HydrogenAudio

Lossy Audio Compression => MP3 => MP3 - General => Topic started by: lozenge on 2020-07-28 16:26:04

Title: Happy 25th Birthday MP3
Post by: lozenge on 2020-07-28 16:26:04
I saw this article and thought of you .... apparently 14th July 2020 was the 25th Anniversary since the MP3 vile extension was chosen.

https://hackaday.com/2020/07/27/mp3-is-25-years-old/
Title: Re: Happy 25th Birthday MP3
Post by: binaryhermit on 2020-07-28 18:03:59
If you believe the media, didn't MP3 die in like 2017?
Title: Re: Happy 25th Birthday MP3
Post by: john33 on 2020-07-28 18:34:55
Rumours of its death are grossly exagerated!! ;)
Title: Re: Happy 25th Birthday MP3
Post by: includemeout on 2020-07-28 19:24:49
It's funny how its detractors have been insisting on flogging this - according to their own expert [cough!] assessment - apparently dead horse for 15 years or so and will keep on doing for the next 15!

A zombie horse perhaps?  :))
Title: Re: Happy 25th Birthday MP3
Post by: includemeout on 2020-07-28 19:26:07
Rumours of its death are grossly exagerated!! ;)
 
 LOL! Was it Groucho who said that?  ;D
Title: Re: Happy 25th Birthday MP3
Post by: john33 on 2020-07-28 21:09:22
It's a paraphrase of words attributed to Mark Twain.
Title: Re: Happy 25th Birthday MP3
Post by: binaryhermit on 2020-07-29 03:44:04
Right, IMO, the "news" that was interpreted as "MP3 was dead" (AKA the last patent expiring) actually made MP3 more alive than it had been in years.
Title: Re: Happy 25th Birthday MP3
Post by: ani_Jackal3 on 2020-08-09 19:35:15
Right, IMO, the "news" that was interpreted as "MP3 was dead" (AKA the last patent expiring) actually made MP3 more alive than it had been in years.

Yep it now a free codec, With a custom decoder would be fun to see a MP3 encoder. That can add new features to fix it limitions & allow 8 ~ 640kbps in VBR, I'm sure their tricks to make older MP3 decoders not freak out if such a thing was made.
Title: Re: Happy 25th Birthday MP3
Post by: binaryhermit on 2020-08-09 20:00:07
I think you'd end up with something that does worse quality-per-bit than and is less compatible than, say, Ogg Vorbis or Opus or a reasonably well supported variant of AAC.

EDIT: if you tried to make something that was "better than" and also "close to but not actually" mp3 technique-wise.
Title: Re: Happy 25th Birthday MP3
Post by: ani_Jackal3 on 2020-08-10 08:39:01
I think you'd end up with something that does worse quality-per-bit than and is less compatible than, say, Ogg Vorbis or Opus or a reasonably well supported variant of AAC.

EDIT: if you tried to make something that was "better than" and also "close to but not actually" mp3 technique-wise.

Modern players can use 3rd party decoders for the 8 ~ 640kbps bit i said, But safe mode just locks it too 32 ~ 320kbps for older players. Using modern codec features would be a big gain if it means it can be tranparent at 160kbps instead of 192 ~ 320kbps on current encoders while still being compatible with 99% of mp3 players.

Many people still stick with 192kbps Lame MP3 in 2020, Despite AAC/Opus being around.

Title: Re: Happy 25th Birthday MP3
Post by: kode54 on 2020-08-12 01:14:23
Out of spec MP3 is not supported in a VBR configuration, because free format MP3 specifies a packet length of 0xFFFF, which means length is not knowable. You literally have to scan the stream forward for sync headers and process each packet one by one to determine their size.
Title: Re: Happy 25th Birthday MP3
Post by: celona on 2020-08-13 18:34:17
I saw this article and thought of you .... apparently 14th July 2020 was the 25th Anniversary since the MP3 vile extension was chosen.

https://hackaday.com/2020/07/27/mp3-is-25-years-old/


Check the date of the MPEG-2 AAC: https://webstore.iec.ch/preview/info_isoiec13818-3%7Bed2.0%7Den.pdf

https://web.archive.org/web/20050324041934/http://www.chiariglione.org/mpeg/standards/mpeg-1/mpeg-1.htm
Title: Re: Happy 25th Birthday MP3
Post by: C.R.Helmrich on 2020-08-15 13:20:11
Check the date of the MPEG-2 AAC: https://webstore.iec.ch/preview/info_isoiec13818-3%7Bed2.0%7Den.pdf
I believe ISO/IEC 13818-3 is some MPEG-1 backward-compatible extension, MPEG-2 AAC is ISO/IEC 13818-7 (originally from 1997, made that mistake myself once).

By the way, the original email (from Friday, July 14, 1995) with the decision to use .mp3 as file extension can be found in this tweet (https://twitter.com/fraunhofer/status/620853977920458752). Rough translation from German:
Quote
Hello,
given the overwhelming opinion of all respondents:
the ending for ISO MPEG Audio Layer 3 is .mp3. That means we should, for upcoming WWW pages, shareware, demos, etc., make sure that .bit endings won't be sent out anymore. There's a reason for it, believe me :-)
J├╝rgen Zeller

Chris
Title: Re: Happy 25th Birthday MP3
Post by: ani_Jackal3 on 2021-05-20 12:17:22
Out of spec MP3 is not supported in a VBR configuration, because free format MP3 specifies a packet length of 0xFFFF, which means length is not knowable. You literally have to scan the stream forward for sync headers and process each packet one by one to determine their size.

Being 8 ~ 640kbps still within standard which a new decoder could fix, But old ones would be capped at 32 ~ 320kbps frames. I've had DAP that capped Vorbis/Opus to 320kbps only despite the both codecs could do VBR 512kbps. There are already a few old MP3 decoders that can do 8 ~ 640kbps with no issues.

I've noticed when doing V0 encodes with LAME, the bitrate can reach 328kbps for 1 second. Showing that within the actual frames could be 400kbps?.



Title: Re: Happy 25th Birthday MP3
Post by: Porcus on 2021-05-20 13:27:33
Being 8 ~ 640kbps still within standard which a new decoder could fix, But old ones would be capped at 32 ~ 320kbps frames. I've had DAP that capped Vorbis/Opus to 320kbps only despite the both codecs could do VBR 512kbps. There are already a few old MP3 decoders that can do 8 ~ 640kbps with no issues.

Yeah, there are old ones that can handle this obsolete format. But just to give you a picture:
Take a software suite whose claim-to-fame is its ability to handle pretty much any useful format - ffmpeg.
* It cannot decode free-format mp3 whatsoever.
* Slightly over 320 kbit/s it cannot even re-encapsulate. Strangely enough it does not cap it at 320; I just tested 325 (OK) and 345 (failed).

Trying to extend this dead horse from CBR to VBR ... there is no way for me to know whether it would even be possible, but if you want a useful audio format, look elsewhere.
SimplePortal 1.0.0 RC1 © 2008-2021