Re: Possible serious Issue with AWB & M9?

[larrylgreenhill]

Advertisement (gone after registration)

Hi All:

I continue to find problems that with the new M9 around shooting indoors. All the White Balance settings work as long as I shoot outdoors. Inside, it seems as if the sensor does not pick up much light so the histogram shows serious underexposure and the shadows are filled with digital noise, even shooting at F/2 at ISO 2500. I've gone so far as to do a manual set of the White Balance pointing the camera at bed sheets for a white reference, which results in a normal histogram and image with none of the green-yellow tinge. But then I point the camera into the room, away from the sheets, and it's as I've shooting in a dark cave. By the way, the incadescent bulbs are all the energy saving type.

 

Perhaps this is a case of an inability of the sensor to handle the new energy-saving bulbs. Do others have the same experience?

 

Larry Greenhill

[gravastar]

.......................A color temp meter looks more and more enticing as time goes on.

 

That's what I thought when I purchased a Minolta Color Meter some years ago during my large format days. It certainly works very well with tungsten light sources but isn't so good with fluorescents. The reason being of course the meter is calibrated for a black body/incandescent radiation source with a well defined continuous spectrum. Unlike CFLs with a mix of phosphors which give an uneven discontinuous spectrum and may "look" like the same color temperature as an incandescent source. The situation isn't helped by the CFL/fluorescent manufacturers who can be somewhat crafty in publishing the spectral characteristics of their lamps. By using a spectrometer with a wide sampling bandwidth intense sharp emission lines such as those due to mercury are averaged out and don't show on the spectrogram, making it appear smoother than it is.

 

Bob.

[Advertisement]

The Leica Q3 is available for pre-order at B&H Photo Video and Adorama

[jklotz]

As a professional interior and architectural shooter, the first thing I do when going into any interior is to remove those terrible "earth saving" light bulbs, and replace with standard incandescents. The color cast you get from those things is hideous. I keep telling myself that I am going to go buy a as many cases of standard incandescent bulbs as I an afford one of these days, before they are completely gone.

[Antonius]

Issue or not. The in-camera white balance measure and processing are there for some reason and for a camera in this price tag it's not too much to ask that it's as accurate as any other high-end picture-taking machine. Why should it be worse than my M8 or Nikons? OK auto white balance is a challenge in mixed light conditions, but what about the preadjusted alternatives like the one for "tungsten". Even this gives me more or less the same yellowish result as "auto" with the M9 I am using (a demo camera from my dealer while I'm waiting to get mine back from reapir at Solms). Sometimes I'm in a hurry needing to use jpgs and sending files as they are straight from the camera. This is when I like the white balance to be at least OK. The best (and only reliable) way to help the M9 with the white balance in difficult light conditions, from my experience, is to set it to "manual" and measure the on a sheet of white paper. It's a bit slow but i works.

[nryn]

For what it's worth, my indoor xmas photos with the M9 _and_ the m8 are way off. Outdoor, tungsten, shade--these are more or less in the ballpark. Throw CFLs into the mix and all bets are off.

[neli]

Hi All:

I continue to find problems that with the new M9 around shooting indoors. All the White Balance settings work as long as I shoot outdoors. Inside, it seems as if the sensor does not pick up much light so the histogram shows serious underexposure and the shadows are filled with digital noise, even shooting at F/2 at ISO 2500. I've gone so far as to do a manual set of the White Balance pointing the camera at bed sheets for a white reference, which results in a normal histogram and image with none of the green-yellow tinge. But then I point the camera into the room, away from the sheets, and it's as I've shooting in a dark cave. By the way, the incadescent bulbs are all the energy saving type.

 

Perhaps this is a case of an inability of the sensor to handle the new energy-saving bulbs. Do others have the same experience?

 

Larry Greenhill

 

Hi Larry

 

Yes, same problems for me, which is really irritating.

I wonder if it's just software, like a primitive look up table or if the sensor that measures the light is just primitive. I hope it's the former so firmware update where art thou, etc, and if the latter that leica will offer a free fix. I like the slowness of the rangefinder system, but I don't appreciate unnecessary hurdles.

What's the best way to get Leica to take notice of these kinds of issues, are there sort of collective dear Leica letters and I can add my signature?

 

blue skies

 

Cornelius

[gravastar]

Advertisement (gone after registration)

Interesting. I've wondered about this as well. Actually I first noticed the problem on a Panasonic GH-1 when diferent images shot at the same table in a restaurant exhibited different colour balance characteristics. I wonder if it related to the "flickering" nature of fluorescent and compact fluorescent lights certain exposure times. I wonder: does the frequency of the light source result in different colours at slightly diferent shutter speeds?

 

- Vikas

The light output of each color component in the mix of tube phosphors can have a different persistence or decay time. For the old type of inductor ballasts used in lamps that results in a color change during each half cycle of the power supply (10ms or 8.3ms) if the phosphor persistence is short. For comparable shutter speeds the color can be different from frame to frame. A similar problem (but due to a slightly different cause) exists for stadium sports photographers.

