taking out frames from a 'fake timelapse'?

D

dylansmith

Well-known member
I typically do the real timelapse, i.e. shoot 1 image every 5s, and import as an img sequence to Premiere Pro CC. Recently, I tried shooting continuous video and speeding up in post to get similar effects.

However, I feel that it looks very choppy and inorganic using the latter method. As such, is there a quick way to automatically take out extra frames from my video, to make it into a 'real timelapse'? Otherwise, other suggestions to make it look more organic are welcome too!
 
Just spitballin' ideas here, but what about exporting as an image sequence at 1 frame per second and then reimporting that? If that's not time lapsed enough, just repeat with the new sequence until you get it to the speed you like.
 
When you speed it up on the timeline, you are essentially doing exactly the same thing - taking out extra frames to make it a timelapse. You just need to find the right amount speed to get the effect you are looking for. To recreate a timelapse shot 1 frame every 5 seconds, you need to speed the video up 12,500% (for a 25fps project). In Premiere Pro the Clip Speed effect will only let you go to 10,000%, but you can combine this effect with the rate stretch tool to go even faster than that and get the speed you're looking for.

The choppiness is likely due to the faster shutter speed required to shoot continuous video. With a photo-timelapse, you have the option to use a slow shutter-speed to add motion blur to the frame for smooth movements. This is not something you can recreate or effectively fake in post. Many video cameras will still let you shoot video at slow shutter speeds so it is still something you can do without having to use stills camera & intervalometer, but the video will look horrible until you speed it up.
 
Last edited:
You probably dont want to drop frames.. because you are then going to get a set of distinct sharpish (1/50) photos.

You need to merge frames together not drop them.

I cant go through the options in Prem because I dont use it.

Many softwares have options..
nearest frame
optical flow

Usually the latter will create the desired effect.

S
 
deltoidjohn said:
When you speed it up on the timeline, you are essentially doing exactly the same thing - taking out extra frames to make it a timelapse. You just need to find the right amount speed to get the effect you are looking for. To recreate a timelapse shot 1 frame every 5 seconds, you need to speed the video up 12,500% (for a 25fps project). In Premiere Pro the Clip Speed effect will only let you go to 10,000%, but you can combine this effect with the rate stretch tool to go even faster than that and get the speed you're looking for.

The choppiness is likely due to the faster shutter speed required to shoot continuous video. With a photo-timelapse, you have the option to use a slow shutter-speed to add motion blur to the frame for smooth movements. This is not something you can recreate or effectively fake in post. Many video cameras will still let you shoot video at slow shutter speeds so it is still something you can do without having to use stills camera & intervalometer, but the video will look horrible until you speed it up.
Click to expand...

Hi there, my shutter speed is maintained at 1/50s all the way. speeding up the video still doesn't look the same as a standard timelapse merged using stills, but can't point out why.
 
It shouldn't look choppy. If anything it should look smoother than a regular timelapse. What are your settings? Have you rendered the shot?
 
dylansmith said:
Hi there, my shutter speed is maintained at 1/50s all the way. speeding up the video still doesn't look the same as a standard timelapse merged using stills, but can't point out why.
Click to expand...

ther is no such thing as a standard TL using stills..

Using your example the still camera could have the shutter at any thing from 1/8000 to 1/2.5 or less each producing a different look
 
Were you also shooting the photo time-lapses at 1/50th? If not, there is no way make them look the same in post.

Time-lapses shot using still photographs will always look better than video (given the same conditions). They'll have more resolution, less/no compression, better DR (especially with RAW), finer grain & better noise reduction, etc. However, given the same conditions (ie shooting stills at 1/50th shutter speed) the playback motion is not something that will be a differentiator. No amount of dropping/blending/merging frames in post will give a true replication of a photo time-lapse shot with a shutter speed of, as an example, 2 seconds.

If you were shooting the still photographs at a shutter speed of 1/50th, then the differences you are seeing are either down to image quality - resolution/noise/colour/etc - or operator error during the post-processing.
 
The problem here is most likely you're using low shutter speed in photographic timelapse and it's giving you a different look from the shutter speed used in video being a minimum of 1/30s (and you probably shot 1/50s of more)
 
Like deltoidjohn says. In this case nothing to do with shutter speed and you obviously do want to drop frames. Make sure that frame blending is off and that frames are removed at even intervals i.e, use percentages so that your frames are dropped at identical intervals rather than duration to set your speed and there shouldn't be any difference with regular timelapse (barring, say, your camera running at an uneven frame rate).
 
Last edited:
I did a pseudo timelapse-with-video. Yea, video will stutter when it's sped up dramatically. But in my silly case, I did this because I wanted a time lapse AND real time video from the same shot, so I compromised. Not sure if you're able to post an example but this is how it turned out for me:

[video]https://dl.dropboxusercontent.com/u/4917215/BMPCC/Spider v2.mp4[/video]

Ignoring the real time video sections, I don't think it's possible to produce a "smooth" time lapse, unless I'm not understanding what you are asking in the op. Do you have any examples you can post comparing the two you've tried and how they differ?
 
You must log in or register to reply here.
Back
Top