Pages: 1 [2]   Go Down

Author Topic: White balance with filters - less noise in blue?  (Read 14079 times)

Guillermo Luijk

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2005
    • http://www.guillermoluijk.com
White balance with filters - less noise in blue?
« Reply #20 on: May 02, 2009, 07:41:01 pm »


Can exposure, and hence noise, and hence captured dynamic range be optimised through optical filtering? of course it can.

Is it worth? in my opinion, it is not. Let's say the G channel clips 1,5 f-stops before the B channel, and with the appropiate filter we manage to make both clip at the same time. Exposure could then be optimised with ETTR. But will 1,5 extra f-stops of exposure in the B channel be critical in the final noise and quality of image? IMO it won't.

In addition to this the behaviour of the filter is not ideal, so according to the spectral characteristics of the scene's light colour casts can be introduced in the image making final colours change.

If you really want to improve your landscape photography, just shoot twice with 3 or 4 f-stops apart and the improvement will be miles ahead of what you get with the filters. Most tripod photography allows to do this even if there are moving elements, that just need to be properly masked not to appear blurred.

BR

Panopeeper

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1805
White balance with filters - less noise in blue?
« Reply #21 on: May 02, 2009, 09:48:36 pm »

Quote from: GLuijk
Is it worth? in my opinion, it is not. Let's say the G channel clips 1,5 f-stops before the B channel, and with the appropiate filter we manage to make both clip at the same time. Exposure could then be optimised with ETTR. But will 1,5 extra f-stops of exposure in the B channel be critical in the final noise and quality of image? IMO it won't
1.5 EV makes a difference like day and night. Look at the following capture; the difference in intensity between the top left and top middle patch is 1.5 EV; the difference in noise is 25% vs 50%.

1.5 EV is not realistic, but about the half of that is. Look at the patches at the bottom left and bottom middle. The difference in intensity is 0.75 EV, and in noise 47% vs 68%.

Nothing against integrating bracketed shots, but that is not always possible or practicable.


Logged
Gabor

Rob C

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 24074
White balance with filters - less noise in blue?
« Reply #22 on: May 03, 2009, 05:26:22 am »

Perhaps I shouldn´t post this here - is so, then I apologise now.

However, trying to follow this thread I come to the conclusion that photography as she was known is dead.

I can´t remember a time in the past era - say up to the mid 80s - where photography and its artistic nature had been so subsumed by factors beyond the basic requirement of exposing and developing a film correctly. The current atmosphere of scientific/psuedo scientific gobbledegook has so overtaken the conversation that used to be photography that I am in a state of perpetual wonderment that there remain any practitioners with either time or interest to perform the otherwise simple operation of framing a shot and clicking the shutter.

It seems to me that we have entered a realm where the art has fled and the science run amok.

Far from the era of digital making the art more democratic, it has, if anything, created a far wider social gap in both the technical and financial levels that impinge on the ability of people to join in the game in any meaningful way.

I began with an apology and end with another; neither do I want to hi-jack this thread - I just felt the need to sigh out aloud.

Rob C

cmi

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 492
White balance with filters - less noise in blue?
« Reply #23 on: May 03, 2009, 08:23:26 am »

To end it from my side,  I got my infos and I will just try it. Thanks once again! I think further theoretical discussion of how ineffective the technique is is just wasted energy. I will report back my findings.

Garbor: I was asking for advice, and you gave me yours, and I thank you for that. Maybe Im inconsiderate with the following especially since Im new here, but I just HAVE to say it.

Im a bit bewildered, everybody uses google and wiki. Why you are telling me such matter of course? It suggests you have not much to say for yourself, that would not throw a very good light on you, dont you think?

To overdraw it: Doubts, difficult to resolve issues, peppered with superstring theory, there to not to be understood, thrown at photographers, workers, who are rather empirical beeings. If that does help someone, FINE. But then, maybe Im just dumb?

And dont be upset! Im just telling what goes thru my mind!

All the best,


Christian



Quote from: Panopeeper
I should not have mentioned it :-)

Seriously, this is not a small topic (nor is it on-topic). You can find all possible and more information in internet by searching for "black body radiation". The essence is, that the spectrum of radiation of the sun and of incandescent (i.e. glowing) light sources is near-continuous.

See the solar spectrum (it is not so clean on the earth level, but the visible part is still quite continuous).

If the illumination is modified by such a filter as the Redhancer, the originally continuous spectrum becomes "ragged"; this illumination can not be described the conventional way via "temperature" and "tint" (some fluorescent sources too create this problem). The only way I know of to counter this would be to modify the color conversion from the camera's color space in sRGB, etc., specifically customized to the filter.

This is theoretically possible with the Adobe DNG Profile Editor (note: it is not only for DNG raw files!), but I don't think the effort is justifiable.
Logged

Guillermo Luijk

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2005
    • http://www.guillermoluijk.com
White balance with filters - less noise in blue?
« Reply #24 on: May 03, 2009, 11:26:46 am »

Quote from: Panopeeper
1.5 EV makes a difference like day and night. Look at the following capture; the difference in intensity between the top left and top middle patch is 1.5 EV; the difference in noise is 25% vs 50%.

