Nice, I'm looking forward to seeing how this performs in practice. FFmpeg's previous AAC encoder produced poor quality output and often had irritating chirping artifacts, so I've always had to install Apple's Core Audio encoder on any computer I do video recording on to get decent sound. I've done A/B/X comparisons and found that a 320kbps MP3 sounds better than a 320kbps AAC encoded by FFmpeg, but about the same as a 256kbps AAC encoded by Core Audio. If installing Core Audio is no longer necessary, that'll be a huge improvement and people who use something like OBS to do screen recordings or streaming will get a massive sound quality boost the next time they update.
In the Hydrogenaudio discussion thread's metrics table, the new encoder scores better than Core Audio. But this is at constant bitrate (CBR) [edit: maybe not? see lesscraft's reply below]. Core Audio also has variable bitrate modes (TVBR) which the new encoder lacks.
So maybe Core Audio will continue to be the best when TVBR is available, but I'm hopeful the new FFmpeg encoder will be "good enough", especially if more folks find and contribute problem samples to help tune it.
You can care about quality to the extent that a lossy codec allows. Lossless is not always necessary or wanted. This is like saying “why care about transcoding quality when you can keep the video as is?”. There’s a myriad of use cases and preferences at play here.
Don't get me wrong, this sort of thing is a valuable exercise and we are better off with better encoders for these older codecs. But look at the numbers for Opus on this benchmark. It simply blows all the AAC encoders out of the water even at 64 kbps.
The biggest advantage for having a good AAC encoder isn't efficiency, it's that for nearly the past 2 decades the de facto standard for live streamed video has been RTMP with H.264 video and AAC audio. There is basically no support for any other codecs. If you want to send a video stream to Youtube or Twitch, you will be sending H.264 and AAC. If you want an idea of how ubiquitous this is, I just checked in OBS and it will not even let you select different video and audio codecs in streaming mode, it just (correctly) assumes that anybody who's streaming will be streaming H.264 and AAC.
Plus, at 96+ kbps (assuming an Apple-quality AAC-LC encoder) Opus loses its quality advantage. So at higher bitrates, the benefit of choosing Opus is that encoders/decoders are royalty-free.
That’s going a bit far. I’m in the games industry and have used opus regularly, it’s a great codec for games, often the hardware decoding is so restricted that we’re using software regardless so we might as well use something like opus.
The licensing restriction is unfortunate, but only restrictive for those with very specific goals, under normal conditions BSD is a wonderful license for game devs since you’re free to use the code and only have to add an acknowledgement somewhere.
I suppose a public domain game might hit the same limitation, though as a non-lawyer I would guess the chance of anyone with standing trying to sue anyone implementing from this spec is realistically zero (though I don’t fault stb for being unwilling to roll those dice!)
This essay says it's not possible to make a public-domain implementation of Opus. But it could be released under BSD (as libopus is), which is fine for games, as evidenced by the Licenses section of the credits in many games.
I would like Opus, but I’m using a subsonic client on iOS and my choice has been Flac (Alac?), MP3, or AAC. Opus wouldn’t play (There are some that supported it, but I didn’t like their UX).
I read almost all the way through your comment thinking there was a decent probability you were saying this new AAC encoder was written with Claude Opus.
I've never been AI guy, and have more fascination with audio. I've long stopped being excited when I read "Opus" on HN. It's refreshing when it turns out to be the audio codec.
Choosing a lossy audio codec has become such a no brainer. Either use opus and be done with it or if for some reason opus cannot be used then use aac for compatibility with insane high bitrate for good quality without having to do research on what encoder and mode to pick.
Still having a good quality and default aac encoder is great. Though I don't get why it is mainly CBR.
I have fully switched to opus for lossy since I cannot be bothered to find the sweet spot for aac encoders and bitrate. Opus simply is too good and convenient and has been for ages.
What other lossy codec is better and for what reason? Under what circumstances and use cases? I really need put effort in looking for edge cases to not choose opus.
Aac is good too, but way too many choices to make for storing mass material for the long term and be sure the quality is always good enough.
> The encoder was mainly optimized for 48Khz audio. Get over it. It's 2026, resampling is free, 48Khz is the standard. 44.1Khz will work, and so will 96Khz but use 48Khz if you want the best quality.
I think the closest thing to an actual "standard" is AES5-2018, "Recommended practice for professional digital audio".
