H3DII-31 1600ISO sample

[nikf]

Yesterday I was out with the mountain bike and my H3DII-31 in a back pack. The weather and light conditions were horrible here in Berlin.
Good conditions to test 1600 ISO with this camera and the new firmware update. I think it gives a good idea what ISO1600 looks like.
This is really just a snapshot, handheld with the 50-110 zoom. It's not meant as anything of photographic value. Standard settings in Phocus,
moderate sharpening, minimal EV correction.
Maybe it has some informational value for some. Full shot and 100% crop:

[Dustbak]

Not bad (even though the lighting appears to be pretty forgiving). Now, if I could just get the ISO800 on my CF39 I would be very happy. Even more so if I would get my 64sec's exposure time.

[Panopeeper]

This shot has *absolutely nothing* to do with the demonstration of the high ISO capability of the camera. Read http://luminous-landscape.com/forum/index....c=32410&hl= to understand the connection between noise, ISO and exposure.

[jimgolden]

looks good
...any chance we could see a 100% crop of the lower right darker area?

[David Grover / Capture One]

This shot has *absolutely nothing* to do with the demonstration of the high ISO capability of the camera. Read http://luminous-landscape.com/forum/index....c=32410&hl= to understand the connection between noise, ISO and exposure.

Regardless of the mathematics, most photographers want to know IF they set their camera to ISO1600, will it

A - Give an acceptable result, without the need for manipulation in software to 'correct' the exposure
B - Show an acceptable image on the LCD screen at correct exposure
C - Show an acceptable histogram at correct exposure

Whatever is happening behind the scenes to most is not useful if the above conditions are met.

...and of course how camera A compares to camera B with similar ISO ratings.

Best,




David

[jjj]

This shot has *absolutely nothing* to do with the demonstration of the high ISO capability of the camera. Read http://luminous-landscape.com/forum/index....c=32410&hl= to understand the connection between noise, ISO and exposure.

Somewhat of an overreaction!  
Particularly as this is how 99% of people will use a camera, i.e. setting an ISO and taking a photograph at the 'correct' exposure, it seems quite a sensible thing to do. Do you take photographs or just analyse data?
I used a H3-39D and it was pretty awful at 400ISO. This image shows a marked improvement on the 400ISO.
Shame the new back doesn't do 1600ISO.

[nikf]

This shot has *absolutely nothing* to do with the demonstration of the high ISO capability of the camera. Read http://luminous-landscape.com/forum/index....c=32410&hl= to understand the connection between noise, ISO and exposure.

My example was not meant as "demonstration of high ISO capabilities" - and I have nowhere stated such a remark. I'm a scientific educated and thinking human being and therefore I wouldn't
show one example as something demonstrating complex technical interactions. I'm not a camera 'researcher' but a photographer interested getting some practical applications of technology.
Therefore higher ISO capabilities are of value for me even in daylight  to extend the ranges of shooting opportunities, e.g. I can stop down a bit more and not worry about excessive noise.
I know the H3DII-31 is no D3 (which I have too) but it seems the firmware upgrade and the processing in Phocus extends the shooting possibilities for me in ISO ranges I hesitated to use before that.
If more people are convinced that my informal test is worthless and it's a waste of bits and time we can delete this. I was not sure to post it anyway because I have no posting history here and people will
try to jump on me to 'educate' me ...

[jjj]

Regardless of the mathematics, most photographers want to know IF they set their camera to ISO1600, will it
A - Give an acceptable result, without the need for manipulation in software to 'correct' the exposure...<snip>

Isn't that what the 50MP back does to get 800ISO? I'm sure the lady at Hasselblad told me that you needed Phocus to get 800ISO when trying back out the other day.
And with the test conditions, native 1600ISO would have been very useful.

[Dustbak]

This shot has *absolutely nothing* to do with the demonstration of the high ISO capability of the camera. Read http://luminous-landscape.com/forum/index....c=32410&hl= to understand the connection between noise, ISO and exposure.

If I would have been in the 'sensor-testing industry' this remark might have been very usefull to me but I am in the photograph-making industry and as an example of a camera that can now actually make any image at ISO1600 it is great. It previously could only do ISO800 max.

I think that is a fantastic demonstration of ISO capability don't you?...

Now, I do assume ISO1600 produces better results than underexposing at 800 and than raising in post, etc... Which is probably where you are coming from, but for many the possibility of using it during the process of photography itself is preferable than underexposing 800. Even if it is just a mere mental thing.

[csp]

still not canon ;-)

[Carsten W]

still not canon ;-)

That is exactly why people are interested.

[Guy Mancuso]

Regardless of the mathematics, most photographers want to know IF they set their camera to ISO1600, will it

A - Give an acceptable result, without the need for manipulation in software to 'correct' the exposure
B - Show an acceptable image on the LCD screen at correct exposure
C - Show an acceptable histogram at correct exposure

Whatever is happening behind the scenes to most is not useful if the above conditions are met.

