Welcome to our community

Be a part of something great, join today!

Log in
  • Hey all, just changed over the backend after 15 years I figured time to give it a bit of an update, its probably gonna be a bit weird for most of you and i am sure there is a few bugs to work out but it should kinda work the same as before... hopefully :)

Need Help : What is best Gamma for LUSTER CC

Biba said:
Good day , I want to make CC on Luster . In what Gamma is better to render out DPX files. RED LOG or PD LOGXX .
Any experiences with Luster ?
Thanks.
Click to expand...

Autodesk's recommendation, if you're going to be correcting for film output, is to use PDLog985 and lower the exposure setting by two stops, as well as increase the saturation by a small amount. This produces a curve that is very close to that of a properly scanned negative. My own experience confirms this, the result is a very good starting point for film color correction.
 
That's an interesting comment Mike. To date I haven't been able to figure out what the pdlog985 curve was built for, so I've pretty much ignored it.... will have to have a play with it again and note the "drop 2 stops" comment! thanx Chris
 
weirdcrew said:
You can use the same settings for straight into telecine as well I have all along, in Lustre. My filmouts was into DI with Lustre and then filmout.

But Mike has greater knowledge then any of us and maybe can give us some more tricks in the trade?
Click to expand...

Thanks for the undeserved flattery.

Most of the film work we've done has been with Scratch as the conform and color correction system, so the need for DPX conversion prior to color correction has been moot. However, in order to supply VFX vendors and others with usable images, we have had to do such conversions from time to time. The Autodesk approach is close to what we wound up with because it provides the best simulation of a scanned film negative, which is what most VFX vendors expect to be working with. While Redlog is quite good in other ways, it is a milder log curve, and requires either a custom LUT or more extensive color correction to arrive at a proper starting point, a good reflection of the photographic intent. Using the PDLog985 LUT with a 2 stop pull gets you an image that when viewed through a "standard" Cineon LUT (typically .6 negative gamma, 95 BP, 685 WP, 1.7 display gamma) yields a very good representation of the original photography. Much less guesswork that way. The difference in the amount of information retained from the original RAW image between this approach and Redlog is very, very minimal, if at all. It's just that the PDLog985 approach yields something much closer to existing, expected standards.

That's my take on it.
 
mmost said:
Thanks for the undeserved flattery.

Most of the film work we've done has been with Scratch as the conform and color correction system, so the need for DPX conversion prior to color correction has been moot. However, in order to supply VFX vendors and others with usable images, we have had to do such conversions from time to time. The Autodesk approach is close to what we wound up with because it provides the best simulation of a scanned film negative, which is what most VFX vendors expect to be working with. While Redlog is quite good in other ways, it is a milder log curve, and requires either a custom LUT or more extensive color correction to arrive at a proper starting point, a good reflection of the photographic intent. Using the PDLog985 LUT with a 2 stop pull gets you an image that when viewed through a "standard" Cineon LUT (typically .6 negative gamma, 95 BP, 685 WP, 1.7 display gamma) yields a very good representation of the original photography. Much less guesswork that way. The difference in the amount of information retained from the original RAW image between this approach and Redlog is very, very minimal, if at all. It's just that the PDLog985 approach yields something much closer to existing, expected standards.

That's my take on it.
Click to expand...

Mike,

This is fantastic empirical data. Would you mind cross-posting this to the SCRATCH list? Or, would you mind if I did and (obviously) credited you as the original author?

Lucas
 
Lucas, check out the Autodesk white paper I linked to above for plots illustrating what Mike is talking about.

Although I guess Autodesk aren't your favourite people at the moment…
 
luki said:
Mike,

This is fantastic empirical data. Would you mind cross-posting this to the SCRATCH list? Or, would you mind if I did and (obviously) credited you as the original author?
Click to expand...

Feel free. I would, however, point out that the Autodesk document explains this very well, with graphics to match. I would also suggest trying it yourself (that comment is addressed to everyone). The only drawback of the approach is that the PDLog985 LUT seems to have been removed from RedAlert (and Red Rushes). It is, however, still available via Redline, which makes it available from Crimson and other programs.
 
mmost said:
The only drawback of the approach is that the PDLog985 LUT seems to have been removed from RedAlert (and Red Rushes). It is, however, still available via Redline, which makes it available from Crimson and other programs.
Click to expand...

Yes. Odd that, given they all share the same render engine.
 
mmost said:
Feel free. I would, however, point out that the Autodesk document explains this very well, with graphics to match. I would also suggest trying it yourself (that comment is addressed to everyone). The only drawback of the approach is that the PDLog985 LUT seems to have been removed from RedAlert (and Red Rushes). It is, however, still available via Redline, which makes it available from Crimson and other programs.
Click to expand...

Mike,

Thanks to share your knowledge with us. What I understand is gamma curve like Redlog and PDlog985 is the gamma conversion of the 12bit linear RGB raw format into a film gamma curve. Color space conversion like cameraRGB or rec709 is RGB representation of same RAW format into 3x3 matrix. So one applied to gamma, the other to color space.

Correct me if I'm wrong (for the rest of this paragraph), when we look at the curve in the autodesk Red white paper figure 3, we can see that in the toe of the curve PD log985 is at 95. Past -6 stop, RedLog tends to have less contrast in the middle of the curve. While on the shoulder (highlight) the two curve seems to be the same since +2.8 stop. The curve exposed in fig 4 PDLog985 exposure set at -2, highlight are placed just above 800. Contrast doesn't mean that we have less information. In fact in the three case, the amount of information remain almost the same. PDlog985 exposure set at -2 seems to react with a lattitude of 12 stops but placed black, mid and high at different values. What does this mean in real world? PDlog985 exposure at -2 seems to be a miror of a typical cineon file in term of contrast. But film remains more details in highlight while Red clamped over +4 stop. And red remain more details in black value.

Do we have less possiblity in during the grading process? Does it affect the quality? To me it doesn't seems to be a wrong method. But I will love to get your input Mike. Those concept are complex too. Please tell me if I'm out the track.
 
Hi

Also Cineon file and DPX are calibrated format. This mean that when you get a DPX file, you apply a the right 3DLUT of the print stock then normaly "WYSIWYG". That's exactly the intend of autodesk white paper. Get the most closest curve from DPX (cineon) file to then apply the calibrate LUT and start grading in a well known color pipeline. Because a color pipeline is really hard to keep straith foward. The lab make a lot of effort in quality control to get the chemical process as closest as possible from the original stock's manufacturer. That's also why you will need usually a one light from timing (chemical color grading process) to get a perfect result at each specific time of a chemical process. I've done this many time when I use to work at Technicolor Montreal. Control chimics is crucial, as well as a controlled grading room (lighting condition), calibrate projector and the right LUT of print stock.

Mathieu
 
You must log in or register to reply here.
Back
Top