Abstract:
> A sampling frequency of 48 kHz is recommended for the origination, processing, and interchange of audio programs employing pulse-code modulation. Recognition is also given to the use of a 44.1-kHz sampling frequency related to certain consumer digital applications, the use of a 32-kHz sampling frequency for transmission-related applications, and the use of a 96-kHz sampling frequency for applications requiring a higher bandwidth or more relaxed anti-alias filtering. This revision further quantifies the preferred choices for higher sampling frequencies.
Edit: From my personal perspective, 44.1kHz is a legacy minor annoyance
AAC has a strange quirk that the window size is dependent on the sampling rate, thus requiring a complete psychoacoustics reoptimization of all encoder parameters for each sampling rate, since a 20msec window sounds very different than a 60msec window, to human ears.
More or less. Streaming is often done with 48, video content has ben 48 for a while now, so unless you still produce content for CDs it is the standard.
44100 Hz had reasons no longer really needed (storing audio in 3 samples per line in VHS: 490 lines × 3 samples × 30 GPS = 44100 sample/s).
Qualitywise both are more than enough snd 99.99% of people would not be able to tell it apart in a blind test. Higher sample rates than 48kHz only needed when you want to pitch down ultrasonic recordings (of whales, bats and other such animals for example).
Aside from this higher than 48 kHz sample rates may have only downsides, like increased size and potential distortion in the ultrasonic frequency range that has sidebands in the audible range. Yet there is a persistent, but unscientific "more-is-better"-crowd in the HiFi-sector.
> Higher sample rates than 48kHz only needed when you want to pitch down ultrasonic recordings (of whales, bats and other such animals for example).
There are numerous use cases for higher sample rates that go beyond this but it's hard to talk about it without starting flame wars filled with junk science.
Say it or don't but "I have evidence otherwise but don't think I should say" is just as bad a flame war gateway as tempting the junk science audiophiles directly.
Higher sample rates are lower latency for the same block size and resampling is not "free" (pick 2: performance, aliasing, latency) so there can be advantages to working with audio archived at higher sample rates.
But all the advantages come down to professional or editing use cases. There's next to zero advantage to using it as a storage format for listening. Just like 24 bit audio (do you have an amp with 96dB SNR?).
Just personally, I have seen little evidence (personally, professionally, or academically) that there is any advantage for lossless audio for consumer applications. For professional applications there are plenty, and it's endlessly tiring to convince people that "no, actually I need 96kHz for my use case."
Where the audiophiles have _some_ argument here is the design of reconstruction filters which I've heard alleged can perform better in the audible frequency range if the stop band is outside of it. But I have never personally tested this, nor cared enough to. But the theory is sound.
Whether or not it's perceptible depends on what you're measuring, though. In theory, there should be perceptual differences in sound localization if your DAC's reconstruction filter is at 24kHz vs 48kHz since it will change the group delay in a critical frequency region, where you'll get sound at >~2kHz arriving later at the lower sample rate. I think it would be extremely hard to test this though, because humans are really shitty at sound localization to begin with, and practically speaking most recorded material is processed to shit in that frequency range to intentionally decorrelate the channels for the perception of "width."
I know that with oscilloscopes, it’s recommended to use 5x instead of nuquist 2x of the highest frequency you want to use., but the most reasonable argument I’ve heard for higher than 48kHz sampling is digital audio effects.
But for the end result 48kHz is more than necessary. I can’t even hear any frequency above 17kHz.
Yes and no. It is the standard for audio in film, which explains the author's focus. But is the audio CD bigger and more "standarder" than DVD and Blu-Ray? I think they're equals, and I personally think this encoder only makes sense for video content. Given all the caveats the author mentions (in particular about the sample rate) I would steer clear from using it when ripping CDs.
>FFmpeg's AAC DEcoder is busted with regards to stereo PNS, and the bug may be in other AAC decoders too, so we work around it in the encoder. Since no other encoder used PNS, the bug was not found until now.
I don't know what PNS is, but I bet this has been bothering someone's niche use-case for 20 years
The issue was twofold, on one hand, using TNS on top of PNS meant the noise that got inserted was shaped by TNS, which is nonsense since the decoder generated the noise, not the encoder. This made PNS explode.
The second, biggest issue was that using PNS in combination with any stereo tools resulted in noise leaking in both channels equally, ruining stereo imaging. So the best and only thing to do was to enable PNS only if the band in both channels is noise (or is sufficiently non-tonal and masked).
