#510293 - 03/13/04 05:06 AM
MP3, anything better?
|
natpub
Senior Member
Registered: 11/25/03
Posts: 96
Loc: Austin, TX
|
Offline
|
|
Does it seem like people become so used to hearing MP3 compression that they come to accept it as being what good music is supposed to sound like?
Surely there is a better way to squeeze digital files into a small space that still sound great. Is there anything on the horizon?
KT
|
|
Top
|
|
|
|
|
#510294 - 03/13/04 11:46 AM
Re: MP3, anything better?
|
adebar
Gold Member
Registered: 07/17/01
Posts: 709
Loc: Wiesbaden, GERMANY
|
|
Offline
|
|
Yes, AAC is better than MP3.
Either better quality at the same data rate or less data rate with the same quality.
More
http://www.vialicensing.com/products/mpeg4aac/standard.html
|
|
Top
|
|
|
|
|
#510295 - 03/13/04 11:54 AM
Re: MP3, anything better?
|
adebar
Gold Member
Registered: 07/17/01
Posts: 709
Loc: Wiesbaden, GERMANY
|
|
Offline
|
|
Forgot to mention this site
http://www.iis.fraunhofer.de/amm/techinf/aac/
which may be more informative.
|
|
Top
|
|
|
|
|
#510296 - 03/13/04 01:03 PM
Re: MP3, anything better?
|
Bob Olhsson
Senior Member
Registered: 05/25/01
Posts: 426
Loc: Nashville,TN, USA
|
|
Offline
|
|
James Johnston who was a principal architect of the MPEG audio codecs and the father of AAC is now developing audio codecs for Microsoft. His expertise combined with a number of audio technology patents that Microsoft owns could easily make Windoze Media the dominant format of the future based entirely on superior technology.
I personally find this a bit chilling but the fact remains that only Microsoft was willing to ante up and develop rather than just licensing other people's technology. At this point they are the only people I'm aware of actively doing research.
|
|
Top
|
|
|
|
|
#510297 - 03/13/04 01:10 PM
Re: MP3, anything better?
|
Lee Flier
10k Club
Registered: 09/13/00
Posts: 15398
Loc: Atlanta,GA,UNITED STATES
|
|
Offline
|
|
That DOES really suck, Bob.
Ogg Vorbis sounds really good (better than MP3 IMO), is still in active development, and is open source.
|
|
Top
|
|
|
|
|
#510298 - 03/13/04 01:11 PM
Re: MP3, anything better?
|
flyscots
MP Hall of Fame Member
Registered: 12/27/01
Posts: 2258
Loc: SW Scotland.
|
|
Offline
|
|
Originally posted by natpub:
Does it seem like people become so used to hearing MP3 compression that they come to accept it as being what good music is supposed to sound like?
|
|
Top
|
|
|
|
|
#510299 - 03/13/04 01:46 PM
Re: MP3, anything better?
|
Bob Olhsson
Senior Member
Registered: 05/25/01
Posts: 426
Loc: Nashville,TN, USA
|
|
Offline
|
|
Originally posted by Lee Flier:
...Ogg Vorbis sounds really good (better than MP3 IMO), is still in active development, and is open source.
|
|
Top
|
|
|
|
|
#510301 - 03/13/04 10:42 PM
Re: MP3, anything better?
|
Pio2001
Senior Member
Registered: 10/05/01
Posts: 63
Loc: Lyon,AK,FRANCE
|
|
Offline
|
|
The following statements are based on double blind listening tests between original files and compressed files, preformed by the HydrogenAudio community .
WMA, MP3pro and RealAudio are not optimized for high quality compression. They excel in strong compression ratio, with a lot of losses from the original sound.
Other codecs (MP3, Vorbis, AAC, MPC) have a point where they reach nearly transparent sound. After this, the filesize grows insanely, for improvements only audible by trained people, and only on very sensitive musical samples.
MP3 is rather complicated. There are a lot of different codecs. Lame and Fraunhofer's are considered as high quality codecs. Lame reach its optimal point (nearly transparent) with the option "--alt-preset standard", in version 3.90.3, 3.93.1 and upper ( infos http://www.hydrogenaudio.org/index.php?act=ST&f=15&t=203 ). The bitrate is then around 190-200 kbps.
I don't know Vorbis nor AAC, but they are considered better than MP3 at handling pre-echo.
MPC (MusePack) is currently the best quality audio encoder. Its optimal point is around 180 kbps.
