Shooting JPEG in color for Black and White Images

[keethjon]

This question is a follow up to my post of 9-29 concerning shooting in color to make black and white images. I understand that shooting in color in the raw mode will give you more digital information to adjust tones, contrast, etc. (once you convert the image to black and white)
My question is if you shoot JPEG in color and then use post processing to convert to black and white, will you get better black and white images that way?
In other words, is there a difference between shooting JPEG in color and then converting, or will you get the same results by using the cameras black and white mode?

Thanks in advance again for your replies,
Keith

 

[480sparky39 38]

When you shoot in color, you have many choices in post that can affect the way your final conversion looks. Shooting in raw gives you far more editing power than shooting jpegs.

When you shoot in "B&W in camera", you only get one choice, and that's the choice made some some software engineer you'll never know.... and that's final.

 

[MLeeK]

I wouldn't be willing to trust the conversion to the camera. When I do it through the color sliders I control how bright or dark each color in the image is. If the subject is wearing a horrific yellow shirt that would be horribly bright and steal the scene I can tone that back...

 

[Dao]

Data available (Not in scale)

BW jpeg: ##########
Color jpeg: #########################
Color RAW: ##################################################

At the end, if my output need ########## amount of data for the end photo, I'd rather starts from Color RAW.

 

[480sparky39 38]

Data available (Not in scale)

BW jpeg: ##########
Color jpeg: #########################
Color RAW: ##################################################..........

To be more precise:

A B&W JPEG will only have 256 shades of gray available for each pixel.
A color JPEG will have 256 shades of red, green and blue available for each pixel (16,777,218 distinct possiblities).
A 12-bit raw will have 4,096 shades of red, green and blue available for each pixel (68,719,476,736 distinct possibilities).
A 14-bit raw will have 16,384 shades of red, green and blue available for each pixel (4,398,046,511,104 distinct possibilities).

 

[panblue]

My question is if you shoot JPEG in color and then use post processing to convert to black and white, will you get better black and white images that way?

Keith, it depends on how skilled you are at making B&W images from digital captures.

"I wouldn't trust the camera to create the B&W" is a commonly-held and dogmatic (in my opinion ) opinion.
"Channel mixing is superior to greyscale conversion" is also a commonly-held and dogmatic opinion.

What you can make with channel-mixing you can usually do also using curves tool (unless also specify 'preserve luminosity') which then will allow some extra/unreal effects.

You can shoot B&W JPEG using physical filters on your lens and you can also even apply digital color filtering even to monochrome images in certain apps that will change the luminance values even without actual color being displayed. You will lose some light using physical filters but often that isn't a big problem unless you are working very low light. (In fact turn that to your advantage and use a filter to suppress light in some situations; allowing wide aperture in strong light etc).

There is a Cambridge in Color article online somewhere which concludes that greyscale conversion is actually the most visually accurate representation of an otherwise color scene.

Some cameras make great B&W straight out the camera. Others not so much. IMO, if you are good with using curves, a straight desaturated image works fine. Especially if you also use an editor where you can duplicate the image and render the second layer in Overlay mode (~50-80% opacity should work). This gives the image a very dense, 'glossy' look.

Some people desire a deep black point in there monochromes (and feel everyone else should have to). Other people only render true blacks where the scene in reality possesses that. IMO people that grew up shooting film appreciate tonal gradation in shadows yet it seems many digital photographers want to render images higher contrast. Just an impression I get.

Using levels tool, lifting the output slider for the black point will soften shadows (and can mask slight posterization ...incidentally, camera picture style setting for JPEGs: some cameras seem to posterize shadow if the picture style contrast setting is too high. Even with in-camera JPEG, set the picture styles low contrast and low sharpness and do those in post).

 

[panblue]

All digital cameras with a Bayer, outputing a B&W JPEG, captured at 12/14bit colour raw.

Keith..try all the options. You can create great monochrome from raw as suggested in this thread and you can also get very nice B&W from some cameras (Canons have often color filters in B&W picture style menus). Experiment!

{QUOTE=480sparky;2737889]

Data available (Not in scale)

BW jpeg: ##########
Color jpeg: #########################
Color RAW: ##################################################..........

To be more precise:

A B&W JPEG will only have 256 shades of gray available for each pixel.
A color JPEG will have 256 shades of red, green and blue available for each pixel (16,777,218 distinct possiblities).
A 12-bit raw will have 4,096 shades of red, green and blue available for each pixel (68,719,476,736 distinct possibilities).
A 14-bit raw will have 16,384 shades of red, green and blue available for each pixel (4,398,046,511,104 distinct possibilities).[/QUOTE]

 

[unpopular]

As I had said before, if JPEG if your only option, there is merit to SOOC, including for b/w.

But really this is about options, and if you don't have the appropriate black and white optical filters, shooting in greyscale is going to be more limiting than any theoretical improvement to IQ.

 

[Studio7Four]

If I may, let me rephrase the OP's question. I'm not sure if this was his intent, but it's at least a related question which I hope someone with appropriate technical expertise can answer.

Let's say that, for whatever reason, you are locked into saving as jpeg. The camera has access to all the RAW data before it makes its B&W conversion. Does this mean that the in-camera conversion can produce subtleties of gray not available if starting from a color jpeg?