Older I get, more it seems it’s possible to ping pong between rewrites for good reasons (ex. here, metric maxes but I find it hard to believe VBR and not-48 kHz are silly things and not worth investing it)
It doesn't let me edit the link, but I'm confused by what even happened here... I posted this from my phone and that wrong link doesn't show up in my clipboard history.
Nice, I'm looking forward to seeing how this performs in practice. FFmpeg's previous AAC encoder produced poor quality output and often had irritating chirping artifacts, so I've always had to install Apple's Core Audio encoder on any computer I do video recording on to get decent sound. I've done A/B/X comparisons and found that a 320kbps MP3 sounds better than a 320kbps AAC encoded by FFmpeg, but about the same as a 256kbps AAC encoded by Core Audio. If installing Core Audio is no longer necessary, that'll be a huge improvement and people who use something like OBS to do screen recordings or streaming will get a massive sound quality boost the next time they update.
A useful project related to Apple's Core Audio is qaac - it wraps iTunes DLL's in a standalone encoding tool with CLI interface. I believe it even works under wine on Linux: https://web.archive.org/web/20250814194428/https://www.andre...
In the Hydrogenaudio discussion thread's metrics table, the new encoder scores better than Core Audio. But this is at constant bitrate (CBR) [edit: maybe not? see lesscraft's reply below]. Core Audio also has variable bitrate modes (TVBR) which the new encoder lacks.
So maybe Core Audio will continue to be the best when TVBR is available, but I'm hopeful the new FFmpeg encoder will be "good enough", especially if more folks find and contribute problem samples to help tune it.
The benchmarks were made using afconvert on OSX with the default VBR settings.
Why not use a lossless codec if you care about quality? Or use Opus, descent for specht and works pretty much anywhere these days.
> Why not use a lossless codec if you care about quality?
(1) Lossy codecs are transparent at half the file size (or less) of FLAC/ALAC.
(2) AAC (strictly, AAC-LC) is universal, where FLAC and Opus are not yet there.
There are a ton of older, but still perfectly usable devices that support AAC well but not Opus.
You can care about quality to the extent that a lossy codec allows. Lossless is not always necessary or wanted. This is like saying “why care about transcoding quality when you can keep the video as is?”. There’s a myriad of use cases and preferences at play here.
Man what a showcase for Opus this is.
Don't get me wrong, this sort of thing is a valuable exercise and we are better off with better encoders for these older codecs. But look at the numbers for Opus on this benchmark. It simply blows all the AAC encoders out of the water even at 64 kbps.
The biggest advantage for having a good AAC encoder isn't efficiency, it's that for nearly the past 2 decades the de facto standard for live streamed video has been RTMP with H.264 video and AAC audio. There is basically no support for any other codecs. If you want to send a video stream to Youtube or Twitch, you will be sending H.264 and AAC. If you want an idea of how ubiquitous this is, I just checked in OBS and it will not even let you select different video and audio codecs in streaming mode, it just (correctly) assumes that anybody who's streaming will be streaming H.264 and AAC.
Plus, at 96+ kbps (assuming an Apple-quality AAC-LC encoder) Opus loses its quality advantage. So at higher bitrates, the benefit of choosing Opus is that encoders/decoders are royalty-free.
Also the fact that hardware-accelerated AAC and even full AAC offload is ubiquitous in modern-ish hardware. I think my rice cooker can play AAC audio
No one really offloads AAC, apart from Apple. Opus can be decoded on very cheap microcontrollers entirely in software using the reference library.
Sample accurate editing is with AAC is a pain though. Especially if you also have video, because frame rates are usually incompatible.
If you want flexibility without fully transcoding both audio and video, Opus is your friend
Opus is your friend as long as the software you’re using supports it—besides, Apple’s AAC-LC can beat out Opus in low bitrates scenarios.
Whether you like it or not, AAC is still the standard.
I think the biggest issue with Opus is the problem with its specification being lacking, see:
https://nothings.org/stb/stb_opus.html
This essentially causes opus to never be used in games or in things in stores that may have issues with specific licenses.
That’s going a bit far. I’m in the games industry and have used opus regularly, it’s a great codec for games, often the hardware decoding is so restricted that we’re using software regardless so we might as well use something like opus.
The licensing restriction is unfortunate, but only restrictive for those with very specific goals, under normal conditions BSD is a wonderful license for game devs since you’re free to use the code and only have to add an acknowledgement somewhere.
