Luminous Landscape Forum
Equipment & Techniques => Beginner's Questions => Topic started by: lowep on July 15, 2016, 07:49:23 am
-
Should any different in camera exposure settings be used for capturing a RAW file intended for converting and processing as a B&W image than for a color image, or is it just the same as photographing color?
-
Same settings, except that I might be able to use higher ISO or push shadows a bit more since noise is less noticeable and color artifacts more or less disappear.
-
Same settings, except that I might be able to use higher ISO or push shadows a bit more since noise is less noticeable and color artifacts more or less disappear.
That sounds reasonable
-
A camera doesn't see colours only luminousity values, therefore a problem doesn't exist.
-
Should any different in camera exposure settings be used for capturing a RAW file intended for converting and processing as a B&W image than for a color image, or is it just the same as photographing color?
Hi,
In everyday practice, the same exposure will do.
When pixel peeping, technically there could be a (small) benefit from shooting through a somewhat magenta filter, to compensate for the denser Red/Blue filters in the Bayer CFA. That would require less boosting of those channels which would help to reduce noise a bit.
The amount of difference also depends on the final postprocessing, and how image colors are used to change the tonal balance.
Cheers,
Bart
-
When pixel peeping, technically there could be a (small) benefit from shooting through a somewhat magenta filter, to compensate for the denser Red/Blue filters in the Bayer CFA. That would require less boosting of those channels which would help to reduce noise a bit.
That's an interesting idea: but since in fact a filter doesn't increase red+blue but decreases green, wouldn't it actually be an issue of boosting G more, rather than R+B less? Unless you were determining exposure by G channel clipping... is that the point I'm missing?
-
That's an interesting idea: but since in fact a filter doesn't increase red+blue but decreases green, wouldn't it actually be an issue of boosting G more, rather than R+B less? Unless you were determining exposure by G channel clipping... is that the point I'm missing?
Hi,
Usually the Green color plane has the highest exposure (Raw DN), and Red/Blue much lower. The Magenta filter will indeed reduce green transmission, so one can increase overall exposure (for the underexposed R/B planes, and the now reduced G plane).
Cheers,
Bart
-
Hi,
In everyday practice, the same exposure will do.
When pixel peeping, technically there could be a (small) benefit from shooting through a somewhat magenta filter, to compensate for the denser Red/Blue filters in the Bayer CFA. That would require less boosting of those channels which would help to reduce noise a bit.
The amount of difference also depends on the final postprocessing, and how image colors are used to change the tonal balance.
Cheers,
Bart
Curious, this is the first time I have heard this one. Kind of makes sense in theory, but have you tried it in a practical application and noticed any real difference?
-
I guess the general idea is to balance light between the red, green and blue filters, since there is twice green pixels as the others. For black and white this should pose no problem at all, but for color you will have issues with tones as RAW converters take for granted the additional green information witch will now be lacking.
I don't think you need a magenta filter that is too intense anyway. Also be aware of how many stops of light you lose with such a filter in front of the lens, as it can be anything between 2/3rds and 2 stops.
-
Should any different in camera exposure settings be used for capturing a RAW file intended for converting and processing as a B&W image than for a color image, or is it just the same as photographing color?
you might want to pay attention, for example, to your raw converter/camera profile(s) of choice for this ... if you intend on doing B/W then sometimes the stronger noise is less visually bad or how converter handles transition from clipped (in raw, 3-2-1 channels) to non clipped areas is visually less distracting... so you in fact can then either underexpose more or overexpose more based on that vs when you think about doing color.
-
I guess the general idea is to balance light between the red, green and blue filters, since there is twice green pixels as the others.
magenta filters are not about "green" sensels being twice more count wise - it is about the illumination spectrum + CFA spectral transmission giving more saturation to sensels behind "green" CFA filters... plus in a studio (or even outside in some cases) you can always gel the light
-
Curious, this is the first time I have heard this one. Kind of makes sense in theory, but have you tried it in a practical application and noticed any real difference?
I did some of this awhile back with a mild magenta filter. In many situations it did pull down the green channel, giving a more spectrally balanced exposure. This let me ETTR more effectively when taking pics with b&w in mind. Nowadays I don't bother…current sensors have enough latitude that it's rarely even a minor issue (for me).
-Dave-
-
Curious, this is the first time I have heard this one. Kind of makes sense in theory, but have you tried it in a practical application and noticed any real difference?
Hi,
I'm not particularly into B/W photography, but in an odd case (depending on the spectral content of the scene, it would help (a bit). Whether it's worth the risk of optical image deterioration (dirt / veiling glare / optical imperfections) remains to be seen on a case by case situation.
Cheers,
Bart
-
When pixel peeping, technically there could be a (small) benefit from shooting through a somewhat magenta filter, to compensate for the denser Red/Blue filters in the Bayer CFA. That would require less boosting of those channels which would help to reduce noise a bit.
Hi Bart, your idea is fine (in fact some people use it), when there is enough light to have a good exposure at base ISO. For instance it will produce cleaner shadows in well lit high contrast scenes. But if there is a lack of light (need to push ISO), it would have an overall negative impact on noise. In this situation the more photons you collect the better.
Aside from that, having channels with different sensitivity (i.e. G ending with higher levels in the RAW file) can have some advantage for B&W photography: while G can get clipped in the highlights, you can still 'recover' a nice grayscale gradient from the R and/or B channels there. The G channel will help you in having sufficiently noiseless shadows. Kind of inter-channel HDR that enhances DR for B&W, specially if some strategy is applied to weight more higher SNR channels on each image area.
That is one of the reasons why paradoxically I don't think monochrome sensors are the best option for B&W photography. When the Leica Monochrome has its highlights clipped, there is nothing to do there: pure white, zero information. This is much harder to happen on a Bayer sensor, where you need to severily overexpose to blow the three channels at the same time. Instead monochrome sensors are dramatically digital, right the opposite concept Leica represents. If I had that camera I would tend to underexpose a bit in order not to ruin the highlights, but this would reduce potential DR.
Another disadvantage of monochrome sensors is that you cannot apply software colour filters, which are very flexible and straightforward.
Back to the topic, in general exposing for B&W is less demanding than doing for colour. If you are lucky to have a mirrorless camera you can even previsualize in B&W.
Regards
www.guillermoluijk.com
-
Should any different in camera exposure settings be used for capturing a RAW file intended for converting and processing as a B&W image than for a color image, or is it just the same as photographing color?
To put it simply - No. Just keep doing what you're doing in colour, then convert to B&W. Do not "desaturated", though. Convert the colour properly from with an editing app like Lightroom.
-
Hi,
In everyday practice, the same exposure will do.
When pixel peeping, technically there could be a (small) benefit from shooting through a somewhat magenta filter, to compensate for the denser Red/Blue filters in the Bayer CFA. That would require less boosting of those channels which would help to reduce noise a bit.
The amount of difference also depends on the final postprocessing, and how image colors are used to change the tonal balance.
Cheers,
Bart
Well, shoot, Bart. I just recommended a CC30M in another similar thread. GMTA?
Jim
-
Kind of inter-channel HDR that enhances DR for B&W, specially if some strategy is applied to weight more higher SNR channels on each image area.
Ah! Yes of course. Do you know if any existing canned software does this?
That said, when you see something like this shot at ISO 4000, you realise that SNR is often not a priority:
https://500px.com/photo/171748293/huddle-in-the-rain-by-morkel-erasmus?from=following&user_id=10643117
-
Well, shoot, Bart. I just recommended a CC30M in another similar thread. GMTA?
LOL, pure physics, and creative thinking.
Cheers,
Bart
-
Hi Bart, your idea is fine (in fact some people use it), when there is enough light to have a good exposure at base ISO. For instance it will produce cleaner shadows in well lit high contrast scenes. But if there is a lack of light (need to push ISO), it would have an overall negative impact on noise. In this situation the more photons you collect the better.
Aside from that, having channels with different sensitivity (i.e. G ending with higher levels in the RAW file) can have some advantage for B&W photography: while G can get clipped in the highlights, you can still 'recover' a nice grayscale gradient from the R and/or B channels there. The G channel will help you in having sufficiently noiseless shadows. Kind of inter-channel HDR that enhances DR for B&W, specially if some strategy is applied to weight more higher SNR channels on each image area.
Correct, although I in principle assume 'correct' exposure, and usually low ISO for ETTR.
You raise a valid point about some potential DR enhancement, being able to choose from different levels of shot noise by channel. I'd add that good Noise reduction software also allows to change the weighting for the Red/Blue and Green bands and the amount of noise reduction.
Bottom line is that there are several benefits to shooting Black and White images as Colour and not Monochrome, and unleash the full potential of tonal control after the fact, during postprocessing. I like Topaz B&W Effects for that, but Googles Silver Efex Pro 2 has somewhat similar capabilities.
Cheers,
Bart
-
I have mentioned this before but Aperture and Capture One allow you to change Hue, Saturation and Luminance in the color channels AFTER conversion to B&W and in addition to the color sliders within the B&W conversion tool. To me, this seems to offer an even greater level of tonal control compared to only using the color sliders in the B&W tool. It also means you can see the effect of the HSL changes in the B&W image as you change them which is handy.
