.jpg + .mp3 one length as .avs another after ffmpeg produces file

[Starduster]

I have been putting audio behind .jpg's for years. I just started using .mp3 for the audio and am getting results I don't understand. I produce the .avs file and AVIsynth 2.6 with:

vd = DirectShowSource("Slide7.mp3", fps = 15, convertfps = True)
vd = vd.ssrc(22050)
#fv=ImageSource("slide7.jpg",fps=15,start=1,end=ceil(15*AudioLengthF(vd)/AudioRate(vd))).ConvertToRGB32().Lanczos4Resize(800,450)
fv=ImageSource("slide7.jpg",fps=15,start=1,end=ceil(15*16.5)).ConvertToRGB32().Lanczos4Resize(800,450)
sld7 = AudioDub(fv,vd)
sld7


It plays for 16.5 seconds and stops like it should. Then I create the .flv file using the latest version of ffmpeg with:

ffmpeg -async 15 -i Slide7.avs -vcodec flv -r 15 -s 800x450 -b:v 512k -ab 96k -ar 22050 -coder 1 -flags +aic+loop+mv4+naq -mpv_flags +cbp_rd -trellis 1 Slide7.flv


But the .flv 29 seconds! What is going on?... how do I fix?

[raffriff42]

1) DirectShowSource does not need "fps = 15, convertfps = True"
2) ffmpeg "-async 15", " -flags +aic+loop+mv4+naq" and "-mpv_flags +cbp_rd" do not apply
3) why reencode the audio? mux the original audio, something like this:## Slide7.avs
ImageSource("slide7.jpg").Lanczos4Resize(800,450).Loop(-1).AssumeFPS(15)

ffmpeg -i "Slide7.avs" -i "Slide7.mp3" -vcodec flv -b:v 512k -acodec copy -shortest "Slide7.flv"

There is actually no need for AviSynth here, since ffmpeg does slide shows natively... I just don't know how to do it off the top of my head.

[Starduster]

Sorry... out of the office all last week. Actually, I do need AviSynth because slide7 is one of many components that will be combined. Each component is different and I'm making them all the same (size, format, fps, etc.) so I can create a single video from all of them: ...slide5 ++ slide6 ++ slide7 ++ slide8 ...

The thing I don't understand is why AudioDuration(vd) or AudioLengthF(vd)/AudioRate(vd) indicate 29 seconds. The .avs plays 16.5 seconds and mediaInfo shows an audio duration of 16.5 seconds. The audio content is 16.5 seconds.

[raffriff42]

Why force FPS on an audio-only clip? This may be the cause of the duration change, I don't know. Again, why reencode when you can copy the original audio?

If you really need to process audio in AviSynth, try FFmpegSource (http://avisynth.nl/index.php/FFmpegSource) or LSMASHSource (http://avisynth.nl/index.php/LSMASHSource) instead. DirectShowSource (http://avisynth.nl/index.php/DirectShowSource) could be doing anything, depending on your system's codec (mis)configuration.

[Starduster]

In order to create a single video from a varying number of dissimilar audio/video components, they must be of the same frame rate (format, size, channels, etc. too). So when the .jpg and .mp3 are combined, they need to match all the other components. I'll give FFmpegSource and LSMASHSource a go and see if there is any difference there. In the mean time, I've used audioTrim to force the .avs output to the duration I want and that seems to work.