ffmpeg won't compile all images in directory into video for stopmotionme drop

I'm trying to compile a stop motion video using ffmpeg. I have 8000+ images. I run a command like this:

Code:

---

ffmpeg -i %04d.jpg -r 30 output.mp4

---

Everything is smooth up until about frame 6000 where it cuts off. The other 2000 are left out of the video.

Could there be something wrong with an image at that point that I'm not seeing?

Is this a limitation of ffmpeg?

Thanks!

I've used this command successfully in the past:

Code:

---

ffmpeg -r 25  -qscale 2 -i image%d.jpeg output1.mp4

---

where all files were like image01.jpeg

Check all your image filenames exactly fit the wildcard expression, this is usually the problem

You could also try making it in two halves if there is a file number limitation, then join them together with mkvmerge.

Quote:

---

Originally Posted by jarednielsen

I'm trying to compile a stop motion video using ffmpeg. I have 8000+ images. I run a command like this:

Code:

---

ffmpeg -i %04d.jpg -r 30 output.mp4

---

Everything is smooth up until about frame 6000 where it cuts off. The other 2000 are left out of the video.

Could there be something wrong with an image at that point that I'm not seeing?

Is this a limitation of ffmpeg?

Thanks!

---

Can you show the console output too? It may have some useful information on the issue. Note that ffmpeg by default reads the input files at -r 25, so you may want to move your -r 30 as an input option (before -i input). The output will then inherit the frame rate of the input. Otherwise ffmpeg may drop or duplicate frames when the input(s) and output(s) do not match frame rates.

Quote:

---

Originally Posted by Jose Catre-Vandis

I've used this command successfully in the past:

Code:

---

ffmpeg -r 25  -qscale 2 -i image%d.jpeg output1.mp4

---

---

I also often suggest using -qscale for MPEG-1/2/4 outputs, but be mindful that more recent ffmpeg (I'm not sure about the status of repository ffmpeg/avconv) now uses the encoder libx264 by default for .mp4 output, and -crf should be used instead of -qscale if that is the case (the default value of -crf 23 will be applied if you omit -crf). However, I can understand that syntax changes and such are hard/annoying to keep up with.

Also, you're applying -qscale as in input option (before -i input) although FFmpeg will probably understand that you're intending on using it as an output option (usually placed after -i input.), but option placement is important.

Quote:

---

Originally Posted by Jose Catre-Vandis

Check all your image filenames exactly fit the wildcard expression, this is usually the problem

---

Good suggestion.

Quote:

---

Originally Posted by Jose Catre-Vandis

You could also try making it in two halves if there is a file number limitation, then join them together with mkvmerge.

---

Alternatively you could use cat to pipe non-properly numbered (but still properly sequential) images to ffmpeg:

Code:

---

cat *.jpg | ffmpeg -f image2pipe -vcodec mjpeg -i - output

---