Jump to content
Sign in to follow this  
nugat

M8 with 2.0 firmware high ISO better noise performance?

Recommended Posts

Advertisement (gone after registration)

I think this issue deserves a separate thread. I moved my post here and invite all to do more tests. The minimum of the para-scientific method is to compare firmwares with all other parameters unchanged. Take low light 2500/1600 ISO pictures with 1.2.1 firmware, then update to 2.0 and take the same pictures again with other parameters unchanged. Don't tell us what pictures are with what firmware (blind test).

I am also curious about the "Noise reduction" message and countdown when taking pictures with shutters longer than 2 sec. Any gives?

Piotr

PS . Dear moderator, I cannot delete the old post somehow, so there's a duplicate ("edit" option disappears?)

 

I am new to M8 (third day, bought used) but also tried some quick and dirty comparisons for high ISO noise before and after FW update.

I took some pictures through my living room onto the dark dining annex. The lights were three floor lamps and one ceiling with "old fashioned" 60 W bulbs. These bulbs imitate the early 20th century electric lighting, they are dim and very warm. There is also a burning fireplace (damn cold September!). We like our lighting really low for a glass of wine and good jazz. So yes, by all standards it's really dark yet contrasty. A,B, C raw originals were given only auto WB and auto development in Lightroom2. The camera AWB chose 3200K, the Lightroom 2800K.

The three pictures A, B, C are same ratio (appx 1/3) crops of the full M8 frame and were taken on a chair+release cable with:

 

1) Ultron 28mm; 1/8 sec; f2.0; ISO 2500; firmware 1.2.1

2) Ultron 28mm; 1/8 sec; f2.0; ISO 2500; firmware 2.0

3) Summicron pre-asph 50mm; 0.7 sec; f5.6; firmware 2.0

 

Match A,B,C to 1,2,3 please.

 

I also found out an intersting phenomenon when shutter was 2 sec or longer. After taking picture the LCD displayed message "Noise reduction" and counter went down 2 sec, 1 sec...Is this new? I could not find anything like that in any literature.

D is the original untouched DNG photo. Guess:before or after fw update?

Piotr

 

D

/applications/core/interface/imageproxy/imageproxy.php?img=http://i168.photobucket.com/albums/u175/nugat63/L1000006.jpg&key=3be7c12fc9ec773338c59e18ee06fb14995db78b82b0b6dfca94600778e74bcb">

 

 

 

 

A

/applications/core/interface/imageproxy/imageproxy.php?img=http://i168.photobucket.com/albums/u175/nugat63/L1000013.jpg&key=342fe7d456f73db34776a7f3ee521d5c42622eddd0714803b44b8fc76722eb91">

 

 

B

/applications/core/interface/imageproxy/imageproxy.php?img=http://i168.photobucket.com/albums/u175/nugat63/L1000005.jpg&key=64c162dce7d8ef8b26c8bf34503724ad6c452ad9d341e91d8ccf28fef32bfad2">

 

C

/applications/core/interface/imageproxy/imageproxy.php?img=http://i168.photobucket.com/albums/u175/nugat63/L1000029.jpg&key=d7af1c1a345664475fe5b1014de01e2e432bc420d828f25c9a34a973ac4d8332">

 

 

 

Disappointed? Well sirs&madams, I offer you truth, the whole truth and only truth...The purpose of the excercise was comparison of noise and not a pulitzer...

__________________

Leica MP, M8, D-Lux-3; Epson RD-1; Canon DSLRs

Share this post


Link to post
Share on other sites

Yes- the M8 will apply black-mask noise reduction at all long exposures. It is annoying in a sense that every exposure will take twice as long, but it makes for extremely clean night shots - take them from a tripod at ISO 160 - beautiful!

The LCD will briefly show you the amount of heat-generated noise that is eliminated this way.

And btw I think your shots are a bit underexposed. They should be cleaner. When developing in RAW shut off all sharpening.

Share this post


Link to post
Share on other sites

Nugat: Thought I would keep the bandwidth down by piggybacking my results on your thread - hope you don't mind.

 

First up: The chart - raw numbers are in the table at the top: If anything V2.0 is a bit NOISIER at ISO 1250/640 - although that is likely within the range of measurement error.

Share this post


Link to post
Share on other sites
Yes- the M8 will apply black-mask noise reduction at all long exposures. It is annoying in a sense that every exposure will take twice as long, but it makes for extremely clean night shots - take them from a tripod at ISO 160 - beautiful!

