At 60fps, you have less than 17ms to encode each frame. Neither of these approaches are suitable for things like video conferencing, where there is a small number of receivers for each encoded stream and latency is critical. You can spend an hour of compute time optimally encoding a 20-minute YouTube video, with no real downside. This is appropriate for any sort of broadcast (YouTube, television, etc).Īlso, in many applications, it’s suitable to exchange time for memory / compute. Yes - many codecs can be optimized for decoding at the expense of encoding. There's also codecs/configurations the Snapdragon in the phone could play using hardware acceleration that would choke a low powered Celeron or Atom decoding in software. There's codecs/configurations a Threadripper with 64GB of RAM in a mains powered jet engine sounding desktop could handle in software that would kill a Snapdragon with 6GB of RAM in a phone. It also depends on what you mean by "modern hardware". There's plenty of video codecs or settings for them that can choke modern hardware. Even within a spec features are bound to profiles so a file/stream can be handled by less capable decoders/hardware. A lot of those features existed on paper 20-30 years ago but weren't practical on hardware of the time, even custom ASICs. Later MPEG codecs increased their algorithmic complexity to squeeze better quality video into low bit rates. It was well into the Pentium/PowerPC era until desktop systems could play back VCD quality MPEG-1 video in software. The limit on bitrate led to increased algorithmic complexity. One of MPEG-1's design goals was to get VHS-quality video at a bitrate that could stream over T1/E1 lines or 1x CD-ROMs. That was way higher than the hardware of the time supported. While today most people think of MPEG-1 videos as low quality the spec provides the ability to handle bit rates up to 100Mb/s and resolutions up to 4095x4095. Take the MPEG-1 codec for instance, it was new about 30 years ago. Video compression is a calculus of IO capacity, memory, and algorithmic complexity.
0 Comments
Leave a Reply. |
Details
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |