DP2m ISO test

Started Aug 14, 2012 | Discussions
HBowman
HBowman Senior Member • Posts: 1,237
DP2m ISO test

The test is basically done in darkness, the back of my plasma TV at home, in very dim light. I opened in SPP 5.3 and touched nothing. Exported as 16 Bit Tiff Prophoto (my work-flow is in Prophoto). I opened it in LR4 where is reduced 100% Chrominance, and switched directly in B&W without touching anything. Export as JPG sRVB. Photo reduced in a basic free application. I done +0.7 bias to compensate any side effects, all in P mode.

The real scene where the test took place, the back of the TV.

ISO 100

ISO 200

ISO 400

ISO 800

ISO 1600

ISO 3200

ISO 6400

What can I say ? It is good, at least for me The real problems come at 1600 ISO, at a colour point of view. There is a huge break between 800 and 1600 ISO in term of colour rendering and chrominance.

The banding, horizontal in my case, start to be really present at 3200 ISO ... in this very test .

I will investigate more and will chase this banding to find the exact conditions, where and why he is generated at lower ISO.
--
http://www.hulyssbowman.com

Sigma DP2
If you believe there are incorrect tags, please send us this post using our feedback form.
hawatt New Member • Posts: 4
Re: DP2m ISO test

would like to see this test in color

HBowman
OP HBowman Senior Member • Posts: 1,237
Re: DP2m ISO test

hawatt wrote:

would like to see this test in color

Here we go.

ISO 800 is awesome, as clean as the 400. no pp needed, at all.

ISO 1600 without pp (here is THE gap ...).

ISO 1600 with my PP.

ISO 3200 without pp

ISO 3200 with my PP.

ISO 6400 without pp.

ISO 6400 with my PP.

I can have some good result out of ISO 1600 and 3200 files but it need some PP skills and I understand we are not all equal front a computer.
--
http://www.hulyssbowman.com

PrebenR Veteran Member • Posts: 4,164
Re: DP2m ISO test

Wow! B&W high ISO looks very very promising!

-- hide signature --

AF takes away the Zen in photography

victorgv
victorgv Senior Member • Posts: 2,013
Re: DP2m ISO test

very big problems in shadows on high iso details just disapear.
For example look on the computer back.
http://forums.dpreview.com/forums/read.asp?forum=1027&message=42226729
http://forums.dpreview.com/forums/read.asp?forum=1027&message=42216763

 victorgv's gear list:victorgv's gear list
Sigma DP2 Merrill Sigma DP1 Sigma dp2 Quattro Sigma sd Quattro H
Johan Borg Veteran Member • Posts: 3,218
Re: DP2m ISO test

HBowman wrote:

The banding, horizontal in my case, start to be really present at 3200 ISO ... in this very test .

I will investigate more and will chase this banding to find the exact conditions, where and why he is generated at lower ISO.

Could you compare a regular ISO 800 and ISO 1600 shot with one taken with the very same shutter speed and aperture but at ISO 200 and then pushed back up to 800/1600 brightness in post? My gut feeling says you will get a different noise signature from that, but I may be wrong

 Johan Borg's gear list:Johan Borg's gear list
Sigma dp2 Quattro Sigma fp Sigma 45mm F2.8 DG DN
HBowman
OP HBowman Senior Member • Posts: 1,237
Re: DP2m ISO test

Johan Borg wrote:

HBowman wrote:

The banding, horizontal in my case, start to be really present at 3200 ISO ... in this very test .

I will investigate more and will chase this banding to find the exact conditions, where and why he is generated at lower ISO.

Could you compare a regular ISO 800 and ISO 1600 shot with one taken with the very same shutter speed and aperture but at ISO 200 and then pushed back up to 800/1600 brightness in post? My gut feeling says you will get a different noise signature from that, but I may be wrong

Hi Joan !