The LCD will briefly show you the amount of heat-generated noise that is eliminated this way.

And btw I think your shots are a bit underexposed. They should be cleaner. When developing in RAW shut off all sharpening.

 

Thanks for the noise reduction explanation.

There was no sharpening in LR, standard presets only: amount 25, radius 1.0, detail 25, masking 0. Should I take them down totally? Visually it does not make difference. Only WB was changed from the camera 3200K to 2800K. The original raw unchanged exposition and WB is on picture D-uncropped. For the crops the histogram is good, only two highlight blowouts for the hanging lamp, it's reflection and the smudge of light in the top right corner. If I exposed more, those higlights would be out of whack.

Anyway, "quick&dirty" as can happen in a reportage.

Can you see any noise differences? Where?

Piotr

Share this post


Link to post
Share on other sites
Nugat: Thought I would keep the bandwidth down by piggybacking my results on your thread - hope you don't mind.

 

First up: The chart - raw numbers are in the table at the top: If anything V2.0 is a bit NOISIER at ISO 1250/640 - although that is likely within the range of measurement error.

 

Don't mind at all, that what the thread is for--science, not impressions.

Share this post


Link to post
Share on other sites

Next: Experimental technique

 

Like Nugat, I shot my last ever exposures using fw 1.201, of a Gretag Colorchecker Mini, under both tungsten and full-spectrum(flash), using all 5 ISOs, DNG + fine JPEG.

 

I then upgraded the camera to V.2.0 fw, and reshot the same pictures.

 

I opened the shots in Adobe Camera RAW, using identical ACR settings. Sharpening, and both types of NR were set to 0.

 

In Photoshop, I then selected the darkest gray/black patch in the image of the colorchecker, and used Photoshop's Histogram to read the "standard deviation" ("Std Dev" in the histogram window) of the pixels within the patch for both luminance only, and all 3 colors (RGB) - see picture below

 

I then graphed the numbers that I read. (see chart above)

Share this post


Link to post
Share on other sites

Advertisement (gone after registration)

Thanks for the noise reduction explanation.

There was no sharpening in LR. Only WB was changed from the camera 3200K to 2800K. The original raw unchanged exposition and WB is on picture D-uncropped. For the crops the histogram is good, only two highlight blowouts for the hanging lamp, it's reflection and the smudge of light in the top right corner. If I exposed more, those higlights would be out of whack.

Anyway, "quick&dirty" as can happen in a reportage.

Can you see any noise differences? Where?

Piotr

Yes - there is default sharpening in RAW conversion which must be reduced to 0. That is something different from the sharpening in the standard workflow.

Share this post


Link to post
Share on other sites

IMHO there is no significant difference in overall noise.

 

There might possibly be a tiny change in the character of the noise - slightly less "clumpy" noise with V2.0, and thus slightly better sharpness. Individual noisy pixels rather than larger clumps of noisy pixels. If so it is still extremely minor <1%.

 

If anything I'd think that results from a small REDUCTION in noise reduction (and associated smearing/clumping, if that makes sense) which would explain the small bump-up in noise in the middle ISOs.

 

But realistically - I don't think there is a change.

 

Anyone with LR, Aperture, Capture One or other RAW developers can try to replicate my technique and see if there is a difference.

 

Totally as an aside, it is interesting to note how much lower the luminance noise is than the total noise including color. This is why B&W conversions often look better than color at high ISOs.

 

There was also about a 1-stop increase in noise for those images shot under tungsten light compared to flash at the same ISO, due to the extra amplification needed in the blue channel to get to correct white balance.

 

E.G. under tungsten light, ISO 2500 showed a noise SD of 25.94, compared to 18.33 under full-spectrum light. ISO 1250 tungsten was almost precisely equal to ISO 2500 flash - 18.98/18.33.

Share this post


Link to post
Share on other sites

I agree - I tried and any differences there may be are marginal

Share this post


Link to post
Share on other sites
IMHO there is no significant difference in overall noise.

 

There might possibly be a tiny change in the character of the noise - slightly less "clumpy" noise with V2.0, and thus slightly better sharpness. Individual noisy pixels rather than larger clumps of noisy pixels. If so it is still extremely minor <1%.

 

If anything I'd think that results from a small REDUCTION in noise reduction (and associated smearing/clumping, if that makes sense) which would explain the small bump-up in noise in the middle ISOs.

 

But realistically - I don't think there is a change.

