AI Video Editing Flows

FFmpeg speed up or slow down video, and change the frame rate

Change video speed with setpts and audio speed with atempo, and change both so picture and sound stay the same length. On our 10 s, 300 frame test file, a 2x speed up gave 5.066 s and 152 frames, and a 2x slow down gave 19.988 s. The slow down kept only 300 frames until we added fps=30.

What to choose

Use setpts for the picture and atempo for the sound. For frame rate, use the fps filter to change the number of frames, and keep -r for the output side only. minterpolate invents in between frames and is slow.

JobResult on our 10 s, 30 fps file
setpts=0.5*PTS with atempo=2.05.066 s, 152 frames
setpts=2.0*PTS with atempo=0.519.988 s, 300 frames
Slow down plus fps=3020.000 s, 600 frames

Speed up and slow down

setpts multiplies each frame time. Smaller is faster. atempo changes the audio speed without changing pitch.

Twice as fast: 5.066 s, 152 frames
ffmpeg -i a.mp4 -filter_complex "[0:v]setpts=0.5*PTS[v];[0:a]atempo=2.0[a]" -map "[v]" -map "[a]" fast2.mp4
Half speed: 19.988 s, but 300 frames
ffmpeg -i a.mp4 -filter_complex "[0:v]setpts=2.0*PTS[v];[0:a]atempo=0.5[a]" -map "[v]" -map "[a]" slow2.mp4

The half speed file has the same number of frames as the source, stretched over a longer time, so ffprobe reported an average of 9000/599 frames per second. Add the fps filter to get a steady 30 fps.

Half speed at a steady frame rate: 20.000 s, 600 frames
ffmpeg -i a.mp4 -filter_complex "[0:v]setpts=2.0*PTS,fps=30[v];[0:a]atempo=0.5[a]" -map "[v]" -map "[a]" slow2_cfr.mp4

The atempo limit

One atempo filter has a limited range. Asking for a quarter speed in one filter failed with "[fc#-1 @ ...] Error applying option 'tempo' to filter 'atempo': Result too large".

This fails: atempo=0.25 in one step
ffmpeg -i a.mp4 -vn -af atempo=0.25 slow4_direct.m4a

Chain two filters instead. Our 10 s of audio became 39.955 s.

Quarter speed audio as two half steps
ffmpeg -i a.mp4 -vn -af atempo=0.5,atempo=0.5 slow4_chain.m4a

Frame rate: -r or the fps filter

-r after the input sets the output frame rate. The fps filter does a similar job inside the filter graph. -r before the input changes how FFmpeg reads the file, so the timing changes. We ran all three on the 30 fps file.

CommandVideo framesFile lengthVideo stream
ffmpeg -i a.mp4 -r 15 r15.mp415210.133 s10.133 s at 15/1 fps
ffmpeg -i a.mp4 -vf fps=15 fps15.mp415010.005 s10.000 s at 15/1 fps
ffmpeg -r 15 -i a.mp4 r_input.mp430020.000 s20.000 s at 15/1 fps

With -r before -i the 300 frames were read as 15 fps, so the picture ran 20.0 s while the audio stayed at 10.005 s. That breaks sync. Use -r after the input.

Motion interpolation

To raise the frame rate with new in between frames, use minterpolate. fps=60 is the plain alternative. We ran both on the first 3 s.

minterpolate to 60 fps: 177 frames in 24.4 s
ffmpeg -t 3 -i a.mp4 -vf "minterpolate=fps=60:mi_mode=mci" -an interp60.mp4
fps=60: 180 frames in 1.5 s
ffmpeg -t 3 -i a.mp4 -vf fps=60 -an fps60.mp4

minterpolate gave 177 frames in 24.4 s, 3 fewer than the 180 from fps=60, and took far longer. We did not compare the look of the two.

Method and limits

All commands ran on FFmpeg 8.1.2 (a full Windows build) on a Windows 11 laptop, Ryzen 7 7735U, CPU only, on 2026-11-06. We added -y and -threads 4 to every run; they are left out of the commands shown. In quoted FFmpeg messages the hex address after @ changes on every run, so we replaced it with ...

a.mp4 is a synthetic 10 s test file (testsrc2 picture and a sine tone, 1280x720, 30 fps, a keyframe every 2 seconds), made with the command below.

Make a.mp4
ffmpeg -f lavfi -i testsrc2=duration=10:size=1280x720:rate=30 -f lavfi -i sine=frequency=440:duration=10:sample_rate=48000 -c:v libx264 -pix_fmt yuv420p -g 60 -keyint_min 60 -sc_threshold 0 -c:a aac -shortest a.mp4

Limits: one run of each command, so the sizes and times are single measurements and not averages. Times were taken on a laptop that was doing other work. Results depend on the input file and the FFmpeg version, so check yours. Test pictures were synthetic, so we judged nothing about motion quality.

Related: FFmpeg for video editors, FFmpeg trim video: accurate cuts by time, Join videos with FFmpeg without errors, FFmpeg compress video: CRF and presets, measured with VMAF and FFmpeg command builder.

Page checked on Nov 4, 2026.