop, So on my IQ260, that would be either iso 50 to iso 100 or iso 100 to iso 200 depending on where you feel the real base iso rests.  In normal exposure ranges say 1" to 1/250th most times I will not see much difference between iso 50 or iso 100 on my 260.  What makes the difference to me is say the difference in shutter speeds, say 1/60 vs 1/125 at F11.  Just that one shutter speed difference can make a big difference in the noise in the shadows.  CCD needs light and lots of it and trying to push it on a dim day or late/early morning to me is not a good solution vs taking a longer exposure. 

By the time you get to iso 200 on the IQ260 I feel that the shadows show quite a bit of noise and color loss, but areas that have good light can still allow for a very good clean shot, but I know I won't be able to push the shadows at all.  By iso 400, there is loss in color/sat thought the file.  Areas that were brightly exposed might allow for some recovery in post, but the loss in details can't be pulled back. 

Of course moving to sensor plus changes all of this un to around iso 800.  But the loss in resolution makes this something I rarely do.

I would love to see more information on this subject.

Paul Caldwell

[Ken R]

Then it makes no sense for you to shoot a Nikon D810 at anything above ISO 32 when you can shoot at ISO 32.

You can shoot at ISO 35 for any "proper exposure" you want. I can shoot at ISO 100 for a brighter look of playback and then pull down exposure in post-processing and still achieve as good image quality as yours. You don't have advantage of image quality over me.

Of course I use ND grad and I don't need a lecture on that. Indeed ND grad is vulnerable to non-straightline skyline scenes. For example, if the sun is lower than a building, then your ND grad will cut the building making the building look dark and you would still need to push the shadow out of the building in post-processing. Indeed ND grad will make any cloud above the sun to be underexposed, relying on the dynamic range of the sensor in post-processing.

Example of a scene where an ND grad would fail: https://500px.com/photo/92579427/

Nice image but looks a bit odd, too high unnatural dynamic range boosted shadows and it is easy to see an abrupt transition from sky to building (see crop attached). Kinda looks like an outline / halo. I mean it's your style and looks good but it is not the only style of photography.

[voidshatter]

Nice image but looks a bit odd, too high unnatural dynamic range boosted shadows and it is easy to see an abrupt transition from sky to building (see crop attached). Kinda looks like an outline / halo. I mean it's your style and looks good but it is not the only style of photography.

Please don't make it personal. I never said shooting backlit long exposure is the only style of photography. I only said such style is a good way to show off the dynamic range performance of the Sony IMX094 CMOS sensor for gearhead's self-satisfaction, because it heavily relies on gear which means not many other photographers will be able to do it properly. It simply rules out competition from any CCD user or Canon user.

There are of course advantages of 80MP CCD but I mostly see it inside a studio or for a low contrast scene outdoors where others have photographed many times already (because it does not rely on sensor DR performance heavily). For portrait in studio yes 80MP CCD gives you most joy for pixel-peeping, but for outdoors if you can bracket with a CCD then I can stitch with a CMOS and I can hardly see advantages of CCD other than an advantage in wide-angle compatibility (soon to be removed by larger version of Sony A7R-II's backside-illuminated CMOS sensor).

Regarding the halo effect in my image: that's the limitation of the software I used - it was a very old version ACR. If you push too much shadow in Capture One you also get the halo effect. It can be fixed by careful local adjustments if I had the time. Thanks for the criticism.

[JoeKitchen]

This is where online forums suck, and I was being part of the problem today. 

If everyone in this thread were to meet up at a bar and talk this out while looking at each other, I feel the tone of this discussion would have been different.  Words only account for 17% of communication, and, when we only see words, we tend to assume the most polarized meaning.  Not to mention, Voidshatter's first language is not English, which means I was probably mis-interrupting his intentions due to the nuances of the language, something that would not have happened in person. 

After looking over the posts again, overall I think Voidshatter had some good points, and I am going to assume you (voidshatter) were not being hostel. 

[gagemanning]

I never realized I would cause such a huge debate.  Honestly, it's hard for me to get into the technical stuff (pixel peeping, etc...).  My eyes are not attune to such testing.  I can definitely see when noise is present and that is why I asked the question if there was much difference between the 2 (also wanted to make sure it did not effect the dynamic range). 

Thanks everyone for there input.  I've learned a lot.   

gage

[Ken R]

This is where online forums suck, and I was being part of the problem today. 

If everyone in this thread were to meet up at a bar and talk this out while looking at each other, I feel the tone of this discussion would have been different.  Words only account for 17% of communication, and, when we only see words, we tend to assume the most polarized meaning.  Not to mention, Voidshatter's first language is not English, which means I was probably mis-interrupting his intentions due to the nuances of the language, something that would not have happened in person. 

After looking over the posts again, overall I think Voidshatter had some good points, and I am going to assume you (voidshatter) were not being hostel. 

I concur.