...and of course how camera A compares to camera B with similar ISO ratings.

Best,




David

 I agree it is what the heck I am looking at is what counts and what my clients think. At some point your throw all the technical crap in the big garbage can and look at the image. I sell images to clients and not the science of it. If it looks good to me than i know my client will be happy , the prints will look good and God willing they actually pay me. LOL

Okay to the image it looks pretty good , I do see some color noise though and maybe some tweaking to it may improve upon it. Obviously we all would like to see more and also would like to see it compared to my Phase back but even though I am a Phase shooter it is nice to see Hassy get this in play. Seriously I have no issues with the science and the technical details of a image and everyone wants to know the back door issues and solutions to make the images better. And frankly i don't care how much we have to cheat to get there as long as the image looks good who cares how we get there.

[Panopeeper]

I don't understand you guys regarding this ISO and noise issue.

I downloaded the image and looked at it in PS. My estimation is, that the darkest part of the underside of that grayish-brownish roof is around the 5th EV of the dynamic range. (The railing is too small to judge anything on it.)

Now, even if the camera does not have ISO 1600, not 800, not even 400, this area would be between the 8th EV at ISO 200. Is *that* expected to be noisy with a Hasselblad back? Is the dynamic range of this back 9 EV at ISO 200? And then we have not even mentioned the ISO gain (assumed, that ISO 1600 is real, i.e. it is associated with a specific analog gain).

The high ISO capability has to be judged (if not measured) on the *shadows*, not on well-lit areas.

Although I find it highly suspect, that higher ISO will be introduced by *firmware*; that must be some miraculous firmware. I suspect this ISO 1600 is simply a shortcut for +1 EV in the raw processing, while the real advantage is, that the in-camera display gets brighter than it would be with the underexposed ISO 800.

Is the raw file of this shot available in DNG format?

[lisa_r]

I agree it is what the heck I am looking at is what counts and what my clients think. At some point your throw all the technical crap in the big garbage can and look at the image. I sell images to clients and not the science of it. If it looks good to me than i know my client will be happy , the prints will look good and God willing they actually pay me. LOL

Panopeeper, I think you miss the big picture most of the time because you are looking at it like a science project instead of a tool which people use to produce photographs, and make prints and such. This example works well for me, and gives some idea of what this camera is now capable of at high ISO.

In my opinion, the test as to whether this camera is going to work at 800 ISO does not take calculators, charts, etc. For the most part I would shoot some images like nikf has done - in various conditions - look at the highlights, shadows, etc. and then I would make some prints at appropriate sizes. Do they look good or not??

Would my clients accept the given image's IQ? Er, yes. Is this useful information? Indeed.

[nikf]

looks good
...any chance we could see a 100% crop of the lower right darker area?

Sure - here it is - same NR settings, which are simply all at '0' (I guess there is NR doing its work nonetheless in some form).
When fiddling around with the NR parameters the result can be of course fine tuned a lot. But as I said - it's not meant as a formal test.

[David Grover / Capture One]

Is the raw file of this shot available in DNG format?

Then you are looking at a DNG file, not a 3F file.

[David Grover / Capture One]

My example was not meant as "demonstration of high ISO capabilities" - and I have nowhere stated such a remark. I'm a scientific educated and thinking human being and therefore I wouldn't
show one example as something demonstrating complex technical interactions. I'm not a camera 'researcher' but a photographer interested getting some practical applications of technology.
Therefore higher ISO capabilities are of value for me even in daylight  to extend the ranges of shooting opportunities, e.g. I can stop down a bit more and not worry about excessive noise.
I know the H3DII-31 is no D3 (which I have too) but it seems the firmware upgrade and the processing in Phocus extends the shooting possibilities for me in ISO ranges I hesitated to use before that.
If more people are convinced that my informal test is worthless and it's a waste of bits and time we can delete this. I was not sure to post it anyway because I have no posting history here and people will
try to jump on me to 'educate' me ...

Thanks for taking the time to post an ISO1600 shot from the H3D31.  No science lesson needed from or for me!

Cheers,

David

[David Grover / Capture One]

Although I find it highly suspect, that higher ISO will be introduced by *firmware*; that must be some miraculous firmware. I suspect this ISO 1600 is simply a shortcut for +1 EV in the raw processing, while the real advantage is, that the in-camera display gets brighter than it would be with the underexposed ISO 800.

If it was THAT simple we would have saved ourselves (and most importantly our customers) a lot of trouble and released it a long time ago!

Not that cut and dry I am afraid.

[csp]

everyone who is interested in 1600 should check out RD as raw converter. the grain RD renders is very fine much nicer to my eyes than  phocus which even mit luminace NR set to -10 produces a very slight painterly look.



david how many stops in dynamic range is lost @ 1600 ?

[David Grover / Capture One]

I couldn't tell you right now actually.  I need to have a bit more time with it to accurately state!

David