Thank you Boris. Again, great explanation of the use of the channel mixer. I’ve been using it more the last weeks and cleaned up a lot of muddy colors. It does take some trial and error and practice to apply it wisely. It’s not the easiest tool in the arsenal 😁
That's right, it's not easy but I'm convinced it's part of the basic knowledge when it comes to dealing with colors in the digital world. And that is not only limited to photography. Additive color mixing is essential for color profiles, LUTs, white balance, color grading, etc.
Terrific! Thank you so much for going over this important tool again. You make it seem so easy but it's a hard one to wrap my head around. But this time I think I'm finally catching on! Really, thank you so much for the time and effort you put in to all your videos.
What a master class! This is still very difficult for me to understand, I will need some time at playing around with these concepts to integrate them in a useful way, but I understood for the first time how those levels work. Thank you. Your videos are amazingly instructive.
This method of colorgrading used to drive me crazy, because I could never get it right. It actually took me all instances of your videos on the subject to finally get the hook on this powerful tool😂. Thank you so much for your patience with the lesser gifted among us😅.
Hi Boris, this video is amazing!! Would you consider doing an in depth breakdown of the Color Balance RGB module in the future? You're part of the reason I have grown to love Darktable. Please keep making videos!
I knew how the channel mixer worked, especially the maths aspects, but I've been struggling to use it so far. Thank you for your help. Your video is exemplary. 👌
Thanks Boris for all your videos! They're extremely well thought and helpful and I already learned a lot from them. Would you please consider making a video on denoising? I do a lot of high-ISO night photography, so I have a lot of noisy images and I'm still struggling to find optimal workflow to post-process them. There are so many options to treat the noise in Darktable (raw denoise, denoise (profiled) with wavelets/non-local means/RGB/Y0U0V0, diffuse or sharpen, contrast equalizer) plus the user manual is sometimes contradictory with respect to itself (e.g. regarding the optimal order of different modules within the processing pipeline). I'd really like to see how someone who really knows the tool treats noisy images.
Hi Boris, thanks again for another great video! I still have one thing in my head I think though which I can’t understand. At 15:23 for example when moving the magenta towards red my intuitive thought would be „the flower is predominantly red, that is why I go to red channel of the mixer and remove blue from this channel“ is that also a valid approach? What is the thought process when you go to blue channel instead?
The individual channels, regardless of which input you use, increase or decrease the corresponding color. If you are in the red channel, you can only increase or decrease red there, regardless of whether you use the sliders for input red green or blue. The idea to make the flower with magenta color red was not to decrease red but blue. So I have to go into the blue channel and use input red to decrease blue in the area of the red channel. Decreasing blue in the magenta area (where red and blue are equal) accordingly leads to the red having a higher value than blue and thus the flower becomes more reddish.
I diligently follow your episodes; they are informative, useful and educative. I have only one issue with your videos. The background being gray, and the cursor being tiny, make it difficult to follow the steps. I would go to the extent of saying, they put a strain on the eyes. Would it be possible to change the cursor to make it with a yellow halo, and the background could be a little darker (or with bigger font). It will help immensely. Thanks
What a demonstration! I'm impressed. But since I'm some years older, I think I have to review the video severel times. My mint is not as fast as of a younger man. But it seems the video is able to be a good explanation to me. Thanks a lot.
Very cool technique but hard to get a full grasp on for me. Its like learning a new language to think about colors in this way. What do you think about using something like HSL sliders in Color Zones? Doesn't that achieve much of the same result without the unintended consequences of affecting other colors than intended?
Hi Boris, Thank you for revisiting the Channel Mixer again. I am beginning to establish a better understanding of, however, I note on the 38th minute you move an Exposure instance up the stack. I understood it was always best to leave the stack as set by DT. I don’t doubt your decision to move it, but is there a ‘Rule of Thumb’ as t when modules should be moved or is it documented somewhere. I am not a confident DT user and would not know to make such a decision. Thank you for the time taken to produce such great tutorials.
