--r3mix preset and Command Line for audio newbie

[RathO]

Ive always been using automate solution (like the recent quick2pass..) to encode my ac3 to mp3. But i think i now need to understand what i do with audio.

I must admit that i surely miss something important using the BeSweetGUI since theres soooo much options in there and im never sure of the final output. I read the Doom9 Guides, but im not sure he explains what i need to understand..

1-What exactly the --r3mix preset do to my audio file? (ABR? VBR? 160? 256?... im lost!) :(

2-Whats the command line in besweet if i want to encode my ac3 to mp3 with normalise, downsampling to 44.1kHz and use the --r3mix preset?

[DJ Bobo]

This is --r3mix:
--nspsytune --vbr-mtrh -V1 -mj -h -b96 --lowpass 19.5 --athtype 3 --ns-sfb21 2 -Z --scale 0.98 -X0

Change the lowpass to 18.1 if you have 384kbps 5.1 AC3 or 20.3 if you have a 448kbps 5.1 AC3.
r3mix is a pure vbr command.
you can lower -V to -V2 and -b to -b32 to lower the final average bitrate, this won't cause audible loss (at least for me)

r3mix is a very good preset, but somehow the encoded file doesn't fit well in video. I mean, it plays fine on winamp, but once you play it muxed in a video, you become distortions (especially on wide sound), at least on my win98 configuration.


SOMEONE ELSE PLEASE ANSWER HIS BESWEET QUESTION, because I don't use it.

[pacohaas]

check the audio FAQ, there's some links to recommended settings (of which r3mix is no longer)

[RathO]

Here is what i do now after lots of tests and re-reading some audio-faqs:

For a 2.0ch:

"BeSweet.exe" -core( -input "*.AC3" -output "*.mp3" ) -azid( -s surround -c light --maximize ) -ssrc( --rate 44100 ) -lame( --scale 1 --alt-preset 128 )

For a 3.2ch:

"BeSweet.exe" -core( -input "*.AC3" -output "*.mp3" ) -azid( -s surround -c light -L -3db --maximize ) -ssrc( --rate 44100 ) -lame( --scale 1 --alt-preset 128 )

These settings are almost the same as the settings included in quick2pass except i use --alt-preset 128 instead of --r3mix.

Also apply the Delay depending of the movie.
These settings are for 1 CD rip.

[DJ Bobo]

you don't really need --alt-preset if you're encoding to 128kbps.
In my opinion, it's only usefull for lower bitrates (112kbps and lower).

The main disavantadge of --alt-preset is speed, it's damn slow and the advantadge in audio quality is very marginal compared to CBR.
So you may consider encoding CBR, this will save a lot of time.

[DSPguru]

Originally posted by DJ Bobo
you don't really need --alt-preset if you're encoding to 128kbps.
In my opinion, it's only usefull for lower bitrates (112kbps and lower).just wanted to let you know that not everyone here agrees with Bobo about this.

The main disavantadge of --alt-preset is speedyou're running two-pass azid.exe,then lame.exe, and sometimes uses wavelab in between, and you wanna tell me that you care about speed :D ?

[The Link]

@ DJ Bobo

You do not post on hydrogenaudio forum, do you? For those people your proposals must sound pretty provocative. :D
BTW: What happened to hydrogenaudio.org? I couldn´t access their page whole the day. :confused:

Regards,

The Link

[DJ Bobo]

@ DSPguru
Well, just because of you (and a few others) I keep posting the higher quality methods, so you don't pick on me.
Seems like it doesn't work at all!
I think it's time to tell the truth. This is the method I use usually:
1) azid without gain
2) wavelab for downsampling (on batch mode, even if I have one single file)
I save this step if I go 160kbps mp3 or higher.
3) cbr mp3 encoding in musicmatch (nowhere under 128kbps though)
4) postgain in mp3trim
You may get scared first, because you see many steps, but believe it or not, I benchmarked things and that method was proven faster than HeadAC3he, although I counted the time when I switch between applications. Don't know how it compares to BeSweet though. May be you'll tell me?! :D

@ The Link
Well, I think I looked over there 2 or 3 times, to see what they are saying.
The truth is: it's all psychologic! how does it feel after using --alt-preset? it feels great, just because you think you improved things somewhere ... somewhere!
But the truth is, you lost a lot of time just to get minimal improvement. First, you are barely or not able to tell the difference. Second, the improvement is nowhere near to the time invested.
Why should I bother?! ;)