Understanding BLEND Modes: the AVERAGE blend mode (Affinity Photo)

[@ptst19]

Looking forward to seeing the next, thank you. I think the apply image filter is worth a dedicated tutorial. I am really impressed how much you know about SP

[@Graphicalfinity]

Awesome, thank you! The Apply Image Filter is indeed a beast and is definitely worth a tutorial in the future :)

[@pedropuckerstein4670]

Excellent because you exposed the aspect of calculation priorities which in general is not clear in many affinity photo operations. I wonder if a rule book has been created by anyone.

[@Graphicalfinity]

Thank you!

[@TV-tq5ed]

Loved it! Looking forward to more of them when you have the time...

[@Graphicalfinity]

Awesome! Thank you!

[@nitareed4088]

This will need to be watched several times to understand...

[@Graphicalfinity]

Imaginable... the blend mode theory can be confusing indeed :)

[@simonethoth3371]

This was huge Really accurate Thank you

[@Graphicalfinity]

Thank you!

[@enthusiaticbeginner]

Very well explained, thanks! Just one query - does the priority processing being 1) Blend Mode 2)Blend Ranges, 3) Opacity -- apply to all Blend Modes or just "Average"? I am guessing that it does apply to all.

[@Graphicalfinity]

Yes, correct. It should apply to any blend mode.

[@selenitafm]

Thanks a lot!

[@Graphicalfinity]

You're welcome! Thank you

[@SouthbayCreations]

Thank you for the video!

[@Graphicalfinity]

My pleasure!

[@gossedejong9248]

thank you!!

[@Graphicalfinity]

Welcome!

[@jed1947]

Hello, am I misunderstanding this? You talk about blend ranges and opacities without really emphasizing to which layer or underlying composition to which they apply. What I get out of this is that a blend range applied to group is processed after the child layer opacities within the group are applied. This seems obvious that child layer rendering would occur first, before any group composition rendering, and this order of processing could certainly affect the final render. This does not prove that average blending provides a different render than a 50% opacity reduction of the THE LAYER TO WHICH THE OPACITY BELONGS.

[@Graphicalfinity]

Thank you for your feedback. Regarding your question...your understanding about the group/child rendering priority is totally correct, however this was not the point i was trying to make. I used a group to separate the steps of rendering to get the same result using opacity instead of average blend mode. Maybe i have not understand your last remark correctly, but as you can clearly see in the video, there is a difference in 50% opacity+Normal and 100%opacity+Average when a blend range is applied. The reason is because a layer is processed in the following manner: 1. Blend Mode, 2. Blend Range and last (3) Opacity. This is why i recreated it using a group where the child with 50% opacity+normal (which equals to blend mode average and 100%opacity) and then the blend range on the group, to show that blend mode is applied first and then the blend range. So in the case of 50%opacity+normal, the program first applies the blend range, which removed the bright/white areas (making them black) and then opacity of 50% making the image darker (as the background is black) In the case of 100%opacity+average, the blend mode is applied first, where the white areas have become grey first, then the blend range is trying to remove the white, but as there are less white areas, so the result of the blend range is different then in the first case, so this imho proves that the statement 50%opacity equals average blend mode is not always true.

[@markoehlschlager3862]

Excellent. If I understand correctly, the Blend Range curve adjustment applied to the left (Normal) image operated on the image as if the Opacity of the image were 100%, thus making the white pixels transparent. On the other hand, the Blend Range curve adjustment applied to the right (Average) image operated on the computed average of the image, and because there were no pure white pixels in the averaged layers, the Blend Range curve adjustment produced no transparency. Is that correct? Love the practical application of the Average Blend Mode to grey the model's hair.

[@Graphicalfinity]

Thank you! Regarding your question.. Yes exactly! I could not have said it better :) The sequence of operations is: Blend Mode Blend Range Opacity So just as you said, in the case of the left image, Blend mode was normal (original image), Blend Range made the whites transparent and opacity lowered the transparency overall, making everything dark Right image (average), Blend mode averaged the pixels (making everthing darker), Blend range has much less pixels it can be applied to..

[@markoehlschlager3862]

@@Graphicalfinity As I remarked above, your practical example of finding the average color value between grey and the model's natural hair color is a very good use of the Average blend mode calculation to produce a convincing aging effect. I struggle to think of any other applications. I suppose one way to think about the use case for this particular Blend Mode would be to use it with a masked fill layer to pull selected portions of a base image halfway toward the color of the fill layer. Perhaps, if one is producing flat graphic art, one might want to generate a third color that is the exact average of two other colors in a logo? Other than that, the general observation here is that the average calculation will always reduce the dynamic range of the result (assuming you are not blending black with black, or white with white), generally producing results that are more muted (midtones and less saturated).

[@Graphicalfinity]

@@markoehlschlager3862 Indeed, there are not much use cases of the average blend mode. As you have mentioned, the most common use will be in graphic art, where you could use it for creating pastel like colors or finding color harmonies. I sometimes use it to give objects a little shine/colortone but always in combination with a blend range on the source (as this is where the real advantage lies vs opacity). Also it works well with motion blur like effects you want to lower down with blend range. I will post an example image in the channel.

[@markoehlschlager3862]

@@Graphicalfinity So, in summary, I learned the following essential facts about the Average blend mode: 1. the Average blend mode takes the color information from a top blend layer and the underlying base layers and computes the average color at each pixel. (a + b)/2 2. the order of computational blending effects is Blend Mode, then Blend Range, then Layer Opacity. 3. the result of an Average blend mode will always produce a more muted and less saturated effect. 4. the most practical use cases for the Average blend mode would be either a) to use masked fill layers to pull portions of an underlying image halfway toward the fill layer color (a kind of local color grading), or b) to produce color palettes where two parent colors can produce a third color that is the exact average (or midpoint) of the two parent colors.

[@markoehlschlager3862]

As I wait for your next Blend Mode tutorial with great anticipation, I was thinking that some of the blend modes have more of a diagnostic or analytical purpose than a creative purpose. For example, take the "Darker" and "Lighter" blend modes. Although I wouldn't rule out a creative use for either of these two blend modes, it strikes me that they have more of a diagnostic/analytical purpose whereby a user might use a colored fill layer set to "Darker" as a binary test threshold to identify areas of the underlying image layers that are darker than the test color set in the fill blend layer. The essential test function of the Darker blend mode is to compare the color values of the blend layer pixel with the base layer pixel and display the darker of the two. Show me those areas of the base layer image that are darker than my test color. The same would be true for the use of the "Lighter" blend mode to identify areas of the base image layers that are lighter than the test color set in the fill blend layer. Show me those areas of the base layer image that are lighter than my test color.