Re:Why xHE-AAC is being embraced at Meta
by mgdlbp on Wednesday April 12, 2023 @05:40AM
(#35537110)
Yeah, opus is incredible for speech, but at those bitrates it gives music a kind of sibilant buzzing that's really hard to listen to, whereas (I find) the muffled/underwater effect of mp3 and aac isn't much worse than just lowpassing.
What I found was that explicitly lowpassing beforehand to match the other encoders (around 4-8kHz) got rid of the buzzing. Apparently opus' threshold for automatically doing this is a lot lower than the other encoders; it keeps higher frequencies at the cost of more artifacts. I concluded that decimating or setting the bandwidth to 8kHz/"wideband" or lower was an improvement, and similar to the others, though I still slightly preferred Fraunhofer at 8-24kbps.
By default, 12k and 24k are particularly bad; at 16k the quality goes up from downmixing to mono, and by 8k decimation is automatically applied.
What I found was that explicitly lowpassing beforehand to match the other encoders (around 4-8kHz) got rid of the buzzing. Apparently opus' threshold for automatically doing this is a lot lower than the other encoders; it keeps higher frequencies at the cost of more artifacts. I concluded that decimating or setting the bandwidth to 8kHz/"wideband" or lower was an improvement, and similar to the others, though I still slightly preferred Fraunhofer at 8-24kbps.
By default, 12k and 24k are particularly bad; at 16k the quality goes up from downmixing to mono, and by 8k decimation is automatically applied.