It is indeed the case that the modules have their given place in the hierarchical stack and that they normally do not need/should be moved. Nevertheless, there are situations when that makes sense. There is no rule for this, but it depends on what changes you want to make in the image. It is important to understand that the modules are stacked hierarchically from bottom to top. Accordingly, the change that one module makes to the photo is then the starting point for further changes to the module above it in the hierarchy. If you then understand what the modules do, you can then restructure them as you see fit. In this case, I used the second instance of the color calibration module, to change color of the sky. And because the sky now differentiates in color from the rest of the photo, it's easier to mask it to change its brightness without affecting the brightness of the rest of the photo. And because you control the brightness with the exposure module, I then took a *new instance* of the exposure module, moved it above second instance of color calibration module, masked the sky and changed its brightness. I emphasize that it is important to know what the modules are doing, because otherwise you risk creating artifacts when you restructure them. That's why moving modules is not generally recommended.
Thank you for this material and the work you put into learning to edit in Darktable. I am amazed at the ease with which you use the channel mixer. Unfortunately, my mind is still unable to grasp the logic of how RGB channels work. I think I will stay with the Color Zone module
I have concluded that one has to think in an almost 3D like manner to develop the necessary colour awareness. Whilst I can describe it, achieving it in reality is somewhat more challenging for me 😂
No, because a) it is adapted to my equipment (camera, lenses) and b) people are attached to presets and no longer deal with modules and c) darktable already offers basic presets as starting points for various use cases, which can then be adapted to your own needs.
Boris, I love your videos, I learn so much every time you post one. For this one I would love to know how to create the RGB image in gimp, or maybe you could post it.
Although this is an excellent program that is on the same level as darktabe in terms of functionality, I use it very rarely to be able to make tutorials for it. That's because darktable satisfies all my needs, so I don't have to use other tools.
I'd notice that after all color corrections and the way you feel harmony in your editing, the histogram show the channels almost in perfect harmony too. Before you does editing channels are unevenly distributed and after almost perfect and even. Nice.👍
I tried making 3 overlapping color circles in Gimp but I could not get the overlapped part to blend the colors. Is there some source where I could download a working gimp example of the 3 overlapping color circles?
Hi I have two questions. 1. How do you get "Advanced Color Options" in Chanel Mixer box in GIMP which you are using here? 2.Why is it written "green for green" instead of "green in green" ( in Green Channel ) Thank you for all your work you put in.
@@kayak58I have stable version 2.10.36 and option is there. I don't know why you don't have it. If you want, you can write a bug report: www.gimp.org/bugs/
I think it's a good explanation, but you didn't really cover how to not destroy atleast one colour with this tool. For example with the 27:59 example, you destroyed the sky. Would've liked to see you fix that
I noticed that you don't use either Filmic or Sigmoid in your recent tutorials. Do you use any Filmic or Sigmoid presets that activate automatically for each photo?
I use filmic all the time. What makes you think I don't use it? You can see that it is always active. And I don't use any preset. Only default setting. What may confuse you is that I don't make any settings on filmic because there is no need for it. But it is always active.
I don't know what is meant by "forest green". If that's what it is: www.colorhexa.com/014421 then the green of your forest needs a bit more blue and not green.
Suggestion for a new topic: A video about the sharpening features in Darktable - e.g. how to avoid the mesh like effect when increasing iterations in Diffuse or sharpen with Sharpness presets, or artifacts with other sharpening tools in Darktable, Contrast equalizer and Sharpen. Remember to zoom in on the subtle differences ;O)
I admire how intuitive seems to you to know what to do to change colors in the directions you desire (and, of course, what is desirable in the first place). Thanks for sharing. [Two cents: I find it sometimes difficult to understand what you're doing with channel mixer (particularly in the last series of videos; I noticed you've been much careful here) because you say, e.g., "we have to add input green" while working on the red channel instead of "we have to add red with the green slider (or to the greenish areas)"]
The fact that the mixer changes all colours in the photo is a distinct problem, next to the very many ways to modify them. When you use the mixer to modify a subject of interest colour, the rest of the photo changes too. All hues are affected, also those you don’t want / need to modify. That means you immediately want / need to modify these unplanned changes. The latter is often not (fully) possible and the technique to counterattack the unwanted changes is not easy: see the Winspeare tutorials on the mixer (a dabble in photography). The changes per channel depend on 6 parameters, which makes it very difficult to predict outcomes. The underlying equations (see Winspeare) harbour three R, G and B estimates and three regression coefficients. The estimates will differ from point to point as the colour differs and can be visualised using the colour picker. The regression coefficients are chosen via the sliders. In the red channel the red slider will increase the red pixels, the other two will swap green / blue pixels for red. The more green / blue pixels are present, the stronger the swapping. It boils down to hard to control changes, difficult to predict. For the three channels, we arrive at 18 parameters in total all, influencing the outcome. In other words: a lot of influencing parameters and many possible outcomes. Winspeare explained that the number of solutions for the equations is infinite. The above characteristics makes the mixer very hard to handle, whatever software you are using. The mixer is useful when colours differ a lot, so the subject of interest changes aren’t affecting other parts much: in a predominantly blue area, a moderate increase in the red channel will not deliver strong changes. Those differences render correction of unwanted effects feasible. If colours don’t differ much, the mixer is problematic. Masking and confining the mixer effect to a smaller area of the photo is often the only / best option and should be used frequently.