 

CFLs using high frequency electronic ballast don't suffer from this problem although their effective color temperature can change substantially as they warm up over a period of a few minutes.

 

ETA: I do know that with early versions of the M8 firmware you could take several successive pictures of a scene with a slow shutter speed (to eliminate lighting color flicker) and occasionally a frame would have a different color balance than the others.

 

Bob.

Edited December 28, 2009 by gravastar

[ggits]

Same problem here !

 

I had just today placed an order for ColorChecker Passport in order to decrease this problem.

Shooting without blitz indoor is a yellow experience no matter settings in WB. Need a lot of work in LR. Shooting outdoor make to much red/purple impression.( ok - it can be removed in LR)

 

I love my M9 - but to be honest - the RAW is really "RAW" - the firmware in M9 appear to be "Not for resale - TEST ONLY" Hope they will make an update soon.

 

ggits

[jaapv]

Fluorescent is not too bad, this was just one colour balance click on the headrest behind the lady in green in C5.

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!

[scott kirkpatrick]

 

ETA: I do know that with early versions of the M8 firmware you could take several successive pictures of a scene with a slow shutter speed (to eliminate lighting color flicker) and occasionally a frame would have a different color balance than the others.

 

Bob.

 

(as I understand it, and I am not a color scientist...)

Auto White Balance consists of placing in the raw file a sample of "neutral middle grey" as the camera thinks it would appear in the light in which the scene is taken. From the supplied RGB values, each raw developer program extracts a color temperature and a magenta-green shift, where + is in the magenta direction. Each program gets somewhat different numbers for these, so I'll cite Capture One 5's values.

 

The initial M8 firmware only used three possible sets of RGB values for "neutral," and in mixed light, small changes in the scene would cause the choice of "neutral" to vary unpredictably from shot to shot. People like Ken Tanaka produced series of alternating yellow and blue images to make it crystal clear that this did not make the camera or the photographer look exactly on top of the situation. That's the "flicker" phenomenon. Since release 2.0 of the M8 firmware, auto white balance has produced a calculated "neutral" value, which changes slowly and continuously with variations in the scene content, eliminating flicker, and getting much more plausible results.

 

Where the M8's AWB falls down, in my opinion, is that it has some hard stops. It will not go to very low color temperatures or suggest very large color shifts. In those cases, custom white balance must be done by shooting a neutral grey card (e.g., WhiBal) and saving that as a custom setting, or simply click-balancing on the card in a test shot and applying the result to the rest of the shoot in post processing. Or to taste, if you want some degree of extra warmth or whatever...

 

The M9 has the same limitation. It gives a calculated, smoothly varying "neutral" with limits. The .limits seem to make it unable to deal well with low color temperatures. The M9 handles the horrible energy-saving flourescents less well than the M8 does, because of these limits. I did an experiment, which I'll report in a separate, number-filled post, that shows this.

 

It would be nice if a firmware upgrade stretched the M9 white balance to cover more situations,. Since this is annoying, but not a show-stopper, I would expect to see this get sorted out in a second or third iteration.

 

scott

[scott kirkpatrick]

I've been shooting judo practices and contests lately. These tend to take place in school gymnasiums, with banks of cool white long tube fluorescent lights high overhead. Walls are painted strange colors, and the mats on the floor are strongly colored in red, green, blue, yellow or grey. Color is definitely a part of the picture, and it is a white balance nightmare. I have always needed to click-balance my shots, usually on the white judo coats, to eliminate the strong yellow cast that the M9 leaves, regardless of setting -- AWB, daylight, tungsten, or fluorescent.

 

Tonight, I checked my white balance with a WhiBal grey card. I found that the M9's AWB was giving me "as shot" settings of 5100 to 5500 K color temperatures and shifts of +10 to +13 in the magenta direction, as read out by Capture One 5. Clicking on the WhiBal card shifted this to from 4350 K +18 to 4800 K +22, which looked much better. That's a bigger discrepancy than I used to see shooting in similar settings with the M8. I pulled up one old M8 shot and checked: as shot = 4750 K + 4, best balanced version = 4800 K + 9.

 

To see some worst cases, I tested the white balance resulting from single bulbs in our house tonight, using cold low energy flourescents, warm energy-savers, and low wattage incandescents. For all the warm lights, my M9 gave the same estimate of "neutral:" 2909 K +12. Apparently the warm CFLs are being treated the same as candles and weak incandescents, presumably because the AWB algorithm has hit some lower limit. The M8 is limited in the same way. For each of the three warm light samples, the M8 registered 3114K + 4 as its recommended "neutral." Click balancing on the WhiBal, I found the M8 was not too far off. The low wattage incandescent bulb corrected to 2350 +4 (lower temperature, but no extra shift needed). The two CFLs corrected to 2350 +15, requiring more magenta as well as a shift overall to the lower temperature.