-
To me it makes no sense: the reduction to BW is just a weighted sum of different colours. Hue and saturation are about individually transforming those colour components to something else: it relates to the colour output, which is not seen. Only colour luminance relates to the colour input, so it makes sense that only that is relevant to the B&W result.
Including the extra controls just provides you with 18 controls which provide the same range of output variation as the 6 in LR, so far as I can see.
-
To me it makes no sense: the reduction to BW is just a weighted sum of different colours. Hue and saturation are about individually transforming those colour components to something else: it relates to the colour output, which is not seen.
Hi Graham,
That not necessarily true. Suppose you have a shade of green and a shade of (dark)yellow-green, say some foliage and grasses that you want to get some more separation in. With a Green or Yellow filter one would lighten the tonality of both approximately the same. But when the Green is Hue shifted towards Cyan, and the YellowGreen is Hue shifted towards Orange, it becomes possible to enhance the contrast between those shades by applying a color filter (e.g. a Redfilter would lighten the now Orange, and darken the now Cyan green). Alternatively one could lessen the contrast by Hue shifting both colors closer to the same hue, and filtering with Green for lightening or Blue or Red to darkening. This is of course much easier to do with a dedicated tool than only the generic LR controls and adjustments.
Only colour luminance relates to the colour input, so it makes sense that only that is relevant to the B&W result.
So it's not 'only', but instead yet another way of controlling tonality in the resulting B&W rendering.
Including the extra controls just provides you with 18 controls which provide the same range of output variation as the 6 in LR, so far as I can see.
More controls offer more ways to add or modify to the strength of tone adjustment.
Cheers,
Bart
-
Correct. And in my limited understanding the 'blue' slider in the B&W conversion tool, for instance, affects a specific blue or range(?) of blue. Changing the saturation and hue changes how the 'blue' B&W slider affects the blues in the image. But yes, the luminance slider has the greatest impact.
For folks who are really deep into B&W tone control I recommend trying this approach. I have found it particularly useful in achieving a red filter effect to darken blue skies. Just applying the red filter in Silver Efex or the blue slider in LR will often result in banding and noise before the desired effect is reached. Using the HSL sliders gives a slight, but still significant level of control in darkening that blue without adding artifacts.
-
With a Green or Yellow filter one would lighten the tonality of both approximately the same. But when the Green is Hue shifted towards Cyan, and the YellowGreen is Hue shifted towards Orange,
Thanks Bart,
That means there are two filter steps? Having a separate hue slider on each of the separate colours suggests that there is first a filter applied to define on which wedge of the a*b* plane the hue-shift will apply. So if it were impossible to separate them to adjust luminance, it would be equally impossible to separate them to apply different hue shifts.
Obviously that is an over simplification, since the real issue is that the transition zone of the filter in the angular component is not sufficiently steep, and one would really like a higher-order filter. However ganging together two low-order filters is a pretty ineffective way of doing that, from memory. And then the idea of having ringing artifacts from higher order theta-space filtering is a little disturbing, but maybe I'm too squeamish.
So am I missing something here?
Regards,
Graham
-
I always wondered (but didn't have time to implement) how a locally adaptive B&W algorithm would work, calculating the weight of each RGB RAW channel to build a customized B&W output for every pixel area in order to maximize either:
- SNR (highest RAW channels weight more)
- Micro contrast (highest HPF output channels weight more)
- A mix of both SNR and micro contrast
The obtained luminance can look quite weird in terms of real world perceived luminance but could be rearranged starting from a robust B&W conversion in terms of initial noise and/or microcontrast.
I'll have some 2-3 weeks free in the next times so perhaps I try these strategies.
Regards
www.guillermoluijk.com
-
Thanks Bart,
That means there are two filter steps? Having a separate hue slider on each of the separate colours suggests that there is first a filter applied to define on which wedge of the a*b* plane the hue-shift will apply. So if it were impossible to separate them to adjust luminance, it would be equally impossible to separate them to apply different hue shifts.
Obviously that is an over simplification, since the real issue is that the transition zone of the filter in the angular component is not sufficiently steep, and one would really like a higher-order filter. However ganging together two low-order filters is a pretty ineffective way of doing that, from memory. And then the idea of having ringing artifacts from higher order theta-space filtering is a little disturbing, but maybe I'm too squeamish.
So am I missing something here?
Hi Graham,
The default user interface of most applications is not ideal for such things, and would probably require lots of local selections and adjustments. A dedicated B&W conversion plugin will make life a lot easier. Also, real-time updated previews will make setting the parameter levels a more creative process.
Cheers,
Bart