When you're not trolling, I find it extremely arogant to think you have to lecture me with your theoretical babble about how channel mixer works and especially when it's useful.
ปีที่แล้ว +1
because of the chanel mixer being complex, I especially value the effort of Boris to explain it. as he demonstrates, the mixer can be very powerful.
@@s7habo I do not think it helps stating that someone is arrogant and lectures with theoretical babbles. On the latter: the Darktable manual on the mixer uses matrix algebra to explain the fundamentals. It is solid science and I added nothing to it. Agreed that the equations can be hard to grasp, but for me as a university based physiologist / epidemiologist it works very well. The equations reveal the character of the mixer and it weakness. I simply tried to explain that for the general audience without any intent to lecture or correct you. I have no need to express an opinion on your approach, nor negative or positive.
That's literally amazing. Thank you so much for making this video. Your series is fantastic- so clear and helpful.
Thank you for taking the time to do this. I found it very helpful, not just for darktable. Keep up the excellent work.
thank you, Boris!
this topic is really hard for me to comprehend.
with your last video I'm getting yet another step closer.
Thank you Boris. Again, great explanation of the use of the channel mixer. I’ve been using it more the last weeks and cleaned up a lot of muddy colors. It does take some trial and error and practice to apply it wisely. It’s not the easiest tool in the arsenal 😁
That's right, it's not easy but I'm convinced it's part of the basic knowledge when it comes to dealing with colors in the digital world. And that is not only limited to photography. Additive color mixing is essential for color profiles, LUTs, white balance, color grading, etc.
Terrific! Thank you so much for going over this important tool again. You make it seem so easy but it's a hard one to wrap my head around. But this time I think I'm finally catching on! Really, thank you so much for the time and effort you put in to all your videos.
Very welcome demonstration, Boris. Thank you.
What a master class! This is still very difficult for me to understand, I will need some time at playing around with these concepts to integrate them in a useful way, but I understood for the first time how those levels work. Thank you. Your videos are amazingly instructive.
Thank you Boris. As usual a very good and useful explanation. Merci.
This method of colorgrading used to drive me crazy, because I could never get it right. It actually took me all instances of your videos on the subject to finally get the hook on this powerful tool😂. Thank you so much for your patience with the lesser gifted among us😅.
This is extremely helpful, THANK YOU! Subbed without hesitation
Hi Boris, this video is amazing!! Would you consider doing an in depth breakdown of the Color Balance RGB module in the future? You're part of the reason I have grown to love Darktable. Please keep making videos!
I knew how the channel mixer worked, especially the maths aspects, but I've been struggling to use it so far. Thank you for your help. Your video is exemplary. 👌
thank you for your videos. you are extremely helpful
Super helpful - thanks Boris
Thanks Boris for all your videos! They're extremely well thought and helpful and I already learned a lot from them.
Would you please consider making a video on denoising? I do a lot of high-ISO night photography, so I have a lot of noisy images and I'm still struggling to find optimal workflow to post-process them. There are so many options to treat the noise in Darktable (raw denoise, denoise (profiled) with wavelets/non-local means/RGB/Y0U0V0, diffuse or sharpen, contrast equalizer) plus the user manual is sometimes contradictory with respect to itself (e.g. regarding the optimal order of different modules within the processing pipeline). I'd really like to see how someone who really knows the tool treats noisy images.
Hi Boris, thanks again for another great video! I still have one thing in my head I think though which I can’t understand. At 15:23 for example when moving the magenta towards red my intuitive thought would be „the flower is predominantly red, that is why I go to red channel of the mixer and remove blue from this channel“ is that also a valid approach? What is the thought process when you go to blue channel instead?
The individual channels, regardless of which input you use, increase or decrease the corresponding color. If you are in the red channel, you can only increase or decrease red there, regardless of whether you use the sliders for input red green or blue.