I suppose a public domain game might hit the same limitation, though as a non-lawyer I would guess the chance of anyone with standing trying to sue anyone implementing from this spec is realistically zero (though I don’t fault stb for being unwilling to roll those dice!)
This essay says it's not possible to make a public-domain implementation of Opus. But it could be released under BSD (as libopus is), which is fine for games, as evidenced by the Licenses section of the credits in many games.
I would like Opus, but I’m using a subsonic client on iOS and my choice has been Flac (Alac?), MP3, or AAC. Opus wouldn’t play (There are some that supported it, but I didn’t like their UX).
You might like Poppy (in beta), which supports all media servers (including OpenSubsonic/Navidrome) and Opus as a first-class music format. https://www.reddit.com/r/PoppyApp/comments/1tiyki0/about_pop...
> Man what a showcase for Opus this is.
I take it you mean this Opus (https://en.wikipedia.org/wiki/Opus_(audio_format)) not that Opus (https://en.wikipedia.org/wiki/Claude_(AI)).
I read almost all the way through your comment thinking there was a decent probability you were saying this new AAC encoder was written with Claude Opus.
I've never been AI guy, and have more fascination with audio. I've long stopped being excited when I read "Opus" on HN. It's refreshing when it turns out to be the audio codec.
Choosing a lossy audio codec has become such a no brainer. Either use opus and be done with it or if for some reason opus cannot be used then use aac for compatibility with insane high bitrate for good quality without having to do research on what encoder and mode to pick.
Still having a good quality and default aac encoder is great. Though I don't get why it is mainly CBR.
> Choosing a lossy audio codec has become such a no brainer.
Falser words hath never been spoken.
> Falser words hath never been spoken.
Why? Care to explain?
I have fully switched to opus for lossy since I cannot be bothered to find the sweet spot for aac encoders and bitrate. Opus simply is too good and convenient and has been for ages.
What other lossy codec is better and for what reason? Under what circumstances and use cases? I really need put effort in looking for edge cases to not choose opus.
Aac is good too, but way too many choices to make for storing mass material for the long term and be sure the quality is always good enough.
> The encoder was mainly optimized for 48Khz audio. Get over it. It's 2026, resampling is free, 48Khz is the standard. 44.1Khz will work, and so will 96Khz but use 48Khz if you want the best quality.
Is 48kHz really the standard nowadays?
I think the closest thing to an actual "standard" is AES5-2018, "Recommended practice for professional digital audio".
Abstract:
> A sampling frequency of 48 kHz is recommended for the origination, processing, and interchange of audio programs employing pulse-code modulation. Recognition is also given to the use of a 44.1-kHz sampling frequency related to certain consumer digital applications, the use of a 32-kHz sampling frequency for transmission-related applications, and the use of a 96-kHz sampling frequency for applications requiring a higher bandwidth or more relaxed anti-alias filtering. This revision further quantifies the preferred choices for higher sampling frequencies.
Edit: From my personal perspective, 44.1kHz is a legacy minor annoyance
48kHz makes alignment between video and audio so much easier. (I.e.: Lip synchronization after edits)
AAC has a strange quirk that the window size is dependent on the sampling rate, thus requiring a complete psychoacoustics reoptimization of all encoder parameters for each sampling rate, since a 20msec window sounds very different than a 60msec window, to human ears.
This was of course fixed in Opus.
Pretty much all DACs run at 48Khz by default due to operating systems picking it as a sane default.
I know the opus codec assumes everything is 48kHz and will resample inputs to that.
For one, audio transcription services that use Whisper will sample the input down to 16Khz mono first.
More or less. Streaming is often done with 48, video content has ben 48 for a while now, so unless you still produce content for CDs it is the standard.
44100 Hz had reasons no longer really needed (storing audio in 3 samples per line in VHS: 490 lines × 3 samples × 30 GPS = 44100 sample/s).
Qualitywise both are more than enough snd 99.99% of people would not be able to tell it apart in a blind test. Higher sample rates than 48kHz only needed when you want to pitch down ultrasonic recordings (of whales, bats and other such animals for example).
Aside from this higher than 48 kHz sample rates may have only downsides, like increased size and potential distortion in the ultrasonic frequency range that has sidebands in the audible range. Yet there is a persistent, but unscientific "more-is-better"-crowd in the HiFi-sector.
> Higher sample rates than 48kHz only needed when you want to pitch down ultrasonic recordings (of whales, bats and other such animals for example).