1.5 EV is not realistic, but about the half of that is. Look at the patches at the bottom left and bottom middle. The difference in intensity is 0.75 EV, and in noise 47% vs 68%.
Not really Gabor. An improvement in the B channel of 1 stop (if it's achieved with the filter, and nobody will guarantee that) will have little influence in the final image because B contributes less to it. Perhaps there will be some more colour noise, but this is the easiest to eliminate.

In these 2 shots, the one with filter had 1 extra stop in the B channel. R and G channels remained (G gained some exposure, and R lost some too):

RAW histograms:



And the final improvement was not like day and night:



Crop 100% dark area:



But 4EV of extra exposure in all three channels do make a difference worth the mess:



BR
« Last Edit: May 03, 2009, 11:38:08 am by GLuijk »
Logged

BernardLanguillier

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 13983
    • http://www.flickr.com/photos/bernardlanguillier/sets/
White balance with filters - less noise in blue?
« Reply #25 on: May 03, 2009, 01:22:11 pm »

Quote from: Rob C
However, trying to follow this thread I come to the conclusion that photography as she was known is dead.

I can´t remember a time in the past era - say up to the mid 80s - where photography and its artistic nature had been so subsumed by factors beyond the basic requirement of exposing and developing a film correctly. The current atmosphere of scientific/psuedo scientific gobbledegook has so overtaken the conversation that used to be photography that I am in a state of perpetual wonderment that there remain any practitioners with either time or interest to perform the otherwise simple operation of framing a shot and clicking the shutter.

Rob,

My view is that there really isn't any need to read all this if you are not interested, just like there was no need to become a master of Ansel Adams's theories to be able to shoot great B&W images in the film days.

The standard image quality delivered by recent bodies in jpg - like the D3x - is so incredibly good that the need to shoot raw is in fact a lot less than it used to be. Shoot both jpg and raw, use jpg 98% of the time and raw the remaining 2% when the WB is really way off and you have the best of both world (and make Hard disk manufacturers happy on top of that). Make jpg a 16 bits format, and 98% will become 99.9%.

I am personally interested as I see these posts as useful hints that there are ways to potentially gain a few additional percents of image quality if I need to some day for some special application. Although I don't see myself pre-filtering my images to gain a bit of noise in the red channel since I believe that the current discussions have not necessarily considered all aspects of image quality (including the quality of the color information from a spectral standpoint).

Cheers,
Bernard

Panopeeper

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1805
White balance with filters - less noise in blue?
« Reply #26 on: May 03, 2009, 02:30:55 pm »

Quote from: Rob C
I can´t remember a time in the past era - say up to the mid 80s - where photography and its artistic nature had been so subsumed by factors beyond the basic requirement of exposing and developing a film correctly. The current atmosphere of scientific/psuedo scientific gobbledegook has so overtaken the conversation that used to be photography that I am in a state of perpetual wonderment that there remain any practitioners with either time or interest to perform the otherwise simple operation of framing a shot and clicking the shutter.

It seems to me that we have entered a realm where the art has fled and the science run amok
Well, Rob, apparently you regard mixing chemicals, measuring temperatures and timing the process in the dark room "art". On the very same basis one could regard developing the raw image through a serie of computer program processes "art" as well.

Don't you agree with this?
Logged
Gabor

Kirk Gittings

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1561
    • http://www.KirkGittings.com
White balance with filters - less noise in blue?
« Reply #27 on: May 03, 2009, 02:42:21 pm »

Quote from: Rob C
However, trying to follow this thread I come to the conclusion that photography as she was known is dead.

I can´t remember a time in the past era - say up to the mid 80s - where photography and its artistic nature had been so subsumed by factors beyond the basic requirement of exposing and developing a film correctly. The current atmosphere of scientific/psuedo scientific gobbledegook has so overtaken the conversation that used to be photography that I am in a state of perpetual wonderment that there remain any practitioners with either time or interest to perform the otherwise simple operation of framing a shot and clicking the shutter.

It seems to me that we have entered a realm where the art has fled and the science run amok.

Far from the era of digital making the art more democratic, it has, if anything, created a far wider social gap in both the technical and financial levels that impinge on the ability of people to join in the game in any meaningful way.

I began with an apology and end with another; neither do I want to hi-jack this thread - I just felt the need to sigh out aloud.

