First, thanks for a great product.

I've got a timelapse I shot on the Canon R5. When I select keyframes and do holy grail, the estimated exposure value on the frames on part of the timelapse have a very different exposure than those earlier in the sequence, even though they were shot with the exact same settings. This causes later frames to be blown out or under exposed. I attached a link to the images from the relevant part of the timelapse, any ideas what is going on. I'm on the latest Lightroom, LRTimelapse, and DNGConverter. Any help would be greatly appreciated.

https://www.dropbox.com/s/1qq3kl3ee0nodr...t.zip?dl=0

Thanks!

Here's the image overlay, see how it drops after the car lights in the middle: https://ibb.co/7RrKBgz Thanks!

As you can see the "Holy Grail" compensation curve at the beginning is way above the horizontal middle, which means it makes the images brighter, after the strange part in the middle where you changed camera adjustments (something is wrong there, see below), the orange curve is below the horizontal middle, which means, images will get developed darker there.

The goal with the holy grail compensation is to compensate for the camera adjustments during sunset and sunrise and in result get an even curve.
In your case you have strong headlamps of a car coming in and out, that is nothing the Holy Grail Wizard was meant to compensate for in the first place.

The short sequence that you sent me, I could get the Curve look quite decent by shifting the Stretch slider all the way to the right and leaving the rotate at the middle. But that's a coincidence, it might not work with the full sequence.

Maybe for that sequence leave the holy grail wizard out (you can delete the holy grail compensations alltogether via the button in the toolbar above the table). Then do the editing and use the visual deflicker only. Maybe that can save the sequence.

Personally I think car's putting their lights into a milky way sequence ruin that sequence. If you take out that part, of course you will have a gap. So maybe use the part before and the part after in separate timelapse.

Thanks for the info. It's fine if the headlights frames end up blown out, but you can't really cut them out without getting a jump in time. I'm not understanding why the same camera settings end up with different exposure estimates? The luminance seems to be the same. It's not possible to save it with the full sequence. (I can send you the full sequence if you want) Also, after this part, it does a holy grail to day time, so I need the holy grail ramping. Any other ideas?

Also, I'm seen this on most of the holy grail's I shoot lately (R5 and A7Siii), I can send a few more examples over if you want. I'm not saying this is a bug, but it sure feels like a bug. Maybe you can explain a bit more why the orange line would end up so much lower after the headlights, even though those are shot at the exact same settings as before the headlights. (Sorry, I'm not understanding) Thanks!

Also, just fyi, I've edited sequences like this without a problem in the past. (with headlights and a night to day ramp). Thanks!

I tried removing the smoothing, but it didn't seem to help. I think the root problem here is that even when removing holy grail smoothing, I end up with frames where the "LRT internal use (HG)" layer is at -4 or close, or (+1.8), and that causes clipping internally.

As I said, the corrections are calculated by the orange Holy Grail curve. You need to adjust that to be as close to the horizontal middle line as possible. Anywhere the curve is above that line it will push the exposure up and where it's below it will push it down. That's how the HG-Wizard works and that's always been like that, no changes since LRT 4 or so.