I will try that tomorrow
--
http://www.hulyssbowman.com

PrebenR Veteran Member • Posts: 4,164
Re: DP2m ISO test

victorgv wrote:

very big problems in shadows on high iso details just disapear.
For example look on the computer back.
http://forums.dpreview.com/forums/read.asp?forum=1027&message=42226729
http://forums.dpreview.com/forums/read.asp?forum=1027&message=42216763

Have you tried to convert to monochrome?
--
AF takes away the Zen in photography

balchinian Regular Member • Posts: 294
Re: DP2m ISO test

PrebenR wrote:

Wow! B&W high ISO looks very very promising!

I agree. I'm tempted to try RGB filters in front of the lens, with monochrome, and then reassembling the full color image digitally. I suspect that might give a much better looking image. Maybe it will reduce or remove banding.

PrebenR Veteran Member • Posts: 4,164
Re: DP2m ISO test

My experience is that converting to monochrome wb in post in SPP wprks very well.
--
AF takes away the Zen in photography

victorgv
victorgv Senior Member • Posts: 2,013
Re: DP2m ISO test

Kind of, details were gone anyhow dark areas below certan level become one solid tone even on lowest noise reduction.

 victorgv's gear list:victorgv's gear list
Sigma DP2 Merrill Sigma DP1 Sigma dp2 Quattro Sigma sd Quattro H
Raist3d Forum Pro • Posts: 44,659
This Isnt really a good high ISO test

Ill explain why-

I realize this is all work so sorry to call on a couple of points here. To your good credit thanks for

Picking a situation that required low shutter speeds. The amount of times I see high ISP testing with more than a lot of light aroun on the web is way too
Much and doesn't prove much of anything either

The problem I am seeing with the test is that the subjects you picked are pretty much all on or off. Written black or somewhat shiny white. Not much tones in between

The second problem is, if I understood correctly, you are showing like 800x600 or smaller shots out of a resize down of 15 megapixels. Any camera benefits immensely from such a big resize. In fact I would be willing to bet I could take that subjec with the W and it may even look better with the Q due to no banding

My suggestion is go outside to the streets an try high ISO shots there in light that requires 1/30 f2.8 ISo 1600-6400 and see the difference. Then post at least two megapixel style shots (1600x1xxx).

Once again I need this is more work but honestly
I don't think the test shows
Much of a good high ISO for the camera (or any Camera tested
The same way)

-- hide signature --

Raist3d/Ricardo (Photographer, software dev.)- "You are taking life too seriously if it bugs you in some way that a guy quotes himself in the .sig quote" - Ricardo

PrebenR Veteran Member • Posts: 4,164
Re: This Isnt really a good high ISO test

Don't understand why doing a shot at 1/10s at 6400 isn't a good test...

-- hide signature --

AF takes away the Zen in photography

HBowman
OP HBowman Senior Member • Posts: 1,237
Re: This Isnt really a good high ISO test

I do think the test show what the camera is about. Obviously, the problems are bigger at bigger size but the point is to see what type of noise we could expect in normal actuation of the camera. I'm not here to please PP technician; I just give the average rendering we can expect and I'm pretty much clear that saying that the camera fall off at 1600 ISO. Punkt/

Comparing the DP2m to ... the Pentax Q is, at best, ridiculous. Stop with this ! It is boring y a know ? WTF !!! No offence attended though, this is just to be clear. And I will not even take time to read your posts if you continue speaking about this thing. If I see Q or Pentax Q when I diagonal read your future posts, I will skip them.

I speak about the DP2m and the DP2m isn't a night street reportage camera, at all, and that don't need a master degree in parapsychology to be understood.

I hope I'm clear about this point. Yes we can do B&W in street BUT this camera is about ultimate IQ at base ISO : 100 to 400, 800 in studio under controlled and homogeneous light. For now that's all we can expect from it and I think it is a very good evolution of the foveon sensor. Over 800 it need some skills in PP, but the quality fall off very deep especially at a colour level. But that ... every user of this technology know it, at least I hope for them.

I use him as he should be used to give his best results, like I do at a pro level since the first DP to the SD1 and now, the DPM. As all other DP/SD they shine in good light and with long exposures, at base ISO if possible.

My post is not finished because I investigate the banding in low ISO and this is, yes, a load of work. I do this for every one of us who already have the camera, and the future users who will have the camera.