Then there are lossless codecs, that can reconstruct the original PCM data bit for bit : Flac, Monkey's audio, etc. But their bitrate is rather around 800 kbps ! (The Uncompressed CD PCM bitrate is 1411 kbps).
|
|
Top
|
|
|
|
|
#510302 - 03/14/04 01:46 AM
Re: MP3, anything better?
|
Bob Olhsson
Senior Member
Registered: 05/25/01
Posts: 426
Loc: Nashville,TN, USA
|
|
Offline
|
|
This doesn't sound like a description of windoze media-9 which is pretty darned impressive compared to earlier versions.
|
|
Top
|
|
|
|
|
#510303 - 03/14/04 01:55 AM
Re: MP3, anything better?
|
Uh Clem
MP Hall of Fame Member
Registered: 03/13/00
Posts: 3442
Loc: Atlanta, GA USA
|
|
Offline
|
|
Originally posted by Bob Olhsson:
This doesn't sound like a description of windoze media-9 which is pretty darned impressive compared to earlier versions.
|
|
Top
|
|
|
|
|
#510304 - 03/14/04 02:21 AM
Re: MP3, anything better?
|
Matt.Hepworth
MP Hall of Fame Member
Registered: 03/13/01
Posts: 2972
Loc: Riverdale, UT
|
|
Offline
|
|
Call me crazy, but I would say that wma is superior to mp3 at bitrates of 192 and above. Although you lose some depth, that "crickets in a tornado" sound is not there like it is with mp3.
_________________________
No matter how good something is, there will always be someone blasting away on a forum somewhere about how much they hate it.
|
|
Top
|
|
|
|
|
#510306 - 03/14/04 04:04 AM
Re: MP3, anything better?
|
Curve Dominant
MP Hall of Fame Member
Registered: 10/29/00
Posts: 4223
Loc: Philadelphia USA
|
|
Offline
|
|
The Fraunhofer Institute mp3 encoder is the default ProTools encoder, which I've been using for both my website, and emailing music.
No one has complained about the sonic fidelity - on the contrary. But then again, I'm careful to set the encoder to "maximum quality" which takes up to 20 minutes to encode a 1&1/2 minute song excerpt on my machine.
I've evolved to appreciate the mp3 platform. It will not damage a good mix, not even close. We've done A/B tests of the mp3's against our PT mixes off the system, and the difference is not offensive to the point that we're reluctant to publish mp3's (and we can be a pretty picky bunch when it comes to sonics).
The portability of the mp3 far outweighs any sonic degradation as far as we are concerned here. The fact that we can instantly publish new song mixes, or email them across the globe, is a heavy freedom of expression issue...but that's not really the subject of this thread, so I digress.
But if I may put this subject into context...I'm old enough to remember when cassette tapes were the defacto delivery medium for song demos. Oh, the agony of the cassette tape. As bad as some "golden ears" say the mp3 is, for me personally, the mp3 is a godsend if matched against the bloody cassette tape. At least I know what mp3 is going to do to the sound of my mix. With a cassette tape, you never knew what it was going to sound like on any given cassette machine...what speed it would play, the bias, with or without Dolby, clean heads or filthy...Argh...the horror.
Not so with mp3. We just posted a new mix on our website of a song we are developing, and we know exactly how it will sound for anybody across the globe who listens to it . There's something to be said for consistency.
|
|
Top
|
|
|
|
|
#510307 - 03/14/04 08:03 AM
Re: MP3, anything better?
|
natpub
Senior Member
Registered: 11/25/03
Posts: 96
Loc: Austin, TX
|
|
Offline
|
|
Then, do we mix with MP3 in mind?
I looked at all the links and such shown above, but I am not really heartened.
When you jump from downloading a song to burning it, then play it in the car (which I believe is what most folks are doing), to me there is a noticible difference, and it isnt good. Eveything is slanted to high-mids, which is naseating to
Is the Mac format for i-tunes any different? I have not heard it, I don't own an i-Pod, I guess I should.
I have followed Ogg Vorbis for a while, but I am not inspired, nor has it caught fire.
I agree that highly-compressed formats may be undifferentiated by mass listeners, but to me, that isn't saying much. We deal on a psychological and emotional plane, and those same listeners thought cassette sounded fine too.
I remain worried that degraded formats will become more and more prevelent, in our hurry-up world. I do not feel that is what we should be pushing for, or supporting.
I have Cable TV at home, and when I want a movie, I just click I-Control and order one. I dont drive, I dont rent, and I love it. But, it doesnt use my surround system, and I know the visual and audio quality is suffering.
I believe that High-Def visual will make a difference, (I have no earthly idea how they ever made Congress "require" a format like that). I hope that we in the audio world continue to push for similar definition requirements in our field.
Regards,
Kurt T.
