@jandren will there?
There curves tool that runs in a browser is really coolā¦you can compare filmic with sigmoid. Adjust the parameters and even compare to blender and other curve modelsā¦
At the very least it is a quick visual answer to Terry and and Tony wrt what do the parameters doā¦ from a tone perspectiveā¦
https://jandren-tone-curve-explorer-streamlit-app-xca32q.streamlit.app/
The 12642 PR is a pretty good start ā¦just needs some between the lines and some formatting to fit into the DT manual formatā¦
6 Likes
Well volunteered
1 Like
Both Filmic and Sigmoid are tone-mappers. Their job is to map the linear, scene-referred data to non-linear, display-referred data that your monitor (or print medium) can display correctly. They just do it in slightly different ways, providing different trade-offs.
Since it makes no sense to tone-map twice, only one tone-mapping module should ever be active.
If you donāt understand the scene-referred workflow and the role tone-mappers such as Filmic and Sigmoid play, I highly recommend the following:
3 Likes
Thanks. Useful. I āunderstandā filmic
Actually, the job of mapping tone to the rendition medium is the job of the color profile specific to the medium. Filmic, Sigmoid, and other user-futzed tone curves are an intermediary between linear and input to the rendition transform. If one is not using calibrated display profiles, these tools are probably filling the gap, but one has to be wary of exporting to JPEG and then wondering why the tone looks differentā¦
3 Likes
See this post, where Jakob describes his recommended workflow (heās answering several questions, so youāll have to scroll down a bit):
1 Like
Yes, I will supply the documentation asap! And I will do my best to explain the module in a more approachable summarized manner than the monster thread.
The short answer is that it should essentially be the same as filmic, but faster to work with.
Thanks everyone for quickly filling in with answers 
6 Likes
Yup, specifically compressing the dynamic range of the image into something that āplays niceā with a limited-capability display.
If you have an HDR-capable display, you can actually get really good results with no tonemapping whatsoever. For example, convert scene-linear data directly to HLG and feed to an HLG television with no tonemapping - works amazingly well. (For the time being, the only consistent way to do this is to encode an H.265 video with Rec. 2020 gamut and HLG transfer function, and make sure to set the ATC SEI metadata properly. Some people claim that HLG requires no metadata, but thatās a lie - if you donāt have the ATC SEI metadata itāll be treated as Rec709 and look āmehā because HLGās backwards-compatibility mechanism is actually a kinda poor tone curve with no āfootā.
But for 99% of stills content delivery, you need a good tonemapper. Honestly for the few cases Iāve seen sigmoid not perform great, you need a local and not global tonemapper. dtās local contrast is one option, using enfuse on multiple images exported with different sigmoid settings would be another.
The funny thing about it is that the gamma curve that is part of display/export profiles was a CRT-specific thing; Iāve been told raw LCDs do not manifest this characteristic. So, the display manufacturers build in the gamma TRC for compatibility, and we have to mung our linear captures to conform, stillā¦ 
Thanks, Troyā¦ 
Iāll have to go back and read what we have for tone-mapping and filmic. To me it makes sense to try and integrate sigmoid from the point of how the current tone mapping options are currently explained ie around why we use filmic and now saying why we could also use sigmoid as an option. This will cover both in a big picture sense and then the collective language about how to describe and present those options and the relative differences,merits etc would have to be ironed out. Then a module specific description of the sliders etc would need to be provided for that section of the moduleā¦
3 Likes
Itās been around for months and months . But it raises a discussion where some people thought it should not be included, or it base functions merged into filmic as a separate filmic mode.
Without going into details , that discussion is no more so it go finally approved to be included. So its now in the main DT source code and will be one of the new features for 4.2.
2 Likes
Whatās missing from most of these tone curve discussions is characterization of the tone curve itself. When you read the seminal literature about filmic-type curves, the attention is directed at the toe, sloped such to ācrispā the darkest tones. Sigmoid, to my limited understanding, doesnāt push highlights to saturation as quickly. And so onā¦
All this āgamut protectionā, āexposureā, and āblack/whiteā is secondary, or worse, to the aboveā¦
1 Like
Thanks for developing the sigmoid module. I look forward to testing it and reading the documentation when it is prepared. At least the answers here have got me started. I am also sure as it is becomes implemented in the various builds there will be some constructive discussions and examples showing both the strengths and weaknesses of Sigmoid and Filmic options for tone mapping.