The idea to make the flower with magenta color red was not to decrease red but blue. So I have to go into the blue channel and use input red to decrease blue in the area of the red channel. Decreasing blue in the magenta area (where red and blue are equal) accordingly leads to the red having a higher value than blue and thus the flower becomes more reddish.
@@s7habo ahhh now I got it, thanks :)
I diligently follow your episodes; they are informative, useful and educative. I have only one issue with your videos. The background being gray, and the cursor being tiny, make it difficult to follow the steps. I would go to the extent of saying, they put a strain on the eyes. Would it be possible to change the cursor to make it with a yellow halo, and the background could be a little darker (or with bigger font). It will help immensely. Thanks
Hi Boris, another big thank you. Yes, I'm one of those users struggling with the channel mixer... 🤔
What a demonstration! I'm impressed. But since I'm some years older, I think I have to review the video severel times. My mint is not as fast as of a younger man. But it seems the video is able to be a good explanation to me. Thanks a lot.
Very cool technique but hard to get a full grasp on for me. Its like learning a new language to think about colors in this way. What do you think about using something like HSL sliders in Color Zones? Doesn't that achieve much of the same result without the unintended consequences of affecting other colors than intended?
Thank you, very useful!
Hi Boris, Thank you for revisiting the Channel Mixer again. I am beginning to establish a better understanding of, however, I note on the 38th minute you move an Exposure instance up the stack. I understood it was always best to leave the stack as set by DT. I don’t doubt your decision to move it, but is there a ‘Rule of Thumb’ as t when modules should be moved or is it documented somewhere. I am not a confident DT user and would not know to make such a decision. Thank you for the time taken to produce such great tutorials.
It is indeed the case that the modules have their given place in the hierarchical stack and that they normally do not need/should be moved. Nevertheless, there are situations when that makes sense. There is no rule for this, but it depends on what changes you want to make in the image.
It is important to understand that the modules are stacked hierarchically from bottom to top. Accordingly, the change that one module makes to the photo is then the starting point for further changes to the module above it in the hierarchy. If you then understand what the modules do, you can then restructure them as you see fit.
In this case, I used the second instance of the color calibration module, to change color of the sky. And because the sky now differentiates in color from the rest of the photo, it's easier to mask it to change its brightness without affecting the brightness of the rest of the photo. And because you control the brightness with the exposure module, I then took a *new instance* of the exposure module, moved it above second instance of color calibration module, masked the sky and changed its brightness.
I emphasize that it is important to know what the modules are doing, because otherwise you risk creating artifacts when you restructure them. That's why moving modules is not generally recommended.
Thank you for this material and the work you put into learning to edit in Darktable. I am amazed at the ease with which you use the channel mixer. Unfortunately, my mind is still unable to grasp the logic of how RGB channels work. I think I will stay with the Color Zone module
I have concluded that one has to think in an almost 3D like manner to develop the necessary colour awareness. Whilst I can describe it, achieving it in reality is somewhat more challenging for me 😂
Hey Boris, thank you for your workflow tips, they are amazing! A quick question: do you provide your presets anywhere?
No, because a) it is adapted to my equipment (camera, lenses) and b) people are attached to presets and no longer deal with modules and c) darktable already offers basic presets as starting points for various use cases, which can then be adapted to your own needs.
Hi Boris, would you be able to go through some street photos and discuss colour grading? one think i find very difficult is editing street photos.
Boris, I love your videos, I learn so much every time you post one. For this one I would love to know how to create the RGB image in gimp, or maybe you could post it.
upload.wikimedia.org/wikipedia/commons/thumb/0/05/AdditiveColorMixing.svg/2097px-AdditiveColorMixing.svg.png
Thank you, outstanding knowledge, can apply to other edit tools, but after Darktable, can you make tutorial for Rawtherapee ?
Although this is an excellent program that is on the same level as darktabe in terms of functionality, I use it very rarely to be able to make tutorials for it. That's because darktable satisfies all my needs, so I don't have to use other tools.
I'd notice that after all color corrections and the way you feel harmony in your editing, the histogram show the channels almost in perfect harmony too. Before you does editing channels are unevenly distributed and after almost perfect and even. Nice.👍
I tried making 3 overlapping color circles in Gimp but I could not get the overlapped part to blend the colors. Is there some source where I could download a working gimp example of the 3 overlapping color circles?
commons.wikimedia.org/wiki/File:Venn_diagram_rgb.svg
Hi
I have two questions.