Rob C

You are confusing art and craft. It is possible to do art without craft, but it seems to me that they are essentially related. That was why I spent years mastering the Zone System (with all the asa testing and characteristic curves for different developers etc.) and then pared it down to a minimalist personal version with one film and developer (that would do what I wanted and I used for the next 20 years) based on all that testing that gave me the control I needed without unnecessary fuss.  I don't see digital any different. I don't just want well seen images but well crafted images too and these kind of discussions, though they quickly go over my head, may give me a simple piece of understanding or practice that will greatly improve my workflow or final product. Honor the teckies, they contribute to craft.
Logged
Thanks,
Kirk Gittings

Panopeeper

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1805
White balance with filters - less noise in blue?
« Reply #28 on: May 03, 2009, 03:00:41 pm »

Quote from: GLuijk
Not really Gabor. An improvement in the B channel of 1 stop (if it's achieved with the filter, and nobody will guarantee that) will have little influence in the final image because B contributes less to it
May be, or may not. Shoot for example at non-halogen incandescent light. The low blue will be multiplied by WBing (the blue coeff can go up to five and more), thus the noise too will be multiplied. I just measured it on a shot made in a theater, the blue is 2.65 EV lower than the red and 2.2 EV lower than the green on a white patch.

Quote
Perhaps there will be some more colour noise, but this is the easiest to eliminate
Just the opposite.

Btw, originally there is no "color noise". The noise is always in the intensity within a raw channel. The color noise is the result of raw processing, caused by the "intensity noise".

Quote
In these 2 shots, the one with filter had 1 extra stop in the B channel. R and G channels remained (G gained some exposure, and R lost some too)
This is not a really noisy situation, as far as I can see it.

Quote
But 4EV of extra exposure in all three channels do make a difference worth the mess
Again, I don't dispute the superiority of blanding multiple exposures when it is possible and realistic, but your example does not prove much. You would have to show, how much +1 EV would have enhanced the shot.
« Last Edit: May 03, 2009, 03:04:15 pm by Panopeeper »
Logged
Gabor

Guillermo Luijk

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2005
    • http://www.guillermoluijk.com
White balance with filters - less noise in blue?
« Reply #29 on: May 03, 2009, 03:36:41 pm »

Quote from: Panopeeper
May be, or may not. Shoot for example at non-halogen incandescent light. The low blue will be multiplied by WBing (the blue coeff can go up to five and more), thus the noise too will be multiplied. I just measured it on a shot made in a theater, the blue is 2.65 EV lower than the red and 2.2 EV lower than the green on a white patch.
It was a theater with red velvet chairs? or perhaps they were playing 'The hunt for the Red October' under the typical submarine red lights?
You know these are not normal situations at all, and in any case the B channel would weight less than the others to the final image.

I know there is only a kind of noise, but after demosiacing it visually translates into two axis: luminance and colour, that can be processed separately with separate strategies. If the B channel is clearly noisier than the other two, we can expect more consequences in colour noise than in luminance noise, since B weights less to luminance than R and G.

Regarding the multiexposure technique, why should we care of testing how much 1EV would improve, if we can afford 4EV? that's the power of multiexposure.

BR
« Last Edit: May 03, 2009, 03:38:44 pm by GLuijk »
Logged

Panopeeper

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1805
White balance with filters - less noise in blue?
« Reply #30 on: May 03, 2009, 03:55:52 pm »

Quote from: GLuijk
You know these are not normal situations at all
I get blue coeffs over 3 even under mixed halogen and non-halogen incandescent.

Quote
in any case the B channel would weight less than the others to the final image
? The weight of a channel does not depend on the illumination but on the color of the objects in the frame.

Quote
If the B channel is clearly noisier than the other two, we can expect more consequences in colour noise than in luminance noise, since B weights less to luminance than R and G
Right. So what? Is color noise not noise? In fact, color noise is *worse* than luminance noise.

Quote
Regarding the multiexposure technique, why should we care of testing how much 1EV would improve, if we can afford 4EV? that's the power of multiexposure
The topic is the effect of color filtering. You contrasted the multiexposure blending with the effect of the color filter. One method is superior to another, if the result is superior or the same result can be achieved easier/cheaper.
Logged
Gabor

madmanchan

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2115
    • Web
White balance with filters - less noise in blue?
« Reply #31 on: May 04, 2009, 08:16:26 am »

If you are really trying to optimize the exposure then you would need to pick a different filter for each scene (subject matter and lighting) and camera. Shooting in shade (e.g., ~D75) is very different from shooting under tungsten lighting, for example. For color issues, ideally you'd use a separate camera profile for the case when using the color filter, but then again, ideally you'd use a separate camera profile for each lens you use, and other filters (e.g., polarizers). The spectral transmission characteristics of my lenses are similar, but definitely different. In practice, however, nearly all the variation is factored out by white balance. I suggest that shooting with some type of magenta filter is only worth it in specialized, controlled conditions, and even then I doubt you'd see the result in a print. If in doubt, try it.

(To me this is like folks who aren't sure whether to print to an Epson printer with High Speed On vs Off, or 1440 dpi vs 2880 dpi. Some photographers see the difference in the print, others don't. If in doubt, make test prints and evaluate with your own eyes.)
Logged
Eric Chan
Pages: 1 [2]   Go Up