Ogg decode performance

[bogdanbiv]

Hello,

I have been unlucky enough to enter a flame ware by trying to contradict a person saying "mp3s decode faster than ogg, and that is the reason why we don't see too many embedded ogg players around". The person was also invoking that Ogg files have a certain latency when opened in Winamp on Windows on a Pentium III. (this claim was actually quoted from slashdot, so it's even more bogus testing...)

How can I challenge such... sayings? The moment one starts to hear is something very subjective, so how can I measure it more reliable? I suspect the difference in latency is small enough to be hard to notice with human ears.

I know http://jamendo.com provides albums in both OGG and MP3, so I thought can I reliable test decoding them both to say Wav and then compare decode time? What about that latency issue? How should I proceed? Can you provide me a way to measure this cross platform (I'm using Kubuntu 7.10, I assume he uses Windows XP)?

So far I have managed to put him off by showing that Wav 2 Ogg (8sec) is faster than Wav 2 MP3 (22sec) - I used results from the Phoronix test suite.

[deanjo]

Well, the test suite only does encoding benches, so wav ->ogg/mp3 comparisons really don't show anything. I guess the best way to bench is to do a ogg/mp3 ---> wav conversion and compare the times there. Another way to compare the performance I suppose is to see how much overhead both display on the cpu during playbacks of the same song encoded at the same bitrate. I do however find it hard to believe that a P3 would have any latency issues playing ogg as a lowly iPod (using iPod linux of course) can play them fine using Tremor. Also there are a few commercial games out there that use ogg for their game sound effects. If it induced the amount of latency that the person is claiming I really doubt that commercial game developers would use it.