Re: empirical fact: good mp3 sounds as good as wav.

Yes... but you can see here, that among the lossy codecs, only OGG is almost losless here at 192 kbps

Le mieux est l'ennemi du bien

52 (edited by zip 2008-08-09 16:30:34)

Re: empirical fact: good mp3 sounds as good as wav.

@robotworks: with what kind of sound did you test that?
it wouldn't surprise me if the musical type would make a difference
maybe you should test it with noise or something

Re: empirical fact: good mp3 sounds as good as wav.

...but still not current common standard, if one's got a web site with dj sets on it and wants them to be immediately available for everyone in a world of stimulus satiation and impatience then one still should provide them as mp3...
...ogg is quite ok in that connection..., but true, there can't be any development of common standards if everyone only looks about common standards...
...oh, i just love those on the one hand/on the other hand things...;)

very interesting graphic !

54 (edited by Robotworks 2008-08-09 16:58:53)

Re: empirical fact: good mp3 sounds as good as wav.

I used a sound sample I generated using sines: 20 Hz 40 Hz, 80 Hz, 160 hz, 320 Hz etc. till 22400 Hz (that way I have well separated rays on the spectrum)


But here is a test of the MP3 lame encoder with a white noise:

http://www.fileden.com/files/2006/11/14/378671/noisemp3.jpg

Le mieux est l'ennemi du bien

55 (edited by zip 2008-08-09 17:12:06)

Re: empirical fact: good mp3 sounds as good as wav.

bias wrote:

is this high frequency data usefull anyway?

that's what compression is based on isn't it, the fact that you can't hear such high frequencies? but maybe they affect the sound still in a way you can experience, don't know how that works

bias wrote:

because even my needle has frequency response ending at 15khz(ortofon elektro) and i don't hear any lack of high frequences.
or do i confuse different things?

that would only mean it wouldn't make a difference (regarding frequencies higher than 15khz) when recording mp3's from vinyl, but it still can make a difference with the original music source I think

@robotworks: interesting stuff!
you should test if there's a difference between a minimal electro record and some acoustic disco stuff
i can imagine the loss is worse for the latter than the first, or am I kicking in open doors here?

56 (edited by Robotworks 2008-08-09 18:21:20)

Re: empirical fact: good mp3 sounds as good as wav.

Human ear has a 20 Hz to 20 KHz frequency response, up to 22 KHz in some cases.
Ortofon Concorde electro has a 20 to 22 kHz frequency range at -3 dB.
It only falls at 15 kHz at -2 dB but high frequencies are still here

http://www.djdeals.com/ortofonCONCORDEELEKTRO.htm

@zip: Regarding minimal electro and acoustic disco, I think the loss will just be the same graphically, but the subjective effect of that loss will not be the same. The ambiance of a minimal electo track would resist better to the loss...

Le mieux est l'ennemi du bien

57 (edited by Robotworks 2008-08-09 18:26:53)

Re: empirical fact: good mp3 sounds as good as wav.

-3 db from 0... it's just a matter of volume level (on the original record)... The cartridge can't play high frequencies if they have a level that is over -3 dB. But many records are pressed at -3 dB, even -6dB sometimes

Le mieux est l'ennemi du bien

Re: empirical fact: good mp3 sounds as good as wav.

Here the dB means the same thing than Volume Unit (VU)
Above Zero dB you get clipping and distortion


http://www.fileden.com/files/2006/11/14/378671/VU_Meter.jpg

Le mieux est l'ennemi du bien

Re: empirical fact: good mp3 sounds as good as wav.

This all very instructive stuff ... thanks RobotWorks for the tests and marvelous graphs

Re: empirical fact: good mp3 sounds as good as wav.

What about 24b 96 kHz flac rips? The 96 kHz refers to the sample rate. It means that the audio is being recorded (sampled ) 96 thousand times per second. In essence, the recording device or software is taking a

freakazoids, Robots, Please Report To The Dancefloor

61

Re: empirical fact: good mp3 sounds as good as wav.

Larry Levan wrote:

Frequencies that are unable to be heard by the human ear may also interact with audible frequencies to affect the sound.

yes I thought so, interesting larry

At end, it is all up to your ears and whether you think it's worth it when you listen on your equipment.

that's why I listen to (at least) 160khz mp3's at home
but playing stuff out is different, sound seems to lack power in mp3's format (although I can only guess, never really tried it myself, but I noticed it with dj's playing dubious cdr's)
maybe's got something to do with the interaction between audible and non audible frequencies?

62 (edited by Robotworks 2008-08-09 21:25:57)

Re: empirical fact: good mp3 sounds as good as wav.

Harmonics that disappear or are changed in high frequencies alterate the precision of TONE. It's the texture of sound that become poorer. All frequencies in the graphics I posted belong to the audible spectrum. But human ears are "educated" to give more importance to frequencies that allow voice and language comprehension (between 1.5 kHz and 7.5 kHz). Frequencies above belong to the tone subtility, but it's what make the difference...


