Re: MPEG DASH
Brad Isbell <[email protected]>
| Newsgroups | gmane.comp.audio.icecast.devel |
|---|---|
| Message-ID | <CAELR-fAJZid84mdYvxkQap3vgabmpDbC-9VHaZdRkoqeQ46rbA@mail.gmail.com> |
Daniel, I won't make any points regarding DASH's fit with Icecast, but I wanted to address your statement that there is no purpose for adaptive streaming. Cell phone connections are fairly terrible and inconsistent. Even if you have a solid signal on a 4G technology, you could walk around a corner and it will disappear. What works at 3mbit/s one minute might be 10kbit/s the next. Adaptive streaming is most certainly helpful in these situations, and is one of the reasons I do not exclusively use Icecast for the stations I work on. Brad Isbell [email protected] http://www.musatcha.com On Tue, Feb 11, 2014 at 6:30 AM, Daniel James <[email protected] > wrote: > Hi Thomas, > > I wrote solid IPR evaluation for a reason. > > It might be helpful to ask bitmovin as a starting point, they have > released the reference code under LGPL: > > http://www.bitmovin.net/libdash.html > > The LGPL preamble says: > > "we insist that any patent license obtained for a version of the library > must be consistent with the full freedom of use specified in this license." > > Personally, I don't see the point of adaptive streaming and the implied > variation in quality. Either the quality you are streaming at is > acceptable, or you are wasting bandwidth by streaming with too high a > bitrate, perhaps due to an inefficient codec. > > I suppose the reasoning is that drops in quality are preferable to > playback interruptions, but why tolerate either in a well-designed > system? Adaptive streaming seems to say "we know the Internet is not > really cable TV, but we are going to work around that by providing a > cruddy picture for those times when available bandwidth does not align > with our business model" :-) > > In the Icecast community, we might gain more by promoting Opus as a more > efficient replacement for MP3 streaming, for example. > > Cheers! > > Daniel > _______________________________________________ > Icecast-dev mailing list > [email protected] > http://lists.xiph.org/mailman/listinfo/icecast-dev > _______________________________________________ Icecast-dev mailing list [email protected] http://lists.xiph.org/mailman/listinfo/icecast-dev