 

Anyone with LR, Aperture, Capture One or other RAW developers can try to replicate my technique and see if there is a difference.

Adan I agree with you after doing a test with my chrome M8.

This is OK for me as if I have to shot at 2500 so be it. I'd rather get the shot then not.

Share this post


Link to post
Share on other sites

I tried to replace the pictures with "all standard sharpening down" versions...and they all disappeared. Trying to bring them back.

Share this post


Link to post
Share on other sites

Adan:

 

You should try your measurements with the default noise reduction turned on in the raw converter. The new firmware may create a noise that is easier for Lightroom or ACR to filter out.

 

When comparing, we should be using our normal work flows for processing, and in my case that is with noise reduction onat 25,0 and sharpening at 30 and .5 pixels.

 

Robert

Share this post


Link to post
Share on other sites

I tried to replicate adan's methodology and got quite different results. I exposed a macbeth color chart (the large version) in tungsten lighting with my m8 and v 2.000 firmware at ISO 2500 (a 50mm 'lux at 2.6). I shot raw, and converted using the current version of ACR, using the defaults except that I white balanced by clicking on the light gray square.

 

Here's a screenshot from PS (before applying Noisebuster). The marquee area shows a standard deviation of 7.82 in color and 7.00 in grayscale. Applying Akvis Noisebuster (using the automatic setting) reduces these numbers to 6.08 color and 5.6 grayscale.

 

 

The odd thing about these results is what I see in actual shooting is closer to adan's results - shadow area standard deviations of 10-12 (grayscale).

 

Odd.

 

I have two m8 bodies, but unfortunately I updated the second after I read this thread so I can't do a comparison.

Share this post


Link to post
Share on other sites

I managed to bring the pictures back this time with all default LR2 sharpening off.

Any difference? Any difference between them?

Anybody try to say which is which firmware?

Piotr

 

Can't bring the pictures back...what the ?

 

A-no sharp

Ļ

 

B-no sharp

Ļ

 

C-no sharp

Ļ

Share this post


Link to post
Share on other sites

A is the new firmware.

 

Did you have the noise reduction on in Lightroom?

 

What is the serial number range of the camera?

 

Was there any flourecent lighting? It can flicker and change the results.

 

Robert

Share this post


Link to post
Share on other sites

I am with all the testing, all the shots I did yesterday have been deleted. I just did my normal thing and see a difference in high ISO noise, looks good to me, I am keeping my M8. Time to get back to shooting.

 

But honestly: Where is Sean Reid now that this is going on? What does he think?

 

I bet he is testing like crazy to see if we all are..

Share this post


Link to post
Share on other sites

I have two M8's and except that one is black and one is chrome they are set up identically. I upgraded the firmware in one body but not the other. I have been taking the exact same photos at the higher ISO's using the same lens with both bodies. So far all I can say is there is absolutely no difference in noise performance. If you only have one M8 and have already upgraded the firmware I don't know how you can compare the noise performance.

 

Len

Share this post


Link to post
Share on other sites
I have two M8's and except that one is black and one is chrome they are set up identically. I upgraded the firmware in one body but not the other. I have been taking the exact same photos at the higher ISO's using the same lens with both bodies. So far all I can say is there is absolutely no difference in noise performance. If you only have one M8 and have already upgraded the firmware I don't know how you can compare the noise performance.

 

Len

 

Le me guess and I have a 50% chance of being wrong. Did you update the chrome one?

 

Try updating the other, but make sure you can duplicate your test images using the new firmware.

 

Robert

Share this post


Link to post
Share on other sites
Le me guess and I have a 50% chance of being wrong. Did you update the chrome one?

 

Try updating the other, but make sure you can duplicate your test images using the new firmware.

 

Robert

 

Robert, You win the prize. Yes I updated the chrome one. Is there a difference? The ser#s are fairly close, 3105xxx vs 3110xxx. Chrome is later. I plan on upgrading the chrome body later tonight. I'll be sure to take before and after photos under repeatable conditions and post results. Don't get me wrong, I would be tickled pink if the new firmware improved noise. I just don't see it.

 

Len

Share this post


Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...
Sign in to follow this  

  • Recently Browsing   0 members

    No registered users viewing this page.

×
×
  • Create New...

Important Information

By using this site, you agree to our Terms of Use. We have placed cookies on your device to help make this website better. You can adjust your cookie settings, otherwise we'll assume you're okay to continue., Read more about our Privacy Policy