I ripped a short clip from the album The Rocky Story. It's the first nine seconds from the song Gonna Fly Now. You can download it here: http://www.mtek.chalmers.se/~ekdafred/samp...nna_fly_now.wav (1,55 MB)

This is probably a known problem but I'll post it anyway:
I have encoded the files with PsyTEL AACEnc 2.01

When encoding the clip with any MPEG-4 profile (-pns switch) some very noticeable artifacts/distortion are introduced at all bit rates (My English isn't good enough to describe the sound better than that). I do not hear any of this with MPEG-2 profiles.

Some of the parameters I used are:
-br 512 -pns
-vr -br 512 -pns
-vbrhi -br 512 -pns
-ultra -pns

And also with -profile 1 or -profile 2 added.

Is this something you have noticed also?

Hi, i thought pns was implemented to help out at low bit rates rather than the higher ones you are using, theres a thread recently bout pns that you might wann read, bout pns and winamp,

http://www.hydrogenaudio.org/forums/showth...hp?threadid=638

it seems unwareented to use it in your situation and not using it may help your situation quite abit,

i may be wrong, and if i am, could anyone correct me,

Thanks
Ami

I think *lol* is appropriate for this one... Encoding your sample with PsyTEL AAC 2.02 using profile 1 and the Freeware AAC decoder plugin 01/05/2002, caused Winamp (2.77) to crash horribly, but just before the crash you could hear four loud clicking noises... That´s one major bug... Either in PsyTEL or in in_aac.dll... :/

Profile 1 seems to have the same time bug as I also tend to get with some other tracks using profile 0... It reports the track length as 0 second... My guess is that it´s a plugin error, but who knows?

Neither profile 0 or 2 caused Winamp to crash though...

And like ami said, you can really notice that it should only be used on lower bitrates... -pns does sound better at lower bitrates, whereas it sounds terrible at higher bitrates... Even I could hear it with a pair of crappy $20 Sony headphones and a poorly grounded SB Live! with lots of static background noise...

pns + high bitrate = bad idea

Yup, I understand that. But if I for some weird reason would like to have an MPEG-4 file should I use something like this (to stop pns from destroying the sound):

-pns -ptt 1 -psf 15000

?

hi,
Ok, from what i understand, you want an aac file that has pns enabled, but you want to limit it, so that it aint gonna distort your music as much.

From the command line, i get, that you want pns starting frequency at 15,000khz rather than the deafult of 4khz, for some reason i think this may help your situation a bit, but i still think its alot better to avoid pns althoughter as from other threads, it is apparent that the decoding and implentation of pns is far from optimal,

only real way is if you encode various samples and compare them

hope you find what you trying for

ami, I want the opposite; pns disabled but MPEG-4 profile..
I thought the -pns switch was necessary to get any MPEG-4 profiles, but after playing around a bit I realised that using -profile 2 (Long Term Prediction) without -pns seems to work, but how is it possible (if it is) to get Low Complexity and Main MPEG-4 files without using -pns?

The reason I'm asking this is just because of curiousity. I'm quite pleased with MPEG-2 LC anyway.

hehe, its ok, its cool you trying experimenting with aac, its impressive how much work has gone into the decoder, and how far its come relatively short time,

im sorry i cant help you, bout your issue, as ive only played with the encoder for bout 30mins,
hopefully someone more knowledgeble can answer your enquiry, maybe ivan?

thx
ami

PNS was only just introduced into PSYtel and it is for low bitrates only. All the Mpeg4 profiles work fine without it.

john33

Ok, so what do i write to get MPEG-4 LC and main?

-profile 0 gives MPEG-2 LC
-profile 1 gives MPEG-2 Main
-profile 2 gives MPEG-4 LTP

That is the only profile options I'm aware of.

Adding -pns to profile 0 and 1 makes it MPEG-4, but PNS isn't always wanted..

I don't get why you want a MPEG-4 stream.

A MPEG2 LC stream is identical to a MPEG4 LC (no PNS) stream. The only difference is 2 bits in the ADTS header, acording to Ivan.

The sound quality is identical too.

And some hardware players that can play MPEG2 AAC can't play MPEG4. So, it's better to use MPEG2 always that you can.

profile 2 generates a MPEG4 stream because LTP is part of MPEG4 audio specification, and doesn't exist in MPEG2.

Let me point you to this thread in the audiocoding.com forum:
http://www.audiocoding.com/phorum/read.php?f=1&i=960&t=949

Regards;

Roberto.

Yes, I know but I really don't _want_ MPEG-4. I was just playing around with the encoder and wondering why there isn't an option to have the MPEG-4 header or if I had missed it (for my frontend). Thanks for the clarification anyway.