FAST AAC LC Enc. v2.0 (build Mar 4 2002), Should I use it or not? Options

[budgie]

t hearWell, from the early morning I've been browsing the net to get a "clear picture" about the AAC encoding, but I've ended up in a total mess... One of my friends wants all his collection have converted into AAC -archive. It is approx. 1400 CDs, so it's more than clear he doesn't want to make a mistake. We don't want to use Nero for encoding, definitely. And as fastenc (FAST MPEG-2 AAC LC Encoder v2.0 build Mar 4 2002) is a lot more quick than aacenc (MPEG-4 AAC Encoder v2.15 build Mar 2 2002) we want use the fast one. But we found kind of a bug in winamp 2.81 (I guess it's not a in_aac.dll error but an encoder-related error). Sometimes it shows length 0 but during replay it shows the time correctly...
Besides this I found some dlls on the net (libfac, libfad, libfad2...)...

So my questions are:

1. Is it safe enough to use the fastenc? If it's free of bugs except the afore mentioned, I could live with it. What about the quality? As it's the approx. same date as aacenc v2.15, I think maybe we could use it, because from some listening tests on the same samples we didn't hear anything wrong...
2. What about ADTS header? Shall we disable it in the command line? We would like to have the compatibility with MP4 in the (near) future, as we believe it becomes world widespread format...
3. What encoders, decoders, plugins for winamp and so on to use?

Many thanks in advance for everybody, who would try help us!

[hans-jürgen]

1. Is it safe enough to use the fastenc? If it's free of bugs except the afore mentioned, I could live with it. What about the quality? As it's the approx. same date as aacenc v2.15, I think maybe we could use it, because from some listening tests on the same samples we didn't hear anything wrong...

I would guess that FastEnc hasn't been used that much as AACEnc or AACENC_MPEG4, and furthermore it's declared as a beta version by Ivan, as far as I know. So there might be some hidden problems, but as long as no one tries this program, they will remain hidden... I also tested it yesterday because of Speek's GUI that can use it besides the other two PsyTEL codecs and found it to be twice as fast as AACEnc on my old AMD 5x86/133 MHz. A small "bug" appeared when I tried to use my favorite -resample option that obviously is not known to FastEnc. But it was very interesting to see five different profiles shown in the DOS box while encoding, one of them being Scalable Sample Rate (SSR). B) So hopefully Ivan will read this and tell us if this profile is "for real" or only an error.

2. What about ADTS header? Shall we disable it in the command line? We would like to have the compatibility with MP4 in the (near) future, as we believe it becomes world widespread format...

Don't ever disable the headers in an AAC file, because you will of course get seeking errors then... So was this your idea and your file that "dony" on the audiocoding.com forum has troubles with now or are you "dony" himself maybe? If you want to convert AAC files to MP4, use Speek's GUI and mp4creator for many files and maybe Winamp's in_mp4.dll for a single file.

3. What encoders, decoders, plugins for winamp and so on to use?

Throw away in_aac.dll, it's too old and outdated by now, and download the newest plugins from John33's mirror of RareWares. If you need a seeking bar with MPEG-2 AAC files in Winamp, choose in_faad.dll, if not (because you want to convert all AAC files to MP4), install in_mp4.dll that can also seek through MP4 files, but not through AAC files.

You already wrote that you and your friend did not find any sound problems with FastEnc, so you can probably stick to it. Did you read Ivan's PDF file about this codec and the inherent differences and speed optimizations compared to AACEnc? He also mentioned that there shouldn't be any problems normally...

[budgie]

hans-jürgen:

Thanx a lot. Now it's clear that we will use AAC/MP4 -archive as a format for future (although we'd rather stick with MPC -Braindead...) We've been testing fastenc since last Saturday and it sounds very, but very acceptable on all the songs/compositions we've tried... We tried samples a complicated excerpts from classic music (symphonic, chamber, arias), metal, death metal, hard rock, techno, some kind electronic music, pop, new age, world music, nu jazz, acoustic and electric jazz... It just works fine. Great work, really, from the author(s) of the format!

[hans-jürgen]

Thanx a lot. Now it's clear that we will use AAC/MP4 -archive as a format for future (although we'd rather stick with MPC -Braindead...) We've been testing fastenc since last Saturday and it sounds very, but very acceptable on all the songs/compositions we've tried...

I still don't understand why hardware support is important (assuming you're not considering MPC anymore because of that) when using such a high bitrate, because the files sizes will be beyond anything usable on a portable for example. In other words, if hardware support for a format is important, I would not encode all my CDs with -archive, but rather with -extreme or even -normal, so that I can easily stuff some of them onto the portable player, be it a Philips Expanium or a Flash ROM/memory card based one without another re-encoding lossy process.

We tried samples a complicated excerpts from classic music (symphonic, chamber, arias), metal, death metal, hard rock, techno, some kind electronic music, pop, new age, world music, nu jazz, acoustic and electric jazz... It just works fine. Great work, really, from the author(s) of the format!

Good to know... now do it all again with -extreme, please! If you can find a problem sample with that setting, I really would like to hear it (and I don't mean harpsichord music...)

