Jump to content

I can't find the noise...


jaapv

Recommended Posts

Advertisement (gone after registration)

I have been experimenting with ISO 2500. The only reason this image is a bit soft is because it was handheld at 1/3rd. Summicron 35asph. But where is the noise? I add a crop. Standard C1 developing and no NR software....The only secret is to keep that histogram where it should be. In the middle. This thing is really better than the M8.

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

Link to post
Share on other sites

I have been experimenting with ISO 2500. The only reason this image is a bit soft is because it was handheld at 1/3rd. Summicron 35asph. But where is the noise? I add a crop. Standard C1 developing and no NR software....The only secret is to keep that histogram where it should be. In the middle. This thing is really better than the M8.

 

I totally agree that the M9 represents a significant improvement over the M8 in the domain of high ISO noise.

Link to post
Share on other sites

This particular shot, @ 100%, reminds me of a 100% view of Digilux 2 ISO 200 jpegs. Simultaneously noisy (mostly luminance speckles) and yet with some "fuzzing" due to in-camera noise processing. (Yes, even a DNG can have some processing done to it in-camera - witness Leica's vignetting corrections).

 

Of course with 3.8x the pixels, at 3.6 stops higher ISO, from a sensor 16x the area - so obviously not comparable otherwise...

 

The color noise looks clean. No streaks. Shows what correct exposure can do.

 

Compared to M8 - I'd prefer to see identical exposures side-by-side, but yes, lower "streakiness" makes the difference between "usable" and "damaged".

 

Compared to Canon 5DII - a bit softer, but without the "burlap-weave" texture that shows up when any sharpening is applied to a 5DII CMOS image @ ISO 1600 or above. Both look processed - in different ways.

 

I hope one of Leica's firmware changes is an "on-off" setting for the high ISO noise smoothing I think they have added in the M9. Sometimes smooth is good - sometimes I'd prefer "grittier but sharper" - especially in B&W shots. A choice would be best of all.

Link to post
Share on other sites

I have been experimenting with ISO 2500. The only reason this image is a bit soft is because it was handheld at 1/3rd. Summicron 35asph. But where is the noise? I add a crop. Standard C1 developing and no NR software....The only secret is to keep that histogram where it should be. In the middle. This thing is really better than the M8.

 

Looks good. But on a tripod it would have been incredibly good. I do think as well that the slight movement masked the noise a bit. Can you repeat this shot on a tripod with both, your M9 and M8. At the same apertures, speed and no editing except for size what so ever? Sorry if the request is a bit too long... :o

Link to post
Share on other sites

Advertisement (gone after registration)

Jaap, there are several "grain" sharpeners in Photoshop. If you can't see the noise, you could use one of them. :p

 

Seriously, good illustration. It will be interesting to learn more about the M9's high-ISO noise reduction, if present.

Link to post
Share on other sites

I have been experimenting with ISO 2500. The only reason this image is a bit soft is because it was handheld at 1/3rd. Summicron 35asph. But where is the noise? I add a crop. Standard C1 developing and no NR software....

Not sure what C1 version you've used but the standard settings of mine (C1v4) include noise reduction already.

Link to post
Share on other sites

Looks good. But on a tripod it would have been incredibly good. I do think as well that the slight movement masked the noise a bit. Can you repeat this shot on a tripod with both, your M9 and M8. At the same apertures, speed and no editing except for size what so ever? Sorry if the request is a bit too long... :o
Sorry - no. The snow has melted :(
Link to post
Share on other sites

Camera shake can't mask noise, because it's not 'out there', it's inside the camera!

 

The old man from the Age of Tri-X

 

Is that true? When I first made my comment it was just sort of spontaneous, with no thought. Then I thought as you remark above. But then I thought camera shake causes a light ray to impact over more pixels. A dark area, where noise is normally evident, might during the exposure receive light from a nearby lighter area.

 

Perhaps I should just go and watch telly.

 

Jeff

Link to post
Share on other sites

I've done some test shots, on a tripod, in low room light at 1250 and 2500 on both the M9 (with 28/2.0) and the M8u (using my Zeiss 21/2.8 coded as an Elmarit pre-asph). Same field of view, so the M9 gets an advantage by putting more pixels behind the same image. Will post full frames reduced to fit and 100% crops later tonight, after dinner.