English is not my first language either and I understand that sometimes some things are lost in translation but voidshatter has a quite a history of posting in this and other forums and he has a habit of hammering in his point of view no matter what and has quite a bit of underlying hate towards PhaseOne (other posters have this as well), on purpose or not it is there in a lot of his posts.

Does voidshatter speak the truth about some aspects of sensor performance? Yes, never said otherwise. It is just that he reduces photography gear and photography to just a few parameters and basically anything else than the Sony sensors is junk.

His discourse turns productive people away from posting in this and other forums. Not his information and tests which are nice but the tone and overall online demeanor. That is the personal aspect. You need at least some amount of cordiality and tact. Yes, even when one is behind a computer.

He might better serve a digital sensor / camera company behind the scenes in the technical development department. Back of house, not front of house.

[sbernthal]

This is where online forums suck, and I was being part of the problem today. 

After looking over the posts again, overall I think Voidshatter had some good points, and I am going to assume you (voidshatter) were not being hostel. 

I don't think you were part of the problem, and you are being very positive towards this guy who doesn't deserve it.

A few quotes from this thread:

"For your use case in the studio, the ISO 35 setting is merely a "+1.5EV exposure compensation for flash" for people like you who have no clue about what is called native ISO." (He edited this out later but certainly never apologized for it and pretty much denied it)

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

Why would you dismiss such comments? I wouldn't.

[ErikKaffehr]

Hi Paul,

My take is that there is not a lot of difference between modern CMOS and CCD in that sense. Both are mostly ISO-less. There is some gain in image quality with increasing ISO versus underexposure but it is very small. Canon cameras and Nikon D4 work differently, probably having a lot more analogue gain.

The Sony A7s has a couple of tricks in it's bag, helping at very high ISO's.

My take is that CMOS is improved in two areas, much reduced readout noise and a significant gain i well capacity, that is how many electron charges a pixel can hold.

Best regards
Erik

This is from the DxO Mark scores, on the IQ180.  They do rate the Phase iso's of 35, 50 and  100 as all at iso 29. 

The problem as I see it is that photographers tend to use the same logic for iso increases on CCD and CMOS.  The issue on whether the iso on the Phase One CCD is really an increase? or is it just a metadata setting for C1 to use later on.  Unlike a CMOS iso increase which as I understand it, is an actual increase in gain on the chip. 

There have been several comparisons of IQ180 iso 35 and iso 200 shots.  I tried to find the one done a few years ago, but was not able.  But the net, was that the image looked the same between a under exposed iso 35 by 2.5 stops  and a correctly exposed.  The two images when viewed side by side look pretty much the same noise levels etc.  (Note I may have this backwards, i.e. the iso 35 was even exposed but the iso 200 was under exposed) 

I have put this question forward several times on various forums but it's never really been answered very clearly.  From reading various posts by users (as I have never found anything written officially by Phase One) is that you really don't change the gain on a CCD by pushing the iso but I am not qualified enough to know the answer. 

I my use, (P45+, IQ160, IQ260, and IQ180), I feel that you can "push" the CCD maybe one stop, So on my IQ260, that would be either iso 50 to iso 100 or iso 100 to iso 200 depending on where you feel the real base iso rests.  In normal exposure ranges say 1" to 1/250th most times I will not see much difference between iso 50 or iso 100 on my 260.  What makes the difference to me is say the difference in shutter speeds, say 1/60 vs 1/125 at F11.  Just that one shutter speed difference can make a big difference in the noise in the shadows.  CCD needs light and lots of it and trying to push it on a dim day or late/early morning to me is not a good solution vs taking a longer exposure. 

By the time you get to iso 200 on the IQ260 I feel that the shadows show quite a bit of noise and color loss, but areas that have good light can still allow for a very good clean shot, but I know I won't be able to push the shadows at all.  By iso 400, there is loss in color/sat thought the file.  Areas that were brightly exposed might allow for some recovery in post, but the loss in details can't be pulled back. 

Of course moving to sensor plus changes all of this un to around iso 800.  But the loss in resolution makes this something I rarely do.

I would love to see more information on this subject.

Paul Caldwell

[Steve Hendrix]

I don't think you were part of the problem, and you are being very positive towards this guy who doesn't deserve it.

A few quotes from this thread:

"For your use case in the studio, the ISO 35 setting is merely a "+1.5EV exposure compensation for flash" for people like you who have no clue about what is called native ISO." (He edited this out later but certainly never apologized for it and pretty much denied it)

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

Why would you dismiss such comments? I wouldn't.

I agree - I'm not sure how one can attribute "people like you who have no clue" to language nuances...


Steve Hendrix
Capture Integration

[ErikKaffehr]

Hi Steve,

I definitively see your point. Thanks for chiming in, but I feel that you would be a very good person to put things straight for the original poster. So I hope you share your knowledge on the issue…

Best regards
Erik

I agree - I'm not sure how one can attribute "people like you who have no clue" to language nuances...


Steve Hendrix
Capture Integration