LRTimelapse Forum

Hi. This is my first post here so please be gentle!

My project involves about 1200 frames ranging from very dark (2am) to very bright (11am). It is a sequence of a very busy traffic circle in NYC. For the first ~400 exposures, I didn't use a long enough shutter speed to capture a reasonable amount of activity. What I'd like to do for those 400 frames is to composite them five at a time using PS. The result would be 80 frames that have the desired amount of action and are brighter. (I've tested the idea, I'm happy with how the composites have turned out, and 80 frames is an adequate number for that part of the TL sequence.)

All of the images, composite or not, require some lens correction/distortion/rotation/crop adjustments -- let's just refer to them as "distortion" -- (the same for every frame) as well as various other develop panel adjustments (not the same for every frame -- more of a normal LRTimelapse flow with keyframes).

The output files from PS obviously cannot be .dng files -- they must be psd or tiff or jpg, etc. So the first question is, can I mix, say, jpg's and dng's in LRTimelapse? If not, I'd like opinions as to what to do, e.g., should I export everything into the same format -- jpg, I guess? -- and re-import them, and if so, should I do the distortion adjustments prior to exporting them or will LRTimelapse be able to deal with those adjustments if they are made to the jpg's once I have imported them back into LR4? In fact, will LR4 "see" those distortion adjustments or when I open them in LRTimelapse will they not appear?

For that matter, is there another approach I should consider?

I hope I've done a good enough job explaining what I'm trying to do and I really appreciate any replies.

Hi, this is certainly possible but quite complicated. And I think you will have to deal with a different visual appearance between the stacked images and the non-stacked in the last part of the sequence.

I have two simpler approaches for you:

1. Use the "Select every nth." feature in the HDR menu to select every second image, do this only for the first part (clear the selection for the last part) - for example select every 2nd. image in the first part, delete them. Then select every second image again, delete them. The result would be a first part that plays 4 times as fast as before.
   
2. Do the accelleration in post processing in you video-editing program. So just render the whole sequence, import it in the video program, cut the sequence between the 2 parts and make the first part 4 times faster. In the most video programs you could as well key frame the playback speed so you can make accellerations wherever you want.
   

Hope I could help!

PS: To answer your questions: you could mix a sequence with DNG and JPG but I would recommend exporting all into JPG to have consistency. If you export the RAW to JPG to further process them in LRTimelapse make sure to check "minimize Metadata" or initialize the Metadata of the JPGs in LRTimelapse to get rid of the "old" develop-settings.

(2012-06-29, 10:28)gwegner Wrote: [ -> ]Hi, this is certainly possible but quite complicated. And I think you will have to deal with a different visual appearance between the stacked images and the non-stacked in the last part of the sequence.

I have two simpler approaches for you:

1. Use the "Select every nth." feature in the HDR menu to select every second image, do this only for the first part (clear the selection for the last part) - for example select every 2nd. image in the first part, delete them. Then select every second image again, delete them. The result would be a first part that plays 4 times as fast as before.
   
2. Do the accelleration in post processing in you video-editing program. So just render the whole sequence, import it in the video program, cut the sequence between the 2 parts and make the first part 4 times faster. In the most video programs you could as well key frame the playback speed so you can make accellerations wherever you want.
   

Hope I could help!

PS: To answer your questions: you could mix a sequence with DNG and JPG but I would recommend exporting all into JPG to have consistency. If you export the RAW to JPG to further process them in LRTimelapse make sure to check "minimize Metadata" or initialize the Metadata of the JPGs in LRTimelapse to get rid of the "old" develop-settings.

Gunther, thanks so much for your speedy reply and for your suggestions. The reason I want to compress the early frames is not to accelerate the time-lapse, rather, to get more "life" in each frame. Because the exposures were short and it was the middle of the night, there wasn't much traffic, so I didn't get a lot of light trails in each frame. When I composite them I get frames that have more going on in them. It's possible that accelerating them could have the same effect, so I'll try your suggestion.

To the other question, what I want to know is if I do the distortion/correction/crop adjustments and export the images as JPGs, will those adjustments survive or will I have to do them over when I bring them back into LR/LRTimelapse? I'm not sure what you mean about getting rid of old develop settings, as I want to *keep* the distortion/correction/crop edits. Does that make sense?

Again, thanks very much for your time. Your program is wonderful and your book was very, very helpful!

Quote:To the other question, what I want to know is if I do the distortion/correction/crop adjustments and export the images as JPGs, will those adjustments survive or will I have to do them over when I bring them back into LR/LRTimelapse?

If you make this adjustments before exporting the will be rendered into the JPGs, so yes, the will survive. But if you don't click "minimize metadata" the "instruction" to apply this effekt will stay in the xmp and cause it to be double applied the next time you render out the jpgs again.

Quote:Again, thanks very much for your time. Your program is wonderful and your book was very, very helpful!

That's good to hear, thank you very much!