 

The M9 needed bigger corrections. The incandescent source corrected to 2050 +18, needing added magenta. The two warm CFLs corrected to 1900 +26 and 2030 +23, needing biggers shifts as well as lower temperatures. In the light of these sources, all the shots come up very yellow. Neutralizing the yellow, however, can lead to purplish skin tones.

 

My two cool white CFLs give off very different light. Each registered over 5000 K and +11 or 12 with the M9. Correcting to neutral required lowering color temperature to 4500 and increasing the shift to 25 or 26. The M8 registers these as 4940 K + 15 in one case, and 4300 +17 in the other, and those seem to be the correct values.

 

So the M9 is having more trouble handling flourescent lighting, especially the more recent low energy warm bulbs, than the M8. I think the AWB routines and presets will probably improve with a little more tweaking.

 

scott

.

[abagar6]

3 words:

 

Use manual WB!

 

Cheers

[giordano]

That is exactly why I have a huge mains filter in the power supply of my Hifi.

 

I'm working on a range of special audiophile optical covers for compact fluorescent lamps. They will smooth the harmonic ionisation overload potential curve of the fluorescent tubes, and in conjunction with low decay rate high-stability negative ion generation units will also stabilise the nitrogen atoms in the room's atmosphere, delivering an 83.57% reduction in nano-eddies generated by the unbalanced phosphor interactions, contributing to the cleanest ever listening experience.

[jaapv]

When are you bringing them to the market, John? They will really make my tin-foil hat shine! Unfortunately such a filter is beneficial if one lives not too far from heavy industry

[finkaudio]

That is exactly why I have a huge mains filter in the power supply of my Hifi.

 

Hi,

 

unfortunately, a mains filter can take away some dynamics. Whenever I try one, I get better and smoother top end, but less dynamic excitement. Things are just getting slow

 

And yes, in the evening my system sounds better - not to talk about humming transformers of my NAIM system during the day.

 

So with all those problem in my professional life, any AWB problems with my M9 look tiny.

 

No problem to change the WB with a slider in LR, but try that with a noisy mains transformer

[jaapv]

Anyway, about the importance of AWB - it is important if one wants to enter German wildlife photography competitions. WB changes in RAW developing will count as digital manipulation and disqualify the entry. As will "more than standard" levels and curves - at the jury's discretion:eek: A very luddite bunch over there. Despite 95 % digital entries they would like to turn the clock back to Agfachrome.

[wlaidlaw]

I assume that the AWB will be better in the "very slow to arrive" next generation firmware. It has been out for beta testing for weeks if not months now. I don't know what the hold up is. The M9 is a country mile better than the M8 was at the same stage in its evolution. Manual WB is fine and I used it all the time initially with the M8 but I did find I kept forgetting to update it for different lighting situations. If I had set it for use with a B+W 092 IR filter, the results of not resetting it were kind of interesting - blue and black pictures.

 

Wilson

[giordano]

When are you bringing them to the market, John? They will really make my tin-foil hat shine! Unfortunately such a filter is beneficial if one lives not too far from heavy industry

 

I'd have thought that high-end audio gear would be well shielded and have its DC power supplies adequately smoothed and well isolated from the mains. Do you need the filter to prevent the rest of the mains wiring interfering with the audio signal (presumably between the stylus and pre-amp or in the speaker cables), or am I wrong and it's needed for the sound system itself?

[jaapv]

The problem is high frequency junk that comes superimposed on the 50 Hz mains. It can be caused by electronics in the house like dimmers and fluorescents and it can come from outside sources. The filter I use is basically a 1:1 transformer that isolates the whole system from the mains.

Edited December 29, 2009 by jaapv

[Advertisement]

Hallo,

Look here: Leica M

[larrylgreenhill]

Cornelius:

Thanks for your response to my posting about the white balance challenges the M9 has with the compact flourescent lamps (CFL) energy savers. My dealer suggested I wait one week until Leica Camera USA re-opens after the holiday and send an email to their technical specialist handling the M9, Justin Stailey. I plan to do that to find out what the Leica company's intentions are concerning the M9's white balance limitations.

 

In the meantime, I returned home, and was able to shoot the same indoor picture in my office with the same Leica 35 mm f/2 Sumicron ASPH lens, same white balance setting of tungsten, same exposure (1/90 second) and same ISO (1250), switching only camera bodies between the M8 and M9. I reproduced the dense green-yellow cast in the M9, with the high levels of noise in the shadows, which were totally missing in the M8's images.

 

Following the advice of others on this thread, I then went into the SET menu and switched the M9 to custom WB by focusing on a pristine sheet of xerox paper and clicking the shutter. I shot the same picture, this time with custom WB, and the green-yellow cast was gone, as was the digital noise. I thus validated an approach used by other people on this thread. I plan to save the custom WB as one of the profiles, and label it "CFL indoors."

 

Best,

Larry G