Please do not reply by quoting every of my phrases, I find that boring too and it tend to excite me. So I skip message done in this way. I took time to write this post to be clear about ALL points. I will not post such long post for a very LOOONNGG time. Please, if you want to reply, try to be as concise as possible. Thank you.

PS: I do this also for SIGMA staff who read this forum too. Especially the issue concerning banding at LOW ISO settings. This is a very important point to be examined and it need some extensive tests in controlled environment.

-- hide signature --
PrebenR Veteran Member • Posts: 4,164
Re: This Isnt really a good high ISO test

HBowman wrote:

I do think the test show what the camera is about. Obviously, the problems are bigger at bigger size but the point is to see what type of noise we could expect in normal actuation of the camera.

Thank you for the test! Just to be clear is the photos you showed crops of original photos or resized down from original photo?

-- hide signature --

AF takes away the Zen in photography

PrebenR Veteran Member • Posts: 4,164
Re: This Isnt really a good high ISO test

PrebenR wrote:

Thank you for the test! Just to be clear is the photos you showed crops of original photos or resized down from original photo?

Sorry, I didn't read first post clearly enough.
--
AF takes away the Zen in photography

rf-design
rf-design Contributing Member • Posts: 619
Re: DP2m ISO test

That is a perfect approach for long time exposure.

Foveon has the very specific disadvantage that the 3-layer-silicon spectral sensitivity is not matched to the sRGB color representation. So in camera for jpg and in SPP there is a need for a color conversion matrix operation. This matrix have strong and also partly negative off-diagonals numbers. So seen in sRGB the noise is amplified w/o getting more signal. An ideal sensor/display combination for foveon would be a display with a white background light and three different thin layers of silicon before the LCD shutter.

But using RGB filters in front in a sequential mode allows you also to change the spectral sensitivity and to adapt to a possible color temperature shifted lighting source w/o noise impact of the color temperature correction.

Did you tried this for long time exposure?

balchinian wrote:

PrebenR wrote:

Wow! B&W high ISO looks very very promising!

I agree. I'm tempted to try RGB filters in front of the lens, with monochrome, and then reassembling the full color image digitally. I suspect that might give a much better looking image. Maybe it will reduce or remove banding.

 rf-design's gear list:rf-design's gear list
Panasonic LX100 Sigma dp0 Quattro Sigma DP1 Sigma SD14 Sigma SD15 +68 more
richard stone Veteran Member • Posts: 3,472
Thanks

Ridiculous and boring pretty much sum it up.

You could use the DP2M for night pictures and for street work and night street images. And the images could be excellent and unique. But that would take a lot of talent and desire, maybe by someone who had no other camera available, and is not what the camera is designed for or best at.

For that matter, the next thing we see could be a superb sport image. From someone with talent and determination. The balloon images were striking.

The camera and Foveon sensor produce subtle and exquisite images and color. The portraits I have seen so far are striking too.

My small gallery: http://www.pbase.com/richard44/inbox

 richard stone's gear list:richard stone's gear list
Sigma sd Quattro Sigma SD10 Sigma 17-50mm F2.8 EX DC OS HSM Sigma 30mm F1.4 DC HSM Art
PrebenR Veteran Member • Posts: 4,164
Re: Thanks

Who are you replying to?
--
AF takes away the Zen in photography

Raist3d Forum Pro • Posts: 44,659
Re: This Isnt really a good high ISO test

PrebenR wrote:

Don't understand why doing a shot at 1/10s at 6400 isn't a good test...

I believe I explained why. The light intensity is right but the subject and mega resize down is not. Try this on another real world subject and you'll see. There's another poster that got much worse results than this with slightly more intense light. Why do you think that is?

Not to mention the banding and the resize down this has. And forget about color

-- hide signature --

AF takes away the Zen in photography

-- hide signature --

Raist3d/Ricardo (Photographer, software dev.)- "You are taking life too seriously if it bugs you in some way that a guy quotes himself in the .sig quote" - Ricardo

Keyboard shortcuts:
FForum MMy threads