Hello. First of all, thanx for that great codec. I would encode all my music with Ogg Vorbis in the future, if it hadn't been for some samples I encoded for test purposes: Everything was great as long a I used "natural" sounds/music, but certain pieces of synthesized music (specially techno) contained very audible artifacts even at high bitrates. So I just went on using mp3 for synthesized music, until I recently read your FAQ (the point about reporting artifacts). Afterwards, I encoded my samples again with the latest release version (vorbis-tools 1.0.1 for Windows) and checked the output from both the winamp plugin and oggdec. The artifacts are still there. You can download the encoded and the original samples from people.freenet.de/chaossoft/vorbis/ to check my results. Simeon Maxein <p><p>--- >8 ---- List archives: http://www.xiph.org/archives/ Ogg project homepage: http://www.xiph.org/ogg/ To unsubscribe from this list, send a message to 'vorbis-dev-request@xiph.org' containing only the word 'unsubscribe' in the body. No subject is needed. Unsubscribe messages sent to the list will be ignored/filtered.
> sounds/music, but certain pieces of synthesized music (specially techno) > contained very audible artifacts even at high bitrates.I listened to your samples, and couldn't ABX them even with my $150 headphones on. By subtracting the WAVs from each other, I can hear what you're probably talking about, but I'd say "very audible artifacts" is perhaps a little overstated. Secondly, for someone with golden ears like yourself, I wouldn't consider q5 to be "high bitrate". If you're unhappy with Vorbis' performance on a particular track at a certain bitrate, then keep encoding with higher bitrates until you can't hear the difference. You may be running into the known "treble boost" artifacts at bitrates under q6 or 7. It might be helpful if you could determin the lowest quality setting where the problem (for you) goes away. Say, within .5 or so. You also might be interested in the Garf-tuned version, which gives higher quality at some of the higher bitrates than stock oggenc does. Search the archives for a link. That said, we are always looking for hard-to-encode samples to help us reduce artifacts, and it may turn out that a slight modification can be made to Vorbis to help reduce such artifacts for future versions. I'll let any developers address that. Also, thank you for going to the trouble to put up FLACs containing just the short sample where the artifacts occur, as well as descriptions for what you're hearing. It certainly makes your complaints easier to assess. -- Graham Mitchell - computer science teacher, Leander High School "Several tiers below the surface [of the human heart] is a pervasive, integral force that demands the right to avoid pain and experience self-fulfillment. This self-centered energy is the very essence of what the Bible calls 'sin.'" -- Harry Schaumberg, "False Intimacy" --- >8 ---- List archives: http://www.xiph.org/archives/ Ogg project homepage: http://www.xiph.org/ogg/ To unsubscribe from this list, send a message to 'vorbis-dev-request@xiph.org' containing only the word 'unsubscribe' in the body. No subject is needed. Unsubscribe messages sent to the list will be ignored/filtered.
On Thu, 05 Feb 2004 06:06, Simeon Maxein wrote:> Hello.> So I just went on using mp3 for synthesized music, until I recently read > your FAQ (the point about reporting artifacts). Afterwards, I encoded my > samples again with the latest release version (vorbis-tools 1.0.1 for > Windows) and checked the output from both the winamp plugin and oggdec. The > artifacts are still there. You can download the encoded and the original > samples from people.freenet.de/chaossoft/vorbis/ to check my results.The Product has a definite white noise hiss added in some sections; I wouldn't call it annoying, but it was pretty obvious on my cans. This might be similar in cause to the noise I hear in the crowd applause sections of various live albums I've encoded. I find it goes away at q7, so I encode my live stuff at that, and the rest at 4.99 (for historical/hysterical reasons). You might want to add techno-industrial to that category. Chordian also has something going on; it's subtle as you said. It'd characterize the original's main voice has having a sort of tremelo(?) effect. It sort of warbles between two tones like, maybe, a mandolin, but much faster. In the encoded version this effect seems to be lost as if it were smeared or masked - probably pre-echo. Can't say I know how to fix this, however :-) John --- >8 ---- List archives: http://www.xiph.org/archives/ Ogg project homepage: http://www.xiph.org/ogg/ To unsubscribe from this list, send a message to 'vorbis-dev-request@xiph.org' containing only the word 'unsubscribe' in the body. No subject is needed. Unsubscribe messages sent to the list will be ignored/filtered.