Lame, --nssafejoint and --nspsytune

[seewen]

I wanted to convert a AC3 5.1 into DPL2 MP3 with BeeSweet. And I wanted to know if I could use better than my usual "--preset standard".

I searched the forum, and found this in the MP3 FAQ (#9): http://forum.doom9.org/showthread.php?&postid=75778#post75778

Someone suggest using: joint stereo + "--nssafejoint" and "--nspsytune". But the message is from january 2002.

And when I look in the lame 3.96 helpfile, I could not find --nssafejoint nor --nspsytune.


Are those switch still needed with --preset standard/medium/insane, for better DPL2 compatible result (BeeSweetGUI 0.7b6 won't add those switch with --preset) ?

[KpeX]

They shouldn't be needed for DPL2 compatibility, but your best bet is to test it on your equipment compared to a WAV/PCM DPL2 source.

The history of the lame switches and their inclusion in the --presets is convoluted at best. The presets were designed to have the best of nssafejoint, nspsytune and other obscure switches built in. However the switches included in the presets have changed several times, and newer generations of lame are bringing new settings as well with the new -V VBR presets. If you start searching on hydrogenaudio and google you'll find enough information and debate on obscure lame switches it'll make your head spin.

That being said, in my personal experience, --alt-presets are excellent for DPL(II) encodings.

[seewen]

I followed your advices, and I searched in Hydrogenaudio for lame/Surround.

And I found that:

1) --preset-standard and/or --preset-extreme are best suited for Surround I/II.
2) Some 5.1 receiver owners add that --nssafejoint (replaced by "--nsmsfix 1.0" in recent lame builds) switch should be added in order to remove/hide the noise/glitches from rear speakers.


I usually use --preset-standard/extreme (lame 3.96.1) too. And it works well. But I can hear the glitches/noize on rear speaker during quiet movie part (and/or if you put your hear near the rear speaker).

With "--nsmsfix 1.0" (which must be added AFTER --preset), you get more "ms" and less "ss" than without "--nsmsfix 1.0" (according to encspot, with "Demolition Man").

I'm goind to try if I can really hear a difference.


P.S.
it seems that --nspsytune is already used by --preset-standard/extreme.