rec709 LUTs - Fuji color profiles

[Don Kotlos]

http://menexmachina.blogspot.com/2018/07/rec709-luts-fuji-color-profiles.html

[cantsin]

@Don Kotlos: Nice work, and thanks for sharing it! Tried your LUTs in Resolve, and they create beautiful results. - I encountered only one issue: your LUTs seem to cause color banding in blue highlights. Have you encountered this problem, too?

[Deadcode]

On 3/18/2016 at 5:50 AM, Don Kotlos said:

I did some quick reverse engineering of some of the Fuji profiles and created rec709 LUTs that can achieve similar results. No Fuji camera necessary and no need to shoot with a baked-in profile!

Here is an example applied to a clip from A7rII:

Long story short took advantage of the fact that Lightroom allows color profiles to be used with RAW files. Then I computed the three-dimensional affine transformation that can remap the default color space from adobe to any of the Fuji profiles. 

I can do a much better job if I find color samples that span most of the 3D space. For this I used 4 sample images from dpreview and imaging resource. 

The good thing is that I can use the program that I wrote to compute any transformation that is needed. For example S-log/s-gammut to rec709, or slog to clog, rec709Sony to rec709Fuji etc... I only need two images from the two profiles with the same color information such as a colorchecker and of course neutral WB and same exposure.

 You can find the LUTs attached. I hope you enjoy them! Don't forget you can use them on images too with Photoshop...

Acros.cube

Astia.cube

ClassicChrome.cube

Provia.cube

Velvia.cube

 

I don't understand you perfectly. What do i have to give you to create a proper Fuji LUT? I have several records with colorchecker, few raw files, and lot of .mp4's with different picture profile settings from Panasonic LX100. If i give you a dialed down clip from natural profile (aka. FLAT) you can create a LUT from it to look like a Fuji Provia?

edit: thanks for your hard work!

[cantsin]

@Deadcode: an mp4 file can never span the complete color space needed for a good LUT design since its color is heavily compressed to 8bit 4:2:0 in Rec709 color space.

[Don Kotlos]

12 hours ago, cantsin said:

@Don Kotlos: Nice work, and thanks for sharing it! Tried your LUTs in Resolve, and they create beautiful results. - I encountered only one issue: your LUTs seem to cause color banding in blue highlights. Have you encountered this problem, too?

I did not see any banding on the brief tests that I did. My guess would be that it depends on the rec709 base file and how close it is to the Fuji. I would either try reducing the saturation/contrast of the base file a bit before applying the LUT or you can lower the intensity of the LUT (I don't know how to do that in resolve though...). 

The problem is that the blue channel had the most precision after the affine transformation so it wouldn't matter if I recreated the LUTs with more images. But if it is really bad maybe it needs fine tuning with your specific profile. Do you see it with many LUTs? 

6 hours ago, Deadcode said:

I don't understand you perfectly. What do i have to give you to create a proper Fuji LUT? I have several records with colorchecker, few raw files, and lot of .mp4's with different picture profile settings from Panasonic LX100. If i give you a dialed down clip from natural profile (aka. FLAT) you can create a LUT from it to look like a Fuji Provia?

Deadcode these can be applied to any rec709 image, even with a natural profile. The colors are going to be close but not exactly the same as the Fuji ones since each camera has a unique way of responding to colors. To get exactly the same colors I would need the same image of a colorchecker with a Fuji and your LX100 with the profile that you use. Then I can do the rec709Pana-->rec709Fuji LUT, and when you apply the LUTs on the top of the page (which are actually doing the rec709Fuji-->Fancy color) and then that should be much closer to the Fuji's color. 

I wouldn't worry that much with that though since you would lose the ability to control the image (contrast/dynamic range/saturation). I find these LUTs more pleasing when applied to a less contrasty/colorful image giving a more subtle result than the Fuji. 

[cantsin]

@Don Kotlos: I created the Rec709 image from Blackmagic Pocket CinemaDNG material using an X-Rite Color Chart and Resolve's automatic color chart matching function, then your LUT(s) on a second node. (Yes, it is possible to dial down the effect of a LUT in Resolve by dialing down the strength of the node it sits on.)

[Don Kotlos]

59 minutes ago, cantsin said:

@Don Kotlos: I created the Rec709 image from Blackmagic Pocket CinemaDNG material using an X-Rite Color Chart and Resolve's automatic color chart matching function, then your LUT(s) on a second node. (Yes, it is possible to dial down the effect of a LUT in Resolve by dialing down the strength of the node it sits on.)

Can you send me a frame after the color matching but before the lut?

[cantsin]

8 hours ago, Don Kotlos said:

Can you send me a frame after the color matching but before the lut?

Here are some frames. Admittedly, they are extreme - because they just use a primitive color-chart based color conversion to Rec709, resulting in blown-out highlights (that I could have pulled back/recovered in Resolve). Nevertheless, you see how the overexposed areas of the image create strange blue/green banding artifacts with your Fuji LUTs. I'm posting 50% scaled down jpeg screengrabs here, but could also send you 1080p 16bit TIFFs + the out-of-the camera DNG frame.

- The first image is the 'pure' Rec709, the following images are the same image with your LUTs applied:

 

Rec709

 

Astia

 

Provia

 

Classic Chrome

 

Velvia

[Don Kotlos]

20 hours ago, cantsin said:

Here are some frames. Admittedly, they are extreme - because they just use a primitive color-chart based color conversion to Rec709, resulting in blown-out highlights (that I could have pulled back/recovered in Resolve). Nevertheless, you see how the overexposed areas of the image create strange blue/green banding artifacts with your Fuji LUTs. I'm posting 50% scaled down jpeg screengrabs here, but could also send you 1080p 16bit TIFFs + the out-of-the camera DNG frame.

- The first image is the 'pure' Rec709, the following images are the same image with your LUTs applied:

Velvia

So it took me few hours but I ve managed to fix the most of the crazy colors.

One problem with these color matching techniques is that you don't sample adequately the whole space so some times interpolation can create spurious results. It is similar to the crazy coloring of the highlights in the original files that occur when Resolve tries to do a similar thing . Since there were not any colors like that in the images that I used, the points were off. 

I restricted the LUT to more reasonable values and here is the Velvia example of your original image:

 

I cannot do better than this... I updated the original files so you can download them again...

[cantsin]

@Don Kotlos: So many thanks for this! Great job!

[Attila Bakos]

When you create your own luts, it's always advisable to check color transitions using an image like this:

I applied your ClassicChrome lut to this, and this is what I got:

You want to see about the same smoothness as the original image, but with different colors.

[Don Kotlos]

Yep, these automatic methods of LUT generation have a problem at the boundaries of the color spectrum. I tested with many videos and images and I found it fine, but I did not with a gradient like that. In your test images the first is the Hue to Luma with max Saturation, and the second Hue to Saturation with mid Luminance correct?

[Attila Bakos]

It's not my image but from the looks of it the left part has 0% luminance and 100% saturation at the bottom. As we go up vertically, luminance rises, and at the half of the image it reaches 100%. Then saturation is dropped until the top is reached. The right image has 50% luminance at the bottom, and 0% saturation. As we go up vertically, they both rise, at the top they are both at 100%.