[budgie]

hans-jürgen:

Hardware support... I personally think that MP4 would be important in a near future and then you may find it everywhere, just as today you can find MP3. We don't assume portables, we are not teenagers anymore and we have not much time to walk around with headphones... And what's the reason for using -archive? Just for the feeling of "being secure"... some kind of redundancy can do no harm... WAV is also a sort of overkill, if you take human hearing into account... And for the same reason I use MPC -Braindead and would use lame -insane if there would be no AAC or MPC... Because you can deceive everybody but not your own brain; and my brain just tells me that only these settings are safe enough. But I am a bit "distorted" after all these years in music recording business...

P.S. And the filesize really doesn't matter when you take into account how the capacity of HD grows and prices go down, let alone the prices of CD-Rs... Moreover, in the case you would desperately need to make CDA from your archive, then you have almost no problem using MPC -Braindead to wav (I always made two-pass decoding and I let the program tell me the correct scalefactor when clipping occured...) I made some (just for testing purposes) and gave them some of my friends. I didn't tell 'em what was going on behind and everybody was very satisfied, indeed...

I must a bit disappoint you, but I don't plan any encoding neither with -extreme nor with -normal settings. The two reasons are:
1. the encoding process is really slow and as I see it now, we will do most encodings during the night using batch process from Ivan&Menno v1.6.2
2. I can't understand it, but it is even more slow, when you use -normal and -extreme settings. Both differ very slightly in bitrate (and the same goes for quality, I think...). Maybe it's due to the fact that the encoder must compute the more optimized variant for the given bitrate... when encoding into -archive it has a bit more "free hands"...

So was this your idea and your file that "dony" on the audiocoding.com forum has troubles with now or are you "dony" himself maybe?

Have no idea what are you talkin' about...

This post has been edited by budgie: Jan 16 2003, 16:13

[hans-jürgen]

And what's the reason for using -archive? Just for the feeling of "being secure"... some kind of redundancy can do no harm...
[...]
P.S. And the filesize really doesn't matter when you take into account how the capacity of HD grows and prices go down, let alone the prices of CD-Rs...

Then lossless formats would also be an option... but in the meantime I should have learned to never ever argue with a "peace of mind" follower...

I must a bit disappoint you, but I don't plan any encoding neither with -extreme nor with -normal settings. The two reasons are:
1. the encoding process is really slow and as I see it now, we will do most encodings during the night using batch process from Ivan&Menno v1.6.2

This was meant as a joke, not as a serious suggestion...

2. I can't understand it, but it is even more slow, when you use -normal and -extreme settings. Both differ very slightly in bitrate (and the same goes for quality, I think...). Maybe it's due to the fact that the encoder must compute the more optimized variant for the given bitrate... when encoding into -archive it has a bit more "free hands"...

That's probably right, but Ivan could surely give a more precise explanation, I guess.

So was this your idea and your file that "dony" on the audiocoding.com forum has troubles with now or are you "dony" himself maybe?

Have no idea what are you talkin' about...

OK, forget about it, I meant another thread in the Audiocoding.com forum ("Channel coupling not supported...") where this guy had sent me a non-playable AAC file with no ADTS or ADIF headers, just the raw AAC stream.

Have fun!

[n68]

yup...


@budgie...


found this.. wandering about..

Psytel AAC CLI Encoder [16th December 2002] - A straight replacement for aacenc
(untested for me though.. it says "straight replacement ".. but its only at db site..?)

adress: http://www.dbpoweramp.net/codecs/dMC-AAC-E...Encoder-CLI.exe

[Ivan Dimkovic]

This AAC encoder is old 2.15 from April 2002 with changed date in the binary - current AAC core does not use Intel SPL (ia32math.dll) anymore.

[n68]

yup...


hmmm...
(TnX. for the info)

[budgie]

hans-jürgen:

Did you read Ivan's PDF file about this codec and the inherent differences and speed optimizations compared to AACEnc? He also mentioned that there shouldn't be any problems normally...

Can you provide me a link to it? I'd be glad...

Then lossless formats would also be an option... but in the meantime I should have learned to never ever argue with a "peace of mind" follower...

Lossless "kommt eigentlich gar nicht in Frage", just because it would be perhaps never widely accepted format. And I am not interested in compression about (max.) 1:2, it has no sense then... but the quality of AAC/MP4 is good enough to be accepted by me as casual listening (and by casual I don't mean sitting in my armchair in front of my "vicious" expensive equipment...) As for the "piece of mind follower", I miss the point... Does it mean it has no sense argue with me? That's not nice... (w00t) :'(

This post has been edited by budgie: Jan 17 2003, 09:37

[hans-jürgen]

Did you read Ivan's PDF file about this codec and the inherent differences and speed optimizations compared to AACEnc? He also mentioned that there shouldn't be any problems normally...

Can you provide me a link to it? I'd be glad...

Hmm, let's see... the file name is "di090201.pdf" (Fast implementation of AAC LC Encoder) and Roberto once said he would place all those document files concerning PsyTEL on his FTP server, so you can probably find it with the help of Google (because I don't have the address of Roberto's server). If not, I could also email it to you (178 KB).

As for the "piece of mind follower", I miss the point...  :( Does it mean it has no sense argue with me?  :unsure: That's not nice...  (w00t)  :'(

I thought I had used enough smilies behind that remark the last time. I'm a fan of very low bitrates, because I want to save as much space as possible for my audio files in order to make them usable for a specific purpose, e.g. for homepage up- and downloads. In my opinion that's the only valid reason to compress music at all, because just for listening to it I will always prefer the original tapes or CDs or whatever and don't bother to duplicate it in any other format. But please don't let us start this discussion again, I know there are other opinions...