I want to add slides of a pdf presentation to the audio file of the same presentation with ffmpeg (or whatever).
I split the pdf presentations using pdftoppm into numbered png images and tried to use ffmpeg to sync the audio file with the images. But I am wondering if the output file could be smaller. Example, the following command will split file 'foo.pdf' into file 'output-01.png', 'output-02.png' etc:
$ pdftoppm foo.pdf -png output
Then I listen to the audio file 'foo-audio.mp3' and noted the end of each slide in minutes and second. Using a spreadsheet you have to calculate the exact (in seconds) timing of each slide, which can be expressed as 40 for 40 seconds for the program of the conference or 00:01:37 for 1 minute and 37 seconds for your first slide etc and paste it into a text file 'foo-images.txt' which will look like:
file ./program.png
duration 40
file ./output-01.png
duration 00:01:37
Then the video can be built with the command:
$ ffmpeg -safe 0 -f concat -i foo-images.txt -fps_mode vfr -pix_fmt yuv420p output.mp4
height not divisible by 2 (1500x1125)type here
but I got the above error that the height was not divisible by 2. Then I found (on a the ffmpeg mailing list) that I could use the '-vf scale=1680:-2' option and the following command was successful:
$ ffmpeg -safe 0 -f concat -i foo-images.txt -vf scale=1680:-2 -pix_fmt yuv420p output.mp4
and in order to sync the audio with the images and get the final video 'foo-video.mp4':
$ ffmpeg -safe 0 -f concat -i foo-images.txt -i foo-audio.mp3 -vf scale=1680:-2 -pix_fmt yuv420p foo-video.mp4
It worked, but the video is about 60Mb for 30 slides and 30 minutes presentation. Is is a correct ratio? Although the video can be viewed by an Apple viewer. But is there a benefit to use ffmpeg and command lines compared to the use of softwares like imovie or acrobat to get the same result? Also I am confused with the scale 1680 (which is fine) and the 420p of pix_fmt.