 

The sharpening has to be set off in C1 on the process dialog with a checkbox. Noise sliders, too, but in a different place. I'll double-check to make sure that all are zeroed. I used AWB. For real shots, I would correct each a bit.

 

scott

Link to post
Share on other sites

Here are some side-by-side M9 to M8 comparisons. First at ISO 1250, shots in same light, same exposures, f/5.6 @ 1/45.

 

M9 full frame:

 

L4001981_1websize.jpg

 

and M8 full frame:

 

L1017300_1websize.jpg

 

now a 100% crop from the M9 image:

 

L4001981_1crop.jpg

 

and the same image (but fewer pixels) from the M8:

 

L1017300_1crop.jpg

 

M8 a little lighter than the M9, as others have pointed out.

 

scott

 

(next ISO 2500)

Link to post
Share on other sites

Now the same sequence, but at ISO 2500:

 

M9 full frame:

 

L4001982_1websize.jpg

 

M8 full frame

 

L1017301_1websize.jpg

 

M9 100% crop

 

L4001982_1crop.jpg

 

M8 100% crop

 

L1017301_1crop.jpg

 

scott

 

(P.S. Yes there is a vertical line of white pixels in the M8 shots, which appear sometimes in high ISO with lightsources in the frame. That will need discussion with Solms, not germane to this comparison.)

Link to post
Share on other sites

Great!!! Ok, now I have a question on the ISO 1250 are both shots at the same aperture, speed and lens? If it is so, can you fix or put both on exactly the same temperature setting (WB)?

 

I would like to see if the reason why the low-light pictures on the M9 look darker or have less detail are due to Software and AWB settings.

 

Many, many thanks!

Link to post
Share on other sites

Now the same sequence, but at ISO 2500:

 

 

 

(P.S. Yes there is a vertical line of white pixels in the M8 shots, which appear sometimes in high ISO with lightsources in the frame. That will need discussion with Solms, not germane to this comparison.)

 

 

 

I get that vertical line on shots like this on my M8 as well. I've seen some describe it as a defect and others as an unavoidable "response" to harsh light in a low light situation - similar to the "green blob" effect in response to strong light at edges of an image's frame. With that issue, I've seen most people just live with the green blob because you can position light in the frame in a manner to reduce the likelihood of green blobs - I'm not aware of any method to minimize the vertical line issue in low light shots like this so I'm very interested in a proper fix for it.

 

Do the techs at Solms consider the vertical line to be a fixable sensor defect and if so, do they fix this on their own dime or does the poor M8 user just pay a hefty bill to repair? As most M8's are off warranty now, does anyone know what Leica charges the M8 user to conduct this repair?

 

I have a M9 on order and if it ever arrives, I'd consider sending my M8 in to service to correct this if I hear that Solms has a fix for this vertical line problem. I've been reluctant to send in my only Leica body for a long service but will look to do so once I have the M9. Glad I don't see that in the M9 frames!

 

Cheers

Link to post
Share on other sites

Both shots were taken at f/5.6 and aperture priority. The M9 chose 1/40 sec; the M8 chose 1/45 sec (yet it is a bit lighter). So there is a definite difference in sensitivity,somewhat countered by the fact that the M9 is more center-weighted, and thus more influenced by the bright blown-out spot in the center of the picture.

 

Now, let's look at color temperatures. The M9 comes up "as shot" indicating 5722 degrees with a magenta correction of +7 in Capture One 5.0.2. If I click-balance on the white paper around the creature drawn in the upper left, I shift to 4150 +20, and the colors of the blue sweater and the wooden surfaces in the room are closer to how I perceive them.

 

The M8 shot coes up as 4364 +9. After click-balancing on the same piece of paper, I get 3918 +10, which is a much smaller change. The M8 seems to have gotten closer with AWB than the M9.

 

The resulting M9 picture:

 

L4001981_2websize.jpg

 

and the M8 picture:

 

L1017300_2websize.jpg

 

The indicated color temperatures and offsets are somewhat arbitrary, since there are also effects of the different profiles involved. If I balanced each to the same temperature and shift, they would look more different.

 

scott

Link to post
Share on other sites

Archived

This topic is now archived and is closed to further replies.

  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...