How about using AAC for Music Collection

[Ramirez]

@Rjamorim:

Roberto calm down lol,it wasn't my intention at all to disrespect your work, (I wasn't very clear) as I said I wasn't tried to conduct some "scientific" or even partly serious test, all I meant is that I like to base my opinion on thing based on my own tests and I prefer to use my own ears\eyes for quality evaluation.

my best regards :)

Edit: LOL read above i said "plz advice, I'll give it a go."


LOL! I'm so tired...:D

[JohnV]

Originally posted by Ramirez
So far I've tried AACmachine (psytel)/ Nero AAC plugin and QT6 AAC Export plugin; still in all cases Vorbis produced better sound Quality then corresponding AAC File, and believe me, it was lot more then just "one track"

If you're thinking that I should change something in my setup or use a diff encoder then plz advice, I'll give it a go. Well, one thing is for sure. If you claim that "in all cases" Vorbis produces better results, something is either wrong in your encoding/system, or then you are affected by placebo and I recommend blind testing for you.

My opinion is that Vorbis is not better, rather infact a bit worse codec than QT 6.3 or Nero AAC/-he. I don't know how much testing experience you have, but some here might know something about my history.. :D
I'd say I'm also pretty familiar with Vorbis, since I was feeding results for Monty already over a year ago when he finalized 1.0

[kl33per]

JohnV knows what he's talking about and I agree with him, Vorbis is becoming outdated. However, I think both codecs have a lot of potential still, plus Vorbis isn't locked into any kind of official standard, so it can continue to get better all the time. AAC on the other hand is part of the MPEG-4 standard of A/V codecs and can only have so much done to it before breaking the standard. Still, Vorbis 1.1 is set for release (sometime in the distant future) and Ivan is always working on Nero AAC so things could get mighty interesting.

[mrlipring]

lame mp3 @ 192kbps q0 etc...

My archos jukebox only plays mp3.

Using a higher bitrate means that i'd either have to buy more hard drive space (well, i DO need to, but yeah...) or have less music. no thanks :)

it sounds good. it sounds excellent in fact. no denying that ogg etc are better, but no use to me, unfortunately. I'd love to get a portable player that used a programmable processor to decode music, as opposed to damned codec chips.

[mrlipring]

Originally posted by JohnV
Well, one thing is for sure. If you claim that "in all cases" Vorbis produces better results, something is either wrong in your encoding/system, or then you are affected by placebo and I recommend blind testing for you.

where did he say he was testing it with Placebo tracks?












:D

[kl33per]

Placebo is not a type of track. It's a psychological effect on human beings. Basically it means the results from a test can be easily skewed by our personal preference towards codecs, i.e. a person who likes and uses Vorbis will most likely see Vorbis win in any comparison they do. The only way to get around this is to do a blind test, i.e. you don't know which enocde corresponds with which codec.

[mrlipring]

Oh...


My...


God...


Placebo are a band. It was a joke. Hence the :D A rubbish joke, but a joke nonetheless. Thanks for assuming i'm stupid and explaining the word "placebo" to me though.

You can't make stuff like this up.

[kl33per]

Sorry, I wasn't implying you were stupid. I've never heard of any band called Placebo, but there's probably lots of bands I haven't heard of. And I didn't see your smiley face at the bottom of your post. Plus, a lot of people on doom9 don't understand the 'placebo' effect or 'blind-testing'. Anyway, didn't mean to offend you.

[Mug Funky]

@ kl33per:

placebo were back in the days i thought audiocatalyst and xing were cool...

hmm. profile says you're from brisbane. surprised you've not heard of them, it was all they played in melbourne for about a year on some stations.

just to make my post constructive...

i've played with psytel, but got frustrated at its (IMHO) slightly immature psychoacoustics. no matter how hard i try, things like castanets or some explosions with a strong mid contour (like in a lot of older movies) come out sounding peculiar. for now i'm still using LAME with my own absurdly long command line.

i think the psy model in psytel's aacenc is too closely based on the tunings from the old mp3 psy. i don't think they've hit the ideal model for AAC's different coding type (i may be wrong, but doesn't AAC use DWT instead of DCT?).