There are numerous use cases for higher sample rates that go beyond this but it's hard to talk about it without starting flame wars filled with junk science.
Say it or don't but "I have evidence otherwise but don't think I should say" is just as bad a flame war gateway as tempting the junk science audiophiles directly.
Higher sample rates are lower latency for the same block size and resampling is not "free" (pick 2: performance, aliasing, latency) so there can be advantages to working with audio archived at higher sample rates.
But all the advantages come down to professional or editing use cases. There's next to zero advantage to using it as a storage format for listening. Just like 24 bit audio (do you have an amp with 96dB SNR?).
Just personally, I have seen little evidence (personally, professionally, or academically) that there is any advantage for lossless audio for consumer applications. For professional applications there are plenty, and it's endlessly tiring to convince people that "no, actually I need 96kHz for my use case."
Where the audiophiles have _some_ argument here is the design of reconstruction filters which I've heard alleged can perform better in the audible frequency range if the stop band is outside of it. But I have never personally tested this, nor cared enough to. But the theory is sound.
Whether or not it's perceptible depends on what you're measuring, though. In theory, there should be perceptual differences in sound localization if your DAC's reconstruction filter is at 24kHz vs 48kHz since it will change the group delay in a critical frequency region, where you'll get sound at >~2kHz arriving later at the lower sample rate. I think it would be extremely hard to test this though, because humans are really shitty at sound localization to begin with, and practically speaking most recorded material is processed to shit in that frequency range to intentionally decorrelate the channels for the perception of "width."
I know that with oscilloscopes, it’s recommended to use 5x instead of nuquist 2x of the highest frequency you want to use., but the most reasonable argument I’ve heard for higher than 48kHz sampling is digital audio effects.
But for the end result 48kHz is more than necessary. I can’t even hear any frequency above 17kHz.
Yes, bit depth headroom is very useful for audio production to avoid aliasing. Pro DAWs support 96KHz.
Yes and no. It is the standard for audio in film, which explains the author's focus. But is the audio CD bigger and more "standarder" than DVD and Blu-Ray? I think they're equals, and I personally think this encoder only makes sense for video content. Given all the caveats the author mentions (in particular about the sample rate) I would steer clear from using it when ripping CDs.
Yes, pretty much all new hardware uses it as default output setting as well (by that I mean laptops, phones, smart speakers, etc.)
48kHz has been the recommended setting with Premiere Pro as long as I can remember.
44.1kHz, isn't that what lameMP3 uses as default?
It's what CDs use, so it would make sense for mp3 encoders to follow suit.
A very welcomed addition, hopefully I can replace fdk-aac
>FFmpeg's AAC DEcoder is busted with regards to stereo PNS, and the bug may be in other AAC decoders too, so we work around it in the encoder. Since no other encoder used PNS, the bug was not found until now.
I don't know what PNS is, but I bet this has been bothering someone's niche use-case for 20 years
The issue was twofold, on one hand, using TNS on top of PNS meant the noise that got inserted was shaped by TNS, which is nonsense since the decoder generated the noise, not the encoder. This made PNS explode. The second, biggest issue was that using PNS in combination with any stereo tools resulted in noise leaking in both channels equally, ruining stereo imaging. So the best and only thing to do was to enable PNS only if the band in both channels is noise (or is sufficiently non-tonal and masked).
https://www.audiolabs-erlangen.de/content/resources/aesCodin...
Hah, this sounds like the audio equivalent of Netflix’s grain reconstruction.
Older I get, more it seems it’s possible to ping pong between rewrites for good reasons (ex. here, metric maxes but I find it hard to believe VBR and not-48 kHz are silly things and not worth investing it)
Flagged for the wrong link.
Hopefully they see this - there's still time to edit the submission link.
It doesn't let me edit the link, but I'm confused by what even happened here... I posted this from my phone and that wrong link doesn't show up in my clipboard history.
Link should be: https://hydrogenaudio.org/index.php/topic,129691.0.html
Your options are:
* quick email to HN@ycombinator.com with a "Help Me please!! and link ( mods can edit link in and sideline (hide) these comments )
* Just live with the rotting fish head of public boo boo (we've all made mistakes, as the Dalek said whilst climbing down off the dustbin)
* I can kill the whole thing dead.
It's fixed now.
Our software follows redirs and somehow we got a 302 to our own IP. Perhaps it is someone's idea of a bot detector?