To over simplify an analogy, let's let jpeg save data in the primary colors only (red, yellow, blue) while RAW has the additional ability to save in the secondary colors (orange, green, purple). If you start from a color jpeg, with only three different original colors, only three different shades of gray are possible, whereas if you start from the RAW then six different shades of gray are possible. Or is it that, to complete this analogy, saving the in-camera six-shade B&W as a jpeg would then knock those six shades down to three anyway and the net effect is the same?

 

[KmH]

The in-camera conversion is basically done by just de-saturating the colors. I don't shoot Canon gear so I don't know what B&W picture styles Canon offers.
At any rate, in-camera B&W photos tend to lose a lot of the contrast the color image had, and if in camera conversion offers the photographer any adjustments, they will be rather crude.

Converting to B&W post process from the color image gives more control over the contrast in the final image.
Photoshop offers a non-destructive B&W Adjustment layer that allows the image editor to individually control the density of 7 colors.
Nik software also offers some nice B&W conversion software.

It is my opinion that making a really good B&W image is quite a bit more difficult than making a really good color image

 

[unpopular]

saving the in-camera six-shade B&W as a jpeg would then knock those six shades down to three anyway and the net effect is the same?

No. Because all color images are composed of three b/w channels which can be put together in arithmetical ways to produce a b/w image; channel mixer is the most common. The problem is that these methods are destructive (in the sense that they compress, stretch or clip tonal data), and their execution may result in artifacts, this is the case with RAW, especially if you use blue channel data which tends to be the worst quality.

This is doubly problematic since JPEG is 8-bit and already compressed, usually hiding compression artifacts in areas of smooth graduation without fine detail. Once you adjust these areas' contrast, the compression artifacts become more apparent.

However, the real world, this may be splitting theoretical hairs.

 

[480sparky39 38]

....data in the primary colors only (red, yellow, blue) while RAW has the additional ability to save in the secondary colors (orange, green, purple). .......

Technically, our primary colors are red, green and blue. The 'secondary', or complimentary colors, are cyan, magenta and yellow.

........ Or is it that, to complete this analogy, saving the in-camera six-shade B&W as a jpeg would then knock those six shades down to three anyway and the net effect is the same?

It's hard to explain easily, but reducing an image to only six shades of gray means you're reducing it to a 3-bit image....... far far less than an 8-bit JPEG.

 

[Studio7Four]

Sorry, but it's been a long day and I'm feeling a bit testy.

....data in the primary colors only (red, yellow, blue) while RAW has the additional ability to save in the secondary colors (orange, green, purple). .......

Technically, our primary colors are red, green and blue. The 'secondary', or complimentary colors, are cyan, magenta and yellow.

If by "our primary colors" you mean the colors recorded by a digital sensor (with its 1-1-2 Bayer mask pattern), you are correct. I was referring to the common color wheel (Color Wheel, Color Circle, & Color Relationships), in which the primary colors are those pigments which cannot be made by mixing other colors - red, yellow, blue.

........ Or is it that, to complete this analogy, saving the in-camera six-shade B&W as a jpeg would then knock those six shades down to three anyway and the net effect is the same?

It's hard to explain easily, but reducing an image to only six shades of gray means you're reducing it to a 3-bit image....... far far less than an 8-bit JPEG.

Oh, I understand bit counts. I was admittedly over simplifying for the sake of analogy. My 3 shades of jpeg color and 6 shades of RAW color of the initial premise are also far, far less than their 8-bit and 12-bit analogs.

Keith, sorry for hijacking your thread - poor form, especially to do it to a fellow M@$$hole. I'll sign off of this thread and return it to its original path.

 

[Garbz]

To add something to the original thread there is one area where shooting in JPEG will be severely limited compared to RAW, that is if you attempt to imitate a Wratten 47B (blue) filter. The JPEG algorithm completely destroys the blue channel of an image. This is acceptable in colour images because our eye is far more luminance sensitive to red and green.

A RAW file won't compress the blue channel to buggery so you'll get a nice sharp image. A JPEG file on the other hand will suck for conversion using a blue channel.

 

[Ysarex85 79]

Sorry, but it's been a long day and I'm feeling a bit testy.

....data in the primary colors only (red, yellow, blue) while RAW has the additional ability to save in the secondary colors (orange, green, purple). .......

Technically, our primary colors are red, green and blue. The 'secondary', or complimentary colors, are cyan, magenta and yellow.

If by "our primary colors" you mean the colors recorded by a digital sensor (with its 1-1-2 Bayer mask pattern), you are correct. I was referring to the common color wheel (Color Wheel, Color Circle, & Color Relationships), in which the primary colors are those pigments which cannot be made by mixing other colors - red, yellow, blue.

Physiologically the primary colors are Red, Yellow and Blue. This is actually a fascinating topic because there is compelling evidence to support both positions. The primary colors are Red, Yellow and Blue and the primary colors are Red, Green and Blue. If you explore human perception physiologically then you'll discover that the after color image people see relative to blue is orange and the after color image seen relative to yellow is purple -- hard evidence to conclude the primaries are Red, Yellow and Blue.

On the other hand we measure color temperature under the light of our sun (the ultimate light source) on a scale that is Yellow/Blue not Orange/Blue. Look at any raw file converter and you'll see a temp slider that is Yellow/Blue with a second option for tint that is Magenta/Green -- hard evidence to conclude the primaries are Red, Green and Blue.

Artists et al. have been arguing this one since the day Isaac Newton split light through a prism 300 years ago and drew this:



And then James Maxwell made this:

Joe