i haven't used vorbis very much at all (please, don't hit me, ramirez... i feel like i'm admitting atheism to a baptist), but a while ago i did a few low bitrate tests against lame (96kbit average) and seemed to get more stable and predictable results with lame, but admittedly virbis performed better in certain situations (it worked best on rock songs that were hard limited and not very spatial - dynamic and spatial songs came out with severe MS artefacts).

i am no scientist, so please, any new users: take this as my 2 cents, and not a comprehensive analysis :P

[rjamorim]

Originally posted by Mug Funky
i think the psy model in psytel's aacenc is too closely based on the tunings from the old mp3 psy. i don't think they've hit the ideal model for AAC's different coding type (i may be wrong, but doesn't AAC use DWT instead of DCT?).

Well, Psytel has not been developed in an year and half now. It's sucessor is the Nero AAC plugin - that is being constantly developed.

And no, there are no wavelets in AAC. (At least, not in the LC profile. Maybe some weird and rare profile uses it)

[Mug Funky]

tanks for the response... i'd always wondered aboot DWT - there's been some fuss about it over the last couple of years, but it hasn't really been applied terribly well yet (actually, the rududu video codec is showing some early promise).

yeah, i'm aware of psytel's ceased development. i just can't afford nero 6 atm :(. i'd like to try out it's SBR capabilities, because frankly LC AAC hasn't quite proved itself.

from what i've read about SBR, it should be nearly perfect for classical music, with its predictable upper harmonics.

[kl33per]

Originally posted by Mug Funky
placebo were back in the days i thought audiocatalyst and xing were cool...
I was uneducated then. I missed the whole xing encoder stage.

Like I said, I think AAC (and Vorbis) still have a lot of potential in the future. I just submitted my results for Roberto's test and I must say HE AAC disapointed me. It was very impressive at low bitrates, but wasn't far ahead of Vorbis.

[rjamorim]

Hello.

The test is now finished. Here are the results:

http://audio.ciara.us/test/64test/results.html

[bond]

now, thats the reason why nobody votes for wma ;)

[frodoontop]

I'd like to vote for Musepack(MPC), it's transparant for almost every file at 160-170 kbit/s. This is at q5.

It's the only lossy codec which is designed to be transparent, that I'm aware of. Better not use Musepack for low bitrates.

I'm still waiting for the moment when we can finally mux video with mpc sound. For movies the bitrate is around 140kbit/s, which is reasonable for excellent sound.

[Tuning]

Originally posted by rjamorim
Hello.

The test is now finished. Here are the results:

http://audio.ciara.us/test/64test/results.html

Quote from above link:
Another important note: Keep in mind that Lame MP3 is at 128kbps in this test. There are two reasons for that:
1-To put results into perpective. In this aspect, Lame was a high anchor (and FhG MP3 the low anchor)
2-To test if some of these codecs deliver the marketing plot of "Same quality as MP3 at half the bitrates", specially concerning WMA and MP3pro.



What do u really mean by this.Is Lame encoded file is at 128kbps ABR.
And i have a suggestion: 64kbps is very low bitrate and can u conduct another open test with 128 and 96kbps streams seperately.Then we could find which is better suited for music archieval.:) .Thanks:p

[frodoontop]

Have a look here: http://audio.ciara.us/test/

Rjamorim did arrange another public test @128kbit/s.

[Tuning]

I wasn't aware that the test was already conducted.The test was not done@96kbps CBR.So may i suggest u to conduct once more with all encodes at 96kbpsCBR,Which is the usual encoding bitrate that is preffered for music.

[sillKotscha]

Originally posted by Tuning
So may i suggest u to conduct once more with all encodes at 96kbpsCBR,Which is the usual encoding bitrate that is preffered for music.

:confused: :D

[Teegedeck]

Originally posted by Teegedeck

The only statement that I would consider safe without testing is that MP3pro is worse that both Ogg and AAC.
It seems, I shouldn't have said even that without a blind-test - sorry. (Vorbis scored a little worse than MP3pro in the test.)

