… which is just another sign that Andrew’s clickbaity style worksbetter. Otherwise, he wouldn’t get 10x as many views in only two days with only 2x as many subscribers.
The term “filmic” has been thrown around in CG circles for a while now, so if you knew what to look for, you would’ve found this addon. Most people don’t know what to look for, nor do they look. It is a revelation to them.
If you look at the CG Cookie video, the title sounds boring and you need to watch it for a long time to get to any comparison images. A lot of people will just tune out. It didn’t have the impact it may have deserved.
Meanwhile, if you title your video “Learn this one weird trick Blender developers don’t want you to know!”, you’ll get all the attention from the clickbait, plus all the (negative) attention from the offended commentators for free! If Andrew instead catered to all those uber-professionals who knew it all along, he wouldn’t be successful. He’s doing exactly the right thing for his audience.
That is exactly the point, Andrew did not say he was the one figure this out,but he have his way of get people interested in a technical boring stuff. And guys please let this thread be about the technical stuff and don’t let it be about fan of Andrew or not a fan. I knew that if I mention Andrew Price in the title, people would be interested.Yeah okay I get it, I might have been a little too enthusiastic or dramatic perhaps. I am not into technical stuff, as I said before, I“see what I see” call it what you want.
I read in YouTube that somebody mentioned that from his observation the filmic add-on works similar to the same compared to useing the build in log transformation. The main difference with filmic offering a look option for the contrast.
I have to do some more tests when I have time - I have to admit that so far I did not pay much attention to the tone mapping tools as the look CRF tools worked quite fine in my scenes - but I often do not have bright sunlights or such.
I have to say that the filmic option for me as a creative person is a lot easier to approach than the current build in tools. I really however prefer to fully understand what the options do. So Andrews summarize of data was interesting to follow.
maybe the users manual for blender should include this data as well - often I find the explanations there are too cryptic.
He clearly stated, that it’s Troys work and even that Troy was very helpful in preparing the Tutorial. I’d call that giving proper credits.
I guess, most of the people ranting hadnt even watched it. Don’t know whats the problem with Andrew. Hes making a living with Blender, so what? Dont like his style, feel to advanced for noob tutorials? Dont watch him.
I wonder when BA turned Reddit. No wonder i’m under 10 posts, its no fun anymore here and i miss the Elysiun days
I liked the video - and Andrew Price is targeting beginners here. How many beginners get fed up with Blender because try as they might, they just cannot get decent results.
Andrew’s style may grate a bit and some probably don’t like his ‘clickbait’ style - but his tutorials are good and he has invested a lot of his time into his work. Following his tutorials - you can get nice looking results quickly and simply.
Well to be fair this is a two sided story in how the reputation grew.
However I also do not see the point in calling what he presents wrong - considering that he just summarizes the material he gathered
from others including Troy.
I have the feeling like I stated above that there is a difference in philosophy in how you approach color correction and why people think
that this is wrong. Oddly after all Andrew just talks about filmic he tho is not the creator yet he gets criticized for it.
I guess people are just annoyed that Andrew delivered an old concept in a whiz-bang style. That said, I am pretty sure Troy couldn’t give two flying shits about being credited as long as the information is correct.
The Filmic_log and the contrast looks should be a bit better than a base log encoding because the base log encoding is using a formula, while the filmic log is using a 3D LUT. The difference is that the 3D LUT can have tiny little adjustments that might be a little too difficult to approach using a formula. And, in particular, one of the main thing that this ocio config is trying to solve is trying to desaturate highlights nicely, so it might be that it’s a bit more polished in that section than a formula could be.
Well Andrews statement about the CRF look options made my eye brows go up.
Everybody familiar with CRF should now also as the Blender manual states that those are for artistic looks and not a serious color management tool.
If tho the rest is accurate and I think it is then the filmic options seems to provide a much better approach to tone mapping.
The CG Cookie movie summarizes that pretty good and in less time.
I was never a big fan of the compositor when having to do the CC after rendering - I prefer this during rendering.
What Filmic does seems to be similar to tone mapping in Indigo/LuxRender.
Something that is very true in the CG Cookie movie is that just turning on Filmic will not yield much - you have to rethink your lighting approach. Then you can really see the difference - tried that and this is pretty true.
I also like the fact that it adjust colors of lights and over exposed areas - while this in addition also simulates more film looks and reaches
into artistic territories similar to the standard CRF look approach.
Maybe both the old and new Color management could be an option to choose from.
Curious if they will offer both in upcoming Blender version while I have to admit I mainly use 4 CRFs only.
I’ve been trying the new filmic color on a old scene and everything looks ok except one object i’m using with glass material, it has a lot of fireflies. On the old scene changing the indirect clamp to 0.03 fixed this but now that setting makes everything way to dark, it’s like the setting can’t go low enough as if you put in a number too low it changes to 0… Am I missing something obvious?
Basically you’re telling cycles to consider the max intensity of a sample that have bounced already one or multiple times (indirect) as 0.03. (Higher values are clamped to 0.03).
Pratically with a so low value (actually I consider it a really strange value), the contribution given by indirect light to pixel is almost irrelevant so you clamp fireflies energy.
As a general rule in poor words with filmic you shouldn’t clamp to keep everything consistent and smooth from highlights to shadows.
If it gives to you more stops of light, so you have the possibility to crank up the light to huge values without burning and then you clamp the highlights energy, what’s the point of use filmic?
It’s obvious that in an interior scene, where everything comes mostly by indirect light, you’ll end up with a dark situation when you clamp indirect to 0.03 and you have light sources with spitting samples with energy of 10000 for example.
Same for those who clamp for noise, it’s obvious that if you clamp energy samples you get less difference between close pixels’ intensities, but the scene will be dark (and light won’t behave linear) cause pixels brighter were castrated.
At least this is what I think. Right or wrong take it as you wish. I know there are clamping-fans here but it’s not that I have to sell them my believes.
As for fireflies from glass try disabling caustics as a first test if you don’t need them.
ah thanks for the explanation that makes sense. I already tried disabling caustics but I’ll spend some more time on it tomorrow, I just need to test it out a bit more I think.
Why Andrew would get such grief is beyond me, He didnt say he’s Jesus, He just did a little video with the little knowledge he had that im pretty sure would of helped quite a few people either hear about the Colour display issue in Blender or find the plugin available to get more filmic results.
No wonder the guy has stated he NEVER uses this forum, I wonder why. Even after watching this it did make me question if the Node blend types within the normal cycles material editor (not just the compositor) create wrong results with this plugin. Any idea’s anyone, I promise not to acuse you of taking credit for others work, But if the mix node types in material editor are wrong with this plugin its not something that can be said to be ready to go into Master yet anyway.
I dont quite get the loath towards Andrew, now I learnt there’s such a thing for Blender which could make render better, who cares what way he sells things, I actually enjoyed his free tutorials, they are among the best blender tutorials, practical and clear, this tutorial certainly doesn’t bother me either, i think its more like the title is a bit provocative for some people who dont get along with his marketing.
