grayscale vs desaturation

#1

In PS, when I change a photo to grayscale and desaturate it respectively, I got different results.

Anyone can tell me more about their mechanisms?

Which method would u prefer when u need to convert a colored photo to B/W photo?

#2

Desaturate has no regard for the actual color information in the image. If different colors have the same Intensity value, they will all desaturate to the exact same shade of gray.

The Grayscale command has a more realistic conversion map that tends to make a more natural conversion to black and white.

Blue tends to appear darker than green or yellow, even if they have the same Intensity values.

The best method is using the Black & White command (Image > Adjustments > Black & White…)

This gives you complete control over the intensities of the gray values based on the color information in the image.

#3

It helps me. eikonoklastes, thank you for your explanation!

#4

Personally, I prefer working in LAB, channel mixing or desaturate as it give me greater control and options over adjustments overall.

Desaturation is actually a more “realistic” expression of how film and human eyes would interpret natural light desaturation. It is hue independent and relies almost exclusively on brightness. So hues of the same brightness will look identical when completely desaturated. However, it does add a bit of gain to the image.

Another option is working directly in channels either in RGB or LAB. You can desaturate/adjust individual channels for desired effect or pla with the L channel in LAB which is pure brightness.

There is nothing particularly wrong with grayscale, but it will adjust/discard the images natural tones during the conversion, taking those tonal adjustments out of your hand. So it depends on what you are aiming for.

For critical black and whites, most professional photographers I know tend to A) Actually SHOOT in BW (nothing beats it) or B) prefer channel/desaturate with curves/levels, as they like to have complete control of the image and not loose that valuable tonal data. However, you do have to be keenly aware of what you are doing to get the most out of it.

Nothing beats actually shooting in BW but in the end there are many ways to do it and it all depends on what kind of final output you need.

#5

BoBoZoBo, I am interesting to LAB. The main advantage of using LAB is it because it has an independent ligntness channel?

Any drawbacks of using LAB?

#6

LAB has no real drawbacks.

You just need to be aware, that mathematically (theoretically) speaking, LAB has the largest color space but in practicality it means little unless you are converting your correction to CMYK or a very restricted RGB compression. But it’s almost a non-issue.

The best thing about lab is the L channel is pure luminance, allowing you to extract a really pure value base.

What I normally do is open the art, then make a duplicate in LAB, extract or copy the L channel and use it as a base to correct the file in either LAB or back to RGB.

There are a few books on the use of LAB and if you are serious about color correction it is definitely a space you want to familiarize yourself with as it will exponentially expand your options and ability to select ranges you want to change.

#7

just a side not on this topic, I find that when adding a diffuse glow effect on b&w images you get really weird results doing it to a desaturated image. You get the proper effect applying the effect on a gray scaled image. Why is this? anyone know?

#8

aesnakes, as eikonoklastes said…
“Desaturate has no regard for the actual color information in the image. If different colors have the same Intensity value, they will all desaturate to the exact same shade of gray.”

I have tried convert a colorful flower photo to B/W photo by “grayscale” and “desaturate” respectively. When compared the results, I won’t use the “desaturate” method anymore.

BoBoZoBo, as I know, LAB mode has the largest color space when compared to sRGB and adobeRGB etc…so, it should have a larger file size in PS, right? But when I convert the same photo to “RGB Color” and “LAB Color” (both still psd format), their file size nearly the same, but the LAB one is a little bit smaller why don’t the LAB one bigger?

In PS, there are many working color spaces for RGB, CMYK and even Gray. Why LAB doesn’t have?

#9

Lab is a linear working space, you will notice, if you get a pms swatch in PS, the colors will also be listed in Lab (in the pallete) for their exact appearance.

Lab is an amazing colorspace, but you wouldn’t know it until you learn more about it. I highly recommend this book by “The Lab” expert.

http://www.amazon.com/Photoshop-LAB-Color-Adventures-Colorspace/dp/0321356780

-Jim

#10

cheung, the LAB size is similar because you still only have 3 channel values and as 3dj pointed out, it is linear. The expanded color space is purely mathematical. Plus, there is more to the way PS native formats code their files than color space. Its not as obvious as it is with JPEG or TIFFs.

As far as Grayscale vs Desaturation. Simply converting is not enough. You need to understand how the colors are changed in each space, and adjust the file accordingly for best results. Different art with different hues and brightness levels will give you different results.

Remember that with grayscale you only have one channel to mix and play with, desaturating will keep your 3 RGB channels to give you better ability to adjust.

However, again, it all depends on what your end needs are.
There is nothing wrong with grayscale if it gives you the results you want/need.

3dj recommends an EXCELLENT book for anyone seriously interested in color correction. I have it myself and he is right, it opens a new world for color correction if you are not familiar with it.

aesnakes - I have not experienced your issues, can you be more specific?

#11

You guys broaden my horizon, Thanks a lot!

LAB is really interesting, I will try that book:)

#12

Glad to hear that. When you get the book, the first 5 chapters are quite readable, the rest, even the experts are recommended to read at-least twice, so dig in and enjoy :wink:

-Jim

#13

Surely the Channel Mixer (under Adjustments) is the way to go. Then you can use Curves if necessary to fine-tune the results. It is only then that you go to Convert to Grayscale (which uses the settings that you have chosen). Personally, I haven’t used Grayscale on its own for years.

You could also look at the SilverOxide filters (www.silveroxide.com). They produce good results and are available for a wide range of emulsion effects. For example, PanX, PlusX and TriX, plus some other more esoteric types. They work in PS CS3, but I haven’t tried them in CS4.

Hope this helps - TL

#14

No Problemo - Glad to hear it. Good luck

#15

The usage of one or the other is dependent on the look you’re trying to achieve. If it’s for the web, I like to use grayscale as it provides better contrast when you incorporate the text, but for all other uses, I usually use desaturation, as someone mentioned above, it’s just more realistic.

#16

Lab’s expanded gamut notwithstanding, does nobody here use the Black & White adjustment layer?

#17

Until PS introduced the black and white conversion, I used to do it in a similar way, sometimes I still do it for the flexibility. Basically, I take an adjustment layer of Hue and Saturation and lower the sat to 0, then below that I throw in a selective colors adj layer and adjust that to get the image I want. also used to use the channel mixer which is probably the best but can get away from you pretty quickly if you don’t watch it.

-Jim

#18

This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.