Edit: Besides eyeing the overall results, I find it always useful to read about the results of the man that I'd say earns the title 'most educated ears' on hydrogenaudio. I hope you'll forgive me for quoting his comment from the thread on HA:

Originally posted by guruboolez
I obtained different results, with bad notation for vorbis (unfortunately, I forgot the matrix on another computer). I'm not at ease with vorbis at this bitrate during a blind test : it sounds too particular (hiss, desquilibrated tonal range : more treble, poor low-medium, and limited stereo), and it's easy for me to detect the encoder. I'm rating vorbis, and not an unknow encoder. So it isn't blind anymore.

HE-AAC was my favorite : often the best - never the worst. Sometimes betrayed by a grainy texture, the same as mp3pro one. No noise packets, as heard with the first releases of the encoder.

Lower anchor was rarely the worse file I rated : on 8 files, I rated other encodings as worst one. I prefer an excessive lowpassed sound without artifacts than a richer sound, but destroyed by flanging. Personnal taste.

WMA9 (I hated this encoder) was as often the best file than HE-AAC. But it was three time the worse for me. So it isn't a reliable encoder, but on some situation, it works very well.

LC-AAC was first on two sample (02 and 09), last on one (11).. Vorbis best on one (04), and last on two (06 and 09).

With these results (talking about overall results and guruboolez') it seems to narrow down to the choice of HE AAC or MP3pro for extremely low bitrate encodings. Still not the thing to go for when you seriously back up. But for movies I'd feel tempted to do another blind-test of my own in order to find out whether I hear a difference between LC and HE AAC for a relevant portion of a movie.

[Tuning]

what i mean is that not to test quality @ 1/2 bitrate of LAME file.Test which codec is more near in quality to source @96kbps CBR(As i think 64kbps is lower than what required (exception:mp3PRO) so that a audio cd compatible WAV can be created without much loss in quality).if possible compare between AAC,Vorbis and mpc only.

[Teegedeck]

You can always prepare samples yourself. After all, you want to decide what suits your ears best, don't you?

[SeeMoreDigital]

Well after looking at the test results Mp3pro is not so bad after all - at 64kbps anyway!

Although I have to say, the encodes I did yesterday with AAC std @ 96kbps sounded better to me than Mp3pro @ 96kbps!

Cheers

[Tuning]

Whenever i tried to comparison i found AAC @ 96kbps is better than mp3@96kbps and I'm transcoding all my music to mp4.And i think mp4 is going to be as popular as mp3.
Edit:Test complete!
Well finally i conducted my own tests(blind ones, with 1000w 4.1speaker system "ON" at optimum volume).The three contestants were AAC,Ogg vorbis and Mpc.The test file was a WAV file ripped from AudioCD and of length 5:30.I used the various plugins provided by dBpowerAMP for creating mp4 and ogg.I used MUSE drop and mppenc to create mpc.The bitrate considered was something variable between 95-105kbps(I changed my mind of testing CBR as i came to know AAC doesn't really have CBR and infact i was using VBR mp3 for DVDrips and want to check what size will these codecs produce.)Thus all was done.

The results

WAV AAC Ogg MPC
-------------------------------------------------------------

size in MB 55.5 3.72 3.67 3.96
-------------------------------------------------------------

ABR(kbps) 1411 93 94 100.7
--------------------------------------------------------------

Note :ABR is as shown in winamp2.92.


Quality:
At this bitrate mpc really found to be lacking in high frequency spectrum.I found the quality of

AAC and Ogg indistinguishable from each other.( in all range of frequencies as i used winamps

equalizer to determine each band('0dB') of spectrum.while turning other bands to '-12dB'.)


Conclusion:

As both files were approximately same in bitrate and quality,the size has to be considered.The

size of Ogg was less than AAC.Thus i think Ogg is more suitable for DVDripping.And need to wait

on using MPC and AAC.(Though we cant mux them to avi till now).


:D

[Sirber]

hum... at this time MP3/mp3PRO is far more supported by standalone players.

[SeeMoreDigital]

Originally posted by Sirber
hum... at this time MP3/mp3PRO is far more supported by standalone players. Hi Sirber,

I've been looking at specifications of some of the new portables. And there appears to be more Mp4/AAC players than Mp3pro players on the market at the moment.