This is underway in the main discussion which got long and the new one I started for posts that were evaluating the sigmoid moduleā¦so this is well underwayā¦ I think the real thing that will come to light is filmic fatigue. People might be thinking oh no not another tone mapping module to master expecting a lot of reading and theory. I think ease of use will be the real strength of this module. Where filmic was involved and technically precise and there were many additions renditions, fixes and nuances, sigmoid will be much simpler and straightforward out of the gate.
Because there were so many features in filmic many people were trying to get a finished image from it. This often led to endless rounds of tweaking all the various parameters and then coupling this with how filmic was implemented with various desaturation and hue preservation I think left some people chasing their tail cycling through combinations of settings. In fact for many instances really it could have just been left in default activated with maybe a tweak of the white relative to do its job as a global tone mapping editā¦ This is an over simplification but I think you will see that you will enable sigmoidā¦you might be happy right away. Then jig contrast a little and if necessary skew. I think @Jandren has mentioned to use this more as a touch up or final tweak near the end of the edit. Likely for a few images you will try the color pres at 0 ,50, and 100 to see what it does and decide then how to use it in the imageā¦ so very easy. One thing mentioned was the suggestion by @Jandren to use the contrast slider in rgb CB with the module to add contrast. I have often done this any way even though it was not generally recommended. I found that something between 0-10 % was always a nice tweak and helped to avoid pushing local contrast too hard but that could just be my taste for contrast. So far I have not noticed any weird artifacts or unpredictable color issues. The sliders pretty much do what you would expect. Tone eq will likely be useful when highlights are blown or maybe this will prove to be the time for filmic with its extra tools and controlā¦ I guess the point of my comments are that filmic suffered all the slings and arrows of transitioning away from a base curve and display referred workflow that lead to all that āthis doesnāt look like my jpgā stuff. I think it also came with the new changes and module of color calibration and rgb color balance and diffuse and sharpen and now HLR etc etcā¦ all of which had their own issues that could be impact filmic and the reverse. So I think getting used to Sigmoid will be far less involved with its simplicity and coming in to this more stable platform and this will make some happy and others will likely prefer to stick with filmic and the level of control it provides along with how they have come to master itā¦
3 Likes
Personally, my workflow with filmic is a matter of a few seconds per image, so I find it fairly easy to use. I donāt object to sigmoid but I donāt really intend to use it.
Agreed.
3 Likes
Yeah, I saw when it was first mooted, but compared with filmicā¦
I teach students DT and teaching filmic is a real challenge, but to date they 100% preferred the result compared to base curve. Having a simple alternative to teach sounds good to me. I have learnt to be very comfortable with Filmic and have found that for many images just setting exposure correctly and then using the auto tune levels option in Filmic works really well as an initial editing point.
I will invest the time learning to use the new sigmoid option and try to understand its strengths and weaknesses. I processed one image with it yesterday and plan to do more with it very soon. It did appear easy to use.
Agreed, at first with filmic I thought I had to tweak a bunch of the parameters every time. Then realised that mostly wasnāt really its function and it could largely be left to do its job, beyond the obvious stuff that it couldnāt easily figure out for itself
1 Like
Futzing around with V6 forced me to revisit v5 and I came to like it and even actually understand it a bit moreā¦ For me in v6 maxrgb might be good at times for color but it kills fine detail IMO. And moving to no or luminance is not always great eitherā¦ In v5 for the most part I would use no and then manipulate the curve and saturation to manage this either up or downā¦ In v6 I find you need to have the relative white too wide to get much out of the extreme luminance slider further killing the detailsā¦ but all of this could be on my foiblesā¦
I actually like fairly rich saturated / contrasted imagesā¦no so much so I have solid blotches of color but likely in a way that gamut control is working against meā¦in many images I guess the color is correct using v6 filmic but it can seem out of whack to the eye as badly as what people would complain about for a bad camera profile. It is generally easy to fix and so you can apply and appropriate and targeted correction but there are times esp for most everyday shots that this is not really requiredā¦ I think @bnpndxtr 's tweak of clearing the gamut check would be a nice thing to have as a checkboxā¦just as you can dial in the the compression in color calibration or set it to zero or turn it off in the input profileā¦ Those that want to use it consistently as a check could and those that found it too intrusive to take it awayā¦ but thatās just some random musingā¦