Austin
|
|
Top
|
|
|
|
|
#510308 - 03/14/04 06:59 PM
Re: MP3, anything better?
|
Pio2001
Senior Member
Registered: 10/05/01
Posts: 63
Loc: Lyon,AK,FRANCE
|
|
Offline
|
|
Originally posted by NE-One:
Call me crazy, but I would say that wma is superior to mp3 at bitrates of 192 and above. Although you lose some depth, that "crickets in a tornado" sound is not there like it is with mp3.
|
|
Top
|
|
|
|
|
#510309 - 03/15/04 02:34 AM
Re: MP3, anything better?
|
Matt.Hepworth
MP Hall of Fame Member
Registered: 03/13/01
Posts: 2972
Loc: Riverdale, UT
|
|
Offline
|
|
Windows media player will encode to wma (it's a microsoft format). I am willing to do some testing though. Let me know.
_________________________
No matter how good something is, there will always be someone blasting away on a forum somewhere about how much they hate it.
|
|
Top
|
|
|
|
|
#510310 - 03/15/04 08:46 AM
Re: MP3, anything better?
|
Will Russell
Senior Member
Registered: 04/25/00
Posts: 144
Loc: Newfield, NY
|
|
Offline
|
|
AAC-Advanced Audio Coding sounds much better than MP3 at any bitrate. At 320, it is very listenable even on nice playback systems.
|
|
Top
|
|
|
|
|
#510311 - 03/15/04 10:25 AM
Re: MP3, anything better?
|
Bob Olhsson
Senior Member
Registered: 05/25/01
Posts: 426
Loc: Nashville,TN, USA
|
|
Offline
|
|
You know the irony is that when everything isn't peak-limited to death so that there is still some dynamic range left to hide the distortion in, MP-3, AAC and Windoze Media all sound great requiring no special mix technique at all.
|
|
Top
|
|
|
|
|
#510312 - 03/15/04 04:10 PM
Re: MP3, anything better?
|
Pio2001
Senior Member
Registered: 10/05/01
Posts: 63
Loc: Lyon,AK,FRANCE
|
|
Offline
|
|
To avoid this problem, you must use the replaygain plugin (mp3gain, for mp3), in order to prevent additional clipping introduced by the decoder.
|
|
Top
|
|
|
|
|
#510313 - 03/15/04 06:08 PM
Re: MP3, anything better?
|
Bob Olhsson
Senior Member
Registered: 05/25/01
Posts: 426
Loc: Nashville,TN, USA
|
|
Offline
|
|
Preventing clipping is a good idea but it still doesn't sound nearly as good as a mix with no peak limiting.
|
|
Top
|
|
|
|
|
#510314 - 03/15/04 06:17 PM
Re: MP3, anything better?
|
Lee Flier
10k Club
Registered: 09/13/00
Posts: 15398
Loc: Atlanta,GA,UNITED STATES
|
|
Offline
|
|
Originally posted by NE-One:
Call me crazy, but I would say that wma is superior to mp3 at bitrates of 192 and above. Although you lose some depth, that "crickets in a tornado" sound is not there like it is with mp3.
|
|
Top
|
|
|
|
|
#510316 - 03/19/04 04:35 AM
Re: MP3, anything better?
|
Matt.Hepworth
MP Hall of Fame Member
Registered: 03/13/01
Posts: 2972
Loc: Riverdale, UT
|
|
Offline
|
|
Originally posted by Lee Flier:
That sound isn't there with a properly encoded MP3 either. I think a lot of people judge MP3's by having heard a crappy codec, and probably at low bit rates to boot.  (Kidding, of course!) With WMA at 192, it sounds pretty normal to my ear, and clearer than mp3 (especially in the HF).
_________________________
No matter how good something is, there will always be someone blasting away on a forum somewhere about how much they hate it.
|
|
Top
|
|
|
|
|
#510317 - 03/21/04 12:41 PM
Re: MP3, anything better?
|
Pio2001
Senior Member
Registered: 10/05/01
Posts: 63
Loc: Lyon,AK,FRANCE
|
|
Offline
|
|
I've performed some listening tests between MP3, MusePack, and WMA9. Since we look for audio transparency at a decent size, the following encoding options were used :
MP3 ecoder : Lame
Lame version : 3.90.3
Lame options : --alt-preset standard
...which leads to VBR bitrates around 190 kbps.
Musepack version : 1.14 beta , with the option --xlevel.
Since no quality setting is specified, the encoder uses the "standard" profile, which leads to VBR bitrates around 175 kbps.
WMA Version : Windows Media Player 9. Windows Media Audio VBR. 135-215 kbps.
___________________
I used audio samples known to be very difficult to encode faithfully :
Badvilbel (from Autechre - Anvil Vapre)
Drone (from another Autechre album)
Spahm (?)
Short version for testing purposes are available at FF123\'s sample page
I then used a CD of mine : Bel Canto - White-Out Conditions. It should not present any difficulty for encoding.You can get the sample that I used here .
