Create video using ffmpeg

Question

I have 100 images(PNG) and I want to create a video using these images. I am using the ffmpeg library for this. Using command line I can create video easily. But how do I do it

Answer 1

The reason this comes up again and again is because you're using encoding_example.c as your reference. Please don't do that. The most fundamental mistake in this example is that it doesn't teach you the difference between codecs and containers. In fact, it ignored containers altogether.

What is a codec? A codec is a method of compressing a media type. H264, for example, will compress raw video. Imagine a 1080p video frame, which is typically in YUV format with 4:2:0 chroma subsampling. Raw, this is 1080*1920*3/2 bytes per frame, i.e. ~3MB/f. For 60fps, this is 180MB/sec, or 1.44 gigabit/sec (gbps). That's a lot of data, so we compress it. At that resolution, you can get pretty quality at a few megabit/sec (mbps) for modern codecs, like H264, HEVC or VP9. For audio, codecs like AAC or Opus are popular.

What is a container? A container takes video or audio (or subtitle) packets (compressed or uncompressed) and interleaves them for combined storage in a single output file. So rather than getting one file for video and one for audio, you get one file that interleaves packets for both. This allows effective seeking and indexing, it typically also allows adding metadata storage ("author", "title") and so on. Examples of popular containers are MOV, MP4 (which is really just mov), AVI, Ogg, Matroska or WebM (which is really just matroska).

(You can store video-only data in a file if you want. For H264, this is called "annexb" raw H264. This is actually what you were doing above. So why didn't it work? Well, you're ignoring "header" packets like the SPS and PPS. These are in avctx->extradata and need to be written before the first video packet. Using a container would take care of that for you, but you didn't, so it didn't work.)

How do you use a container in FFmpeg? See e.g. this post, particularly the sections calling functions like avformat_write_*() (basically anything that sounds like output). I'm happy to answer more specific questions, but I think the above post should clear out most confusion for you.