In fact, I have to admit that I'm a little disappointed in the lack of support for Mp3pro. Dispite the fact that this format is over two years old, there are very few portable players and no intergration at all into video encoders!

We live in hope.... Cheers

[Sirber]

When I bought my Lyra, ~6 month ago, there were no support for AAC. Maybe my infos are outdated :D

[The Belgain]

Thanks for yet another great test ryamorim.

What are the plans for the next one? Would it be possible/feasible to do test of surround codecs at fairly low bitrate (say 192 for 5.1)? This is starting to be relevant for movies now, and there are at least a few codecs that can do it now. The samples would be mainly clips from movies that have a lot of surround info (and maybe some well-mixed 5.1 music tracks).

I suppose most of these would probably be transcoded tracks since DVD movies don't have uncompressed 5.1 tracks, but if DTS tracks are chosen, there shouldn't be too much trouble with that.

Codecs to test would be I guess AAC (Nero probably being the best for surround?), Vorbis (though it isn't tuned much), Real, wma, ac3 (as an anchor - it will sound horrible at low bitrates). Any others?

I guess one problem is there might not be that many people able to test it (people would really need 5.1 setups (or at least quadro) to test the surround effects). Anyway, what do people think about this? AFAIK there haven't been many tests done of this.

[rjamorim]

Would it be possible/feasible to do test of surround codecs at fairly low bitrate (say 192 for 5.1)? This is starting to be relevant for movies now, and there are at least a few codecs that can do it now.

Sure, it would be very interesting.

I suppose most of these would probably be transcoded tracks since DVD movies don't have uncompressed 5.1 tracks, but if DTS tracks are chosen, there shouldn't be too much trouble with that.

Actually, we should mix some DTS tracks and AC3 tracks. Remember, these tests are to reflect real-life usage, and most people still transcode from AC3 :)

And your codec choice seems very good.

Anyway, my next planned test is a voice codecs test in Novemver (i'll be travelling in october)

Then, and MP3 @ 128kbps test in December.

Regards;

Roberto.

[bond]

Originally posted by The Belgain
What are the plans for the next one? Would it be possible/feasible to do test of surround codecs at fairly low bitrate (say 192 for 5.1)? This is starting to be relevant for movies now, and there are at least a few codecs that can do it now.yeah, would be great

although i dont have a 5.1 system (till now :D )

mp3 @128kbps is also interesting...

[SeeMoreDigital]

Originally posted by Sirber
When I bought my Lyra, ~6 month ago, there were no support for AAC. Maybe my infos are outdated :D I know what you mean. Six months is a long time in software and harware terms!

I was hoping that the new Philips HDD100 'iPod imitator' would have had Mp3pro support but sadly it does'nt. It's just an Mp3 and WMA player.

There's supposed to be around a million Apple iPod owners now. I wonder how many of them store their music collections in .Mp4/AAC?
Come to think of it. I wonder what ripping application they use?

What ripping application do you use for you Thomson Lyra?

Cheers

[Sirber]

It's a USB HDD, so I can use whatever I want :) I use AltoMP3 to CD-->WAV then CoolEdit Pro (Adobe Audition) for WAV-->mp3PRO, then Eags On! to set the correct naming and ID3 tags :)

[SeeMoreDigital]

Originally posted by Sirber
It's a USB HDD, so I can use whatever I want :) I use AltoMP3 to CD-->WAV then CoolEdit Pro (Adobe Audition) for WAV-->mp3PRO, then Eags On! to set the correct naming and ID3 tags :) Blimey.... I'm worn out just reading that lot... I think I'll stick to MusicMatch for Mp3pro file creation!

Cheers

[Sirber]

MusicMatch... hihihihihi

You can have way more options with CoolEdit Pro :)

[Tuning]

I think [Edit]Nero 6 Wave Editor is better!
MusicMatch... hihihihihi
And for the continuity
hihihihihihi ...hohohohohoh.......mmmmmm......123*987*infinity

[Sirber]

Nero 6 is a burning software while Cool Edit Pro is a dedicated audio processor :)

I fellow in your path: hahahahahaha ... . ..... héhéhéhéhéhé ... .. :D

[SeeMoreDigital]

I have an original version of CoolEdit Pro. But looking at the specification for the Adobe version... it's way different!