____________________
I decoded the encoded files with Foobar2000 0.7.3a with no processing.
I performed the tests in double blind ABX , in order to eliminate any bias that I might have towards one format. It would be too easy to tell that my prefered format is the one that sounds best. But no. At no time I was aware of what file I was listening to.
If the test is successful, it means that I can hear a difference. If not, it means that I can't hear a difference, even if I think I can.
I compared every file against the original wav file.
Listening system :
Windows XP, ABX Comparator player, Wave device = Marc 2 digital out
Soundcard = Marian Marc 2. SPDIF Optical ouput. (non-dithering-non-resampling soundcard, even with XP). 5 meters optical cable. Checked to be bitwise accurate.
DAC : Sony DTC 55ES DAT deck.
Ampli : Arcam Diva A85
Headphones : Senheiser HD600
__________________________
Results.
First the killer samples.
Badvilbel :
MP3 : ABX tests successful (11/11) Lots of pre echo, loss of transients, drop outs
MPC : ABX test successful (7/8) Gurgling on transients, ringing, drop outs, but better
WMA : ABX test successful (8/8) Best of all. Only transient problems, no drop outs at all
Drone :
MP3 : ABX test successful (16/16) toooo easy ! Artifacts during the first half second, and time smearing during the second half
MPC : ABX test failed (3/8) No audible difference.
WMA : ABX test successful (16/16). Unaccepable quality. Worst of all.
Spahm :
MP3 : ABX test successful (16/16). Noisy
MPC : ABX test successful (16/16). Same effect as transients in badvilbel : sounds grungy while MP3 sounds hissy
WMA : ABX test successful (16/17, I went too fast and made one mistake). Worst of all.
Then the normal CD
Bel Canto :
MP3 : ABX test failed (I don't remember my score). No audible difference.
MPC : ABX test failed (stopped at 1/4 for 16 planned sessions, and I wanted a confidence >99%). No audible difference.
WMA : ABX test successful (16/16). Annoying artifacts.
Since I'm currently participating in the new Lame version testing, I will test more samples later. Tell me if you are interested into more WMA test results.
|
|
Top
|
|
|
|
|
#510318 - 03/21/04 06:09 PM
Re: MP3, anything better?
|
gm
MP Hall of Fame Member
Registered: 02/10/00
Posts: 2184
Loc: Williamson County, TN, USA
|
|
Offline
|
|
Originally posted by Bob Olhsson:
James Johnston who was a principal architect of the MPEG audio codecs and the father of AAC is now developing audio codecs for Microsoft. His expertise combined with a number of audio technology patents that Microsoft owns could easily make Windoze Media the dominant format of the future based entirely on superior technology.
I personally find this a bit chilling but the fact remains that only Microsoft was willing to ante up and develop rather than just licensing other people's technology. At this point they are the only people I'm aware of actively doing research.
I talked to Jim about this when Microsoft came to a couple of us to sell us on the benefits of WM9. In fact, the audio component was an easy sell because Quicktime doesn't do multichannel and WM9 does. Jim said at that time last fall that the codecs deployed in the current WM9 aren't his, but that he has hopes of getting his new work in sometime in the future.
Between Jim and and Karlheinz Brandenberg the new codecs have gotten pretty good (MPEG4, which does do multichannel, and AAC).
George
|
|
Top
|
|
|
|
|
#510319 - 03/21/04 07:20 PM
Re: MP3, anything better?
|
philbo_Tangent
Platinum Member
Registered: 10/08/00
Posts: 1175
Loc: Iowa
|
|
Offline
|
|
Originally posted by NE-One: Originally posted by Lee Flier:
That sound isn't there with a properly encoded MP3 either. I think a lot of people judge MP3's by having heard a crappy codec, and probably at low bit rates to boot.
One thing I've noticed with the LAME encoder at decent bit rates (192 K & over) is that you get a real improvement in sound quality if you turn off the low-pass filter that is 'on' by default.
With it off,you get the whole top octave back. With it on, everything above 15 KHz gets attenuated pretty badly.
I don't recall the command line option to do this, offhand; I use the Razor-LAME shell program, and the option is buried in one of the tabs.
Ogg Vorbis - -
The sound quality on Ogg Vorbis beats the pants off the best MP3 I've ever heard, IMHO. The file sizes are fairly comparable to 256K MP3s.
It'd be nice if Ogg supported surround formats. It might then have a shot at usurping MP3 as the defacto standard for compression.
MP4 - -
I haven't been terribly impressed with the MP4/AAC encoder I got with Nero 6. It isn't any better than MP3. Haven't tried any other MP4 codecs yet... But I'm sure the technology will mature and improve with time.
|
|
Top
|
|
|
|
|
|
|
