MP3 Pro codec demo by Thomson to be released tomorrow!

[xortam]

And how about comparisons with 64 and 128 kbps encodings? Aren't these lower rates more frequently used for CDRs? What is the consensus for the lowest rate at which you won't notice a degradation in an automotive environment?

[az]

I think at normal 64 Kbit you would notice artifacts even in a car - 128 K is fine for cars, maybe 64 K Pro would be fine too

And space is not a consideration unless you use pc-cards/sd cards/memory sticks etc. which you really shouldnt do anyway, since mp3 cd-players and mp3 players with built-in harddisks are getting cheaper and cheaper.

AZ

[This message has been edited by az (edited 16 June 2001).]

[xortam]

But memory is better for "vibratory" environments ... such as a car. Though I believe they're still talking about producing large disk based players for automotive use. There's been some talk among the BMW enthusiasts to convince BMW to supply an MP3 solution (or at least provide for an add-on).

[impact]

I have to agree with Rags this time (suprise suprise ), the Frauenhofer VBR encoding is excellent! But that is not the point of this thread.

I will compare the 64kbit mp3, mp3PRO and 128kbit mp3 and will tell you the results soon.

[impact]

And when I say compare I really mean compare

Original track - frequency analysis:



64kbit encoded in mp3PRO played with regular mp3 decoder


64kbit encoded in mp3PRO played back in mp3PRO (BIG difference)


128kbit encoded with regular mp3 played back with regular mp3

[Indiana]

Xortam, those mp3-CD players normally have some RAM integrated for stutter-free playback - the Expanium e.g. has a 100sec shock-buffer which is working VERY well - I'm using this device mostly in my car and didn't encounter a stutter up to now. And you have longer playback-time even with 160Kbps normal mp3 than you would with any of the current RAM based devices with 64Kbps mp3Pro.

For the quality-tests I use a hifi system (not really high-end, but still quite good, mainly consisting of an Onkyo reciever and some nice big Elac speakers). And, btw, I use the Philips Expanium to play back the mp3s, as the SBLives sound quality sucks bigtime, there so much noise even over the digital-out(!!!) when there's actually no sound played at all...

I can't really discern MP3 and CD at 192Kbps and up with both Fraunhofer and LAME (you have to take a newer build of the latter, of course), I'd even call 160Kbps to be roughly CD-quality with those two encoders in most cases. Only with very few tracks the trebles are distorted (try to encode Desireless Voyage, Voyage - mp3-encoders seem to HATE this song, maybe they're right...)



[This message has been edited by Indiana (edited 16 June 2001).]

[Indiana]

Impact, those graphs are quite interesting. So mp3Pro when played back on a normal mp3-decoder seems to cut the frequency band at 8KHz?!? That's just inacceptable.
And you are not using the HQ switch for encoding with your regular mp3-encoder (which one was it), since there the trebles over 16KHz are cut? Or are you using the Xing encoder (Urgh!)?

[Liquid Snake]

I think that most encoders cut off freq. above 16K when you encode at 128kbps.

[Wombat]

Indiana,
How could there possibly be noise on the digital out?

[DentyCracker]

I think I must have been on crack when I said it sounded great on wmp7 and mantis techdemo. Just listened to it again on those players and it sux big time. Sounds great on the mp3pro decoder. Lame is now very fast, it used to be sloooow. It is my preferred encoder. Its vbr is excellent, though some people don't like vbr.

[az]

AFAIK, MP3 PRO is fully backwards compatible to MP3, so it'll sound like a standard mp3 on any non-mp3 PRO enabled device, and only sound better on special MP3 PRO players.

AZ

[impact]

mp3Pro when played back on a normal mp3-decoder seems to cut the frequency band at 8KHz

Yes, but bear in mind it's only encoded at 64kbps... anything encoded at 64kbps with a mp3 codec (non-pro) will sound really bad.

As for the regular codec, I used the default settings of the fraunhofer codec (fast codec(hq)).

[This message has been edited by impact (edited 17 June 2001).]

[xortam]

<font face="Verdana, Arial, Helvetica" size="2">Originally posted by Indiana:
Xortam, those mp3-CD players normally have some RAM integrated for stutter-free playback ...</font>

I was comparing memory based players versus HDD based players. I'm sure that CD based MP3 players will do fairly will based on my experience with buffering and suspension systems in normal CD players. How well will a HDD based MP3 player perform in a car with a stiff suspension? Sounds like a head-crash waiting to happen.

AFA MP3 fidelity ... It doesn't look like it would be very musical on a high end system. I'll have to do some listening but it still seems to me to be a format only suitable for less ambitious HiFi equipment and/or noisy environments.

[Indiana]

<font face="Verdana, Arial, Helvetica" size="2">Originally posted by xortam:
AFA MP3 fidelity ... It doesn't look like it would be very musical on a high end system. I'll have to do some listening but it still seems to me to be a format only suitable for less ambitious HiFi equipment and/or noisy environments.
</font>

You definitely have to try MP3s at 192 or 256Kbps (with a good encoder), I'd be very surprised if you really could hear the difference, no matter what equipment used.

As AZ said in an earlier post there was a comparision between 256Kbps mp3 and CD done by c't-magazine on highest-end equipment and with testers from respected audio-magazine as well as studio producers, and even they couldn't spot the difference in the majority of cases.

The problem with people thinking mp3 is bad are those files you can dl from the internet. Those are most often of terrible quality because of bad encoders, too low bitrates, and sometimes even bad soucre materials (hiss, knacks that are not mp3-artifacts).

[xortam]

I'll certainly give high rate MP3 encodings a try some time. Impact ... any chance at seeing your charts with 192 and 256 Kbps encodings?