ffmpeg combine multiple images - output video incorrect xfade transition time, and total length

Question

I want to merge 5 images with transition at specific times 4.00, 10.0, 16.0, 22.0, with a transition duration of 1 sec, and a 27 sec audio into a 27 sec video clip.

I used the below command, and got output video of 10 min 43 sec

./ffmpeg 
-i 27.m4a 
-loop 1 -t 4.0 -i image01.jpeg 
-loop 1 -t 6.0 -i image02.jpeg 
-loop 1 -t 6.0 -i image03.jpeg 
-loop 1 -t 6.0 -i image04.jpeg 
-loop 1 -t 5.0 -i image05.jpeg 
-movflags faststart -pix_fmt yuva420p 
-filter_complex " 
[1:v] format=pix_fmts=yuva420p, 
scale=1080:1920:force_original_aspect_ratio=decrease, 
pad=1080:1920:(ow-iw)/2:(oh-ih)/2, 
zoompan=z='if(eq(on,1),1,zoom+0.0004170)':x='iw-iw/zoom':y='0':fps=60:d=60*4.000:s=1080x1920, 
crop=w=1080:h=1920:x='(iw-ow)/2':y='(ih-oh)/2', 
setpts=PTS-STARTPTS+0.000/TB[v0]; 
[2:v] format=pix_fmts=yuva420p, 
scale=1080:1920:force_original_aspect_ratio=decrease, 
pad=1080:1920:(ow-iw)/2:(oh-ih)/2, 
zoompan=z='if(eq(on,1),1,zoom+0.0004170)':x='0':y='ih-ih/zoom':fps=60:d=60*6.000:s=1080x1920, 
crop=w=1080:h=1920:x='(iw-ow)/2':y='(ih-oh)/2', 
setpts=PTS-STARTPTS+4.000/TB[v1]; 
[3:v] format=pix_fmts=yuva420p, 
scale=1080:1920:force_original_aspect_ratio=decrease, 
pad=1080:1920:(ow-iw)/2:(oh-ih)/2, 
zoompan=z='if(eq(on,1),1,zoom+0.0004170)':x='0':y='ih-ih/zoom':fps=60:d=60*6.000:s=1080x1920, 
crop=w=1080:h=1920:x='(iw-ow)/2':y='(ih-oh)/2', 
setpts=PTS-STARTPTS+10.000/TB[v2]; 
[4:v] format=pix_fmts=yuva420p, 
scale=1080:1920:force_original_aspect_ratio=decrease, 
pad=1080:1920:(ow-iw)/2:(oh-ih)/2, 
zoompan=z='if(eq(on,1),1,zoom+0.0004170)':x='iw-iw/zoom':y='0':fps=60:d=60*6.000:s=1080x1920, 
crop=w=1080:h=1920:x='(iw-ow)/2':y='(ih-oh)/2', 
setpts=PTS-STARTPTS+16.000/TB[v3]; 
[5:v] format=pix_fmts=yuva420p, 
scale=1080:1920:force_original_aspect_ratio=decrease, 
pad=1080:1920:(ow-iw)/2:(oh-ih)/2, 
zoompan=z='if(eq(on,1),1,zoom+0.0004170)':x='iw-iw/zoom':y='ih-ih/zoom':fps=60:d=60*5.000:s=1080x1920, 
crop=w=1080:h=1920:x='(iw-ow)/2':y='(ih-oh)/2', 
setpts=PTS-STARTPTS+22.000/TB[v4]; 
[v0][v1]xfade=transition=circlecrop:duration=1.00:offset=3.000[v01];
[v01][v2]xfade=transition=hrslice:duration=1.00:offset=8.000[v12];
[v12][v3]xfade=transition=slideleft:duration=1.00:offset=13.000[v23];
[v23][v4]xfade=transition=vdslice:duration=1.00:offset=18.000[v34]" 
-map "0:a" -map [v34] -y output.mp4

====

I think I need to tweak the time values are specified at 4 places to get this right

-loop 1 -t TIMESTAMP1

I set this to expected duration of each image i.e, 4, 6, 6, 6, 5

zoompan=z='if(eq(on,1),1,zoom+0.0004170)':x='iw-iw/zoom':y='0':fps=60:d=60* TIMESTAMP2 :s=1080x1920,

I set this to expected duration of each image i.e, 4, 6, 6, 6, 5

setpts = PTS - STARTPTS+ TIMESTAMP3 /TB

I set this to cumulative offset from 0 i.e, 0, 4, 10, 16, 22

Additionally, I ensured that TIMESTAMP2 + TIMESTAMP3 = time at which each image display should end ie., 4, 10, 16, 22, 27

xfade=transition=circlecrop:duration=1.00:offset=TIMESTAMP4

=======

Few problems

1. The transition between images, is not as per the times i expect
2. Overall clip length is long 10:43 as opposed to expected 00:27
3. Overall processing time is 5+ minutes

Am i missing something ?

question from:https://stackoverflow.com/questions/65867640/ffmpeg-combine-multiple-images-output-video-incorrect-xfade-transition-time-a

Answer 1

Waitting for answers