I've stumbled upon Lancer's compiles based on aoTuV Vorbis, and I like the speed improvment, but are there any quality degradations?
I understand that it was written in the Assembler, but has anyone done a reliable bit-comparation? I've done a few which have shown the audio is bit-identical, but since I don't have enough free time nowdays I am somewhat paranoid.
In other words, are the two Vorbis files, one encoded with Lancer's compile, other with the standard aoTuV always bit-identical?
Thanks.
edit: Typos
I've stumbled upon Lancer's compiles based on aoTuV Vorbis, and I like the speed improvment, but are there any quality degradations?
I understand that it was written in the Assembler, but has anyone done a reliable bit-comparation? I've done a few which have shown the audio is bit-identical, but since I don't have enough free time nowdays I am somewhat paranoid.
In other words, are the two Vrobis files, one encoded with Lancer's compile, other with the standard aoTuV always bit-identical?
Thanks.
[a href="index.php?act=findpost&pid=364698"][{POST_SNAPBACK}][/a]
All bit comparisons I made between Lancers Q6 and the standard compile show literally hundreds of differences, but I am led to believe this should be non-audible.
This issue hasn't been clarified by anyone properly yet though. Additional input from someone in the know about lancer would be welcomed
The SSE and SSE2 optimisations are written in assembly language. I think it's safe to say that bit comparisons between standard aoTuV and Lancer encoded files will ALWAYS differ, but, it's also true to say that as yet no one has been able to demonstrate any audible differences.
I'd suggest you trust your own ears on this one.
I believe there's no quality differences (I recall some thread about this before...), but because of the changes made in Lancer, the audio is not bit identical.
If you were -really- concerned about it, you could always do some ABXing and see if there's any measurable difference
The difference is due:
1. Different headers
2. Roundation errors.
As we all know the first doesn't affect quality.
The second. Someone here made a test you can try yourself. Just create wave difference between 'generic' and Lancer's compiles in any sound editor. There should be some clicks at >3 kHz. As I remember they were not loud. And I doubt anyone can hear them.
Personally - I use Lancer's compile.
Thanks everyone for the replies!
I'll be testing a little more on myself, until I make a proper choice.
(and as for those few bit-comparations... I guess I seriously sKrewed something up.)
Generic being aoTuV, de Mon?
And yeah, I use Lancer's compile too, I can't hear any quality difference, and the speed is just too good to not use it.
Generic being aoTuV, de Mon?
[a href="index.php?act=findpost&pid=364920"][{POST_SNAPBACK}][/a]
I mean original aoTuV compile.
Every Ogg Vorbis encoder will be affected by precision of Floating-point arithmetic.
aoTuV reference encoder is internally processing in 80bit FPU.
The other optimized compile(Lancer,Rarewares...etc) has 32bit precision(SSE).
So each encoder would not output bit-identical result.
But I never heard it cause audible difference.
details in Lancer coder FAQ (http://translate.google.com/translate?u=http%3A%2F%2Fhomepage3.nifty.com%2Fblacksword%2Freadme_j.htm&langpair=ja%7Cen&hl=ja&newwindow=1&ie=UTF-8&oe=UTF-8&prev=%2Flanguage_tools)