P.S. For all graphics the original wav file was 44.1 kHz/ 16 bit

Le mieux est l'ennemi du bien

Re: empirical fact: good mp3 sounds as good as wav.

Robotworks wrote:

Human ear has a 20 Hz to 20 KHz frequency response, up to 22 KHz in some cases.

Isn't that when you are born as a baby? And then it goes down quickly like when you are 20 its more 20Hz to 16KHz and 30 15KHz.

I think in the future people want to have 96kbps MP3's for the vintage sound (and will use it to master it like that) just as some of us like the sound of old cassette tapes (basicly it all comes to no high frequencies then I guess)

Also in theory a track made with an Amiga and buzzing hi frequency monitor hum over it will sound better in 128kbp MP3 then the wav file hehe

Re: empirical fact: good mp3 sounds as good as wav.

Smackos has a very good point. Sometimes I find that a track 192 sounds more together than the wav smile ... and also 7" have lesser bass presence ... right?

Re: empirical fact: good mp3 sounds as good as wav.

Squadra Smackos wrote:

Isn't that when you are born as a baby? And then it goes down quickly like when you are 20 its more 20Hz to 16KHz and 30 15KHz.

yeah and isn't it ironic that those old fogeys spend tons on audiophile equipment

Re: empirical fact: good mp3 sounds as good as wav.

Marco Tulio Thrash wrote:

Smackos has a very good point. Sometimes I find that a track 192 sounds more together than the wav smile ... and also 7" have lesser bass presence ... right?

Probably cause youre ears are ruined like half of the kids these days , listening to music on there cell phones. One thing is low-fi recording to 4 , 8 track another is crap sounding 192 kps mp3 that are that have that shinny over compressed sound youve come to love.

Re: empirical fact: good mp3 sounds as good as wav.

@robotworks, what settings did you use? you can always turn off lowpass filtering if you want those high freqs, but i don't think it would contribute to a better sound. most music only has noise in those freqs anyway, especially vinyl rips.

Monkey see, monkey do.

Re: empirical fact: good mp3 sounds as good as wav.

change topic title to :empirical fact: good mp3 sounds as good as wav to 100 year old.

Showing the way for the modern man to become a Model Man.

Re: empirical fact: good mp3 sounds as good as wav.

320k still has the full hearing range (upto 20khz). the extra 2050Hz on a cd are there to ease the filtering specs.

Monkey see, monkey do.

Re: empirical fact: good mp3 sounds as good as wav.

For the mp3 I used LAME with the following configuration:

-Constant bitrate (CBR)
-encoding="Normal"
-freq="44100"
-channels="stereo"
-copyright_bit="0"
-original_bit="0"
-crc="0"
-force_iso="0"

Distortion in high frequencies on vinyl makes the sound more "dirty" but don't remove anything to its richness and "energy".

Regarding the compact Cassettes, it depends on the recoder/player, I read that best units have the wide range (20 Hz to 20 kHz)

Le mieux est l'ennemi du bien

Re: empirical fact: good mp3 sounds as good as wav.

Forax wrote:
Squadra Smackos wrote:

Isn't that when you are born as a baby? And then it goes down quickly like when you are 20 its more 20Hz to 16KHz and 30 15KHz.

yeah and isn't it ironic that those old fogeys spend tons on audiophile equipment

Some of them have babies. And hearing music from early on creates a different mindset to sound. wink

ppl may think i act like a pig, but in the end i look cute !

72 (edited by Robotworks 2008-08-10 19:21:58)

Re: empirical fact: good mp3 sounds as good as wav.

Here is a wav file with frequencies that are only between 15 kHz and 20 kHz. Each sine sound last 1 second and is separated from the other by 1 second silence:
1) 15 kHz sine during 1 sec
2) 1 sec silence
3) 16 kHz sine
4) 1 sec silence
5) 17 kHz... etc
Next it's 18 kHz and 20 kHz

http://www.fileden.com/files/2006/11/14 … ifreq2.wav

rather use good headphones to listen to it. Increase the volume if you hear nothing.


Edit: I changed the file, with the previous distortion at high volume was creating new frequencies

Le mieux est l'ennemi du bien

Re: empirical fact: good mp3 sounds as good as wav.

Why use mp3 when you can use flac wink

freakazoids, Robots, Please Report To The Dancefloor

Re: empirical fact: good mp3 sounds as good as wav.

Larry Levan wrote:

Why use mp3 when you can use flac wink

Can't use it on mp3-players...

75 (edited by Robotworks 2008-08-10 23:05:41)

Re: empirical fact: good mp3 sounds as good as wav.

VLC can play FLAC. You can do the conversion for portable players, if needed. But OGG seems pretty good too... With the AoTuv codec anyway... and even if it also has to be verified with other kind of sound...

Globally it seems that Open-Source codecs are (or have become) better. But they cannot be sold, so industrials are not interested.

Le mieux est l'ennemi du bien