Very nice.

Sirber, does the Adobe version allow you to generate Mp3pro encodes higher than 96kbps?

Cheers

[Sirber]

nope, only if you go for VBR it can go up to 144kbps :)

[SeeMoreDigital]

Originally posted by Sirber
nope, only if you go for VBR it can go up to 144kbps :) That's bloody good.

Can you do an experiment for me please?

Can you encode the same source file in Mp3pro @ 128kbps and then again in Mp3std @ 128kbps.

And then (this is the important bit) play back both files using a player that can't decode Mp3pro. Like say, WinMedia Player 9. And then let me know if you can hear much difference between the two!

Cheers Sirber

[Sirber]

I can't ask to encode @128, it's vbr :) I tryed with 96kbps, and you can tell the difference :) The mp3PRO miss all high-freq, while the MP3 have some.

[SeeMoreDigital]

Originally posted by Sirber
I can't ask to encode @128, it's vbr :) I tryed with 96kbps, and you can tell the difference :) The mp3PRO miss all high-freq, while the MP3 have some. Yep, I can do 96kbps myself, as that's as high as I can go! And there is a massive difference in the way the two files sound.

As you may have guessed I'm trying to discover how similar a high bitrate Mp3pro encode sounds compared to an Mp3std encode. When played back on 'non' Mp3pro players! The higher the bitrate for both formats the better!

So are you saying you can encode to Mp3pro @ 128kbps but it's VBR only. If so that should play back OK in WMP9 and so should a Mp3std VBR encode @ 128kbps. Got any tracks by Alanis Morissette?

Cheers

[Sirber]

nope, it's not my kind of music :). but any version of mp3PRO played on a non-mp3PRO player will be like a MP3@22kHz, that's the purpose of SBR, to generate the missing freq. :)

[SeeMoreDigital]

Originally posted by Sirber
..... but any version of mp3PRO played on a non-mp3PRO player will be like a MP3@22kHz, that's the purpose of SBR, to generate the missing freq. :) Yep, this is what my thinking of SBR suggests.

However because I'm only able to encode at low bitrates (upto 90kbps) I don't know if this is still the case for higher bitrates.

I've never heard Mp3pro @ 128kbps (or higher) / 22kHz!

Cheers

EDIT...
69 votes and still ZERO for WMA - Amazing!
Good job there are no portable players that only use this format!

[Sirber]

mp3PRO VBR Q100 is 96-144kbps, I have a song that is varying between 128 and 144. Q is excellent :D

[SeeMoreDigital]

Originally posted by Sirber
mp3PRO VBR Q100 is 96-144kbps, I have a song that is varying between 128 and 144. Q is excellent :D Sirber. Are you listening to it using a non Mp3pro player?

Cheers

[Sirber]

on a mp3PRO player :)

[Dr Satan]

Another vote for MPC here. I can't tell the difference from the original, and average bitrates vary between 109kbit and 197kbit (in my experience).

I think it's worked better because all the other codecs have tried to get ok quality at >128kbit, whereas MPC went for excellent quality at ~170kbit.

[duartix]

I'm just waiting for proper tagging suport to switch to AAC.
2 Months ago I was using OGG until I took the 128 kbps listening test and then changed to MPC.

[bond]

Originally posted by duartix
I'm just waiting for proper tagging suport to switch to AAC.you can tag mp4 files with the winamp plugin, foobar player... (all itunes compatible) of course

[crusty]

I've used mp3 @ r3mix values for about 200 CD's then switched over to Ogg Vorbis @ quality 6.00.

When testing for differences between the original wav file and mp3 and Ogg Vorbis I found Ogg Vorbis clearly superior.

The difference between wav and mp3@192kbps vbr is very pronounced to my ears.
With Ogg Vorbis at approximately the same bitrate I can still hear some differences, but they are small and very subtle. Without a blind test you would never notice. You would with mp3 at 192Kbps vbr.