1. How do you get "Advanced Color Options" in Chanel Mixer box in GIMP which you are using here?
2.Why is it written "green for green" instead of "green in green" ( in Green Channel )
Thank you for all your work you put in.
1. click on the plus sign in "Advanced Color Options". 2. This is simply the error in the labeling.
The problem is that I have no "Advanced Color Options" at all. I use the GIMP version 2.10.36(revision 1)
@@kayak58I have stable version 2.10.36 and option is there. I don't know why you don't have it. If you want, you can write a bug report: www.gimp.org/bugs/
Thanks 👍
I think it's a good explanation, but you didn't really cover how to not destroy atleast one colour with this tool. For example with the 27:59 example, you destroyed the sky. Would've liked to see you fix that
I noticed that you don't use either Filmic or Sigmoid in your recent tutorials. Do you use any Filmic or Sigmoid presets that activate automatically for each photo?
I use filmic all the time. What makes you think I don't use it? You can see that it is always active. And I don't use any preset. Only default setting. What may confuse you is that I don't make any settings on filmic because there is no need for it. But it is always active.
When I try to add green to my forest shots it give me a primary green. How do I add a forest green?
I don't know what is meant by "forest green". If that's what it is: www.colorhexa.com/014421 then the green of your forest needs a bit more blue and not green.
thanks, I'll try adding blue.@@s7habo
Suggestion for a new topic: A video about the sharpening features in Darktable - e.g. how to avoid the mesh like effect when increasing iterations in Diffuse or sharpen with Sharpness presets, or artifacts with other sharpening tools in Darktable, Contrast equalizer and Sharpen. Remember to zoom in on the subtle differences ;O)
Good idea, I'll keep that in mind!
I admire how intuitive seems to you to know what to do to change colors in the directions you desire (and, of course, what is desirable in the first place). Thanks for sharing.
[Two cents: I find it sometimes difficult to understand what you're doing with channel mixer (particularly in the last series of videos; I noticed you've been much careful here) because you say, e.g., "we have to add input green" while working on the red channel instead of "we have to add red with the green slider (or to the greenish areas)"]
The fact that the mixer changes all colours in the photo is a distinct problem, next to the very many ways to modify them. When you use the mixer to modify a subject of interest colour, the rest of the photo changes too. All hues are affected, also those you don’t want / need to modify. That means you immediately want / need to modify these unplanned changes. The latter is often not (fully) possible and the technique to counterattack the unwanted changes is not easy: see the Winspeare tutorials on the mixer (a dabble in photography).
The changes per channel depend on 6 parameters, which makes it very difficult to predict outcomes. The underlying equations (see Winspeare) harbour three R, G and B estimates and three regression coefficients. The estimates will differ from point to point as the colour differs and can be visualised using the colour picker. The regression coefficients are chosen via the sliders.
In the red channel the red slider will increase the red pixels, the other two will swap green / blue pixels for red. The more green / blue pixels are present, the stronger the swapping. It boils down to hard to control changes, difficult to predict.
For the three channels, we arrive at 18 parameters in total all, influencing the outcome. In other words: a lot of influencing parameters and many possible outcomes. Winspeare explained that the number of solutions for the equations is infinite.
The above characteristics makes the mixer very hard to handle, whatever software you are using.
The mixer is useful when colours differ a lot, so the subject of interest changes aren’t affecting other parts much: in a predominantly blue area, a moderate increase in the red channel will not deliver strong changes. Those differences render correction of unwanted effects feasible. If colours don’t differ much, the mixer is problematic. Masking and confining the mixer effect to a smaller area of the photo is often the only / best option and should be used frequently.
When you're not trolling, I find it extremely arogant to think you have to lecture me with your theoretical babble about how channel mixer works and especially when it's useful.
because of the chanel mixer being complex, I especially value the effort of Boris to explain it. as he demonstrates, the mixer can be very powerful.
@@s7habo I do not think it helps stating that someone is arrogant and lectures with theoretical babbles. On the latter: the Darktable manual on the mixer uses matrix algebra to explain the fundamentals. It is solid science and I added nothing to it. Agreed that the equations can be hard to grasp, but for me as a university based physiologist / epidemiologist it works very well. The equations reveal the character of the mixer and it weakness. I simply tried to explain that for the general audience without any intent to lecture or correct you.
I have no need to express an opinion on your approach, nor negative or positive.
"promo sm" 😩