LAME: different compilers producing significantly different bitstreams, Split from “LAME 3.99 is out” |
LAME: different compilers producing significantly different bitstreams, Split from “LAME 3.99 is out” |
Jan 28 2012, 15:16
Post
#1
|
|
![]() Group: Members Posts: 1303 Joined: 14-September 05 From: Helsinki, Finland Member No.: 24472 |
The new 32-bit lame.exe seems to work fine for me on XP Pro SP3 (I have an AMD Phenom II CPU that supports SSE2), but it produces entirely different output than LAME 3.99.3 from the previous "main" bundle. I did not expect that. As far as I can see, the change log does not indicate any further "quality tweaking".
In some cases 3.99.4 seems to produce significantly smaller VBR bitrates than 3.99.3. I tested the settings -V1, -V5 and -V8. When the source tracks were quiet and uncompressed (classical, ambient, etc) the bit rates were 3-12% smaller than before. When the source tracks were loud and compressed (hard rock, metal, etc) the bitrates were about the same as before (or just very slightly smaller). This post has been edited by Alex B: Jan 28 2012, 16:08 -------------------- http://listening-tests.freetzi.com
|
|
|
|
![]() |
Jan 29 2012, 12:00
Post
#2
|
|
![]() Group: Developer Posts: 3035 Joined: 2-December 07 Member No.: 49183 |
It seems that the relevant changes are somewhere in quantize.c: http://lame.cvs.sourceforge.net/viewvc/lam...athrev=lame3_99
*pxmin++ = xmin; -> *pxmin++ = Max(xmin, 1e-20f); etc. |
|
|
|
Jan 29 2012, 23:02
Post
#3
|
|
|
Group: Members Posts: 319 Joined: 26-November 04 Member No.: 18345 |
It seems that the relevant changes are somewhere in quantize.c: I ran a test similar as that run by Alex B, except with a sampling of files from my FLAC library, and only encoded at -V2, and found similar results. Unlike Alex, I did find numerous cases where the RareWares 3.99.4 produced higher bitrate (larger) files than 3.99.3. What I don't understand about this explanation is that in my tests your compiles of 3.99.3 and 3.99.4 produced essentially the same bitrate in the output, unlike the RareWares compiles of the same releases. And in nearly all cases the bitrates were lower than those produced by either RareWares version. Some examples (all at -V2) CODE Katia Lebeque - Rhapsody in Blue.mp3
RareWares LAME 3.99.4 196 kbps RareWares LAME 3.99.3 190 kbps lvcql's LAME 3.99.4 178 kbps lvcql's LAME 3.99.3 178 kbps The Black Keys - Set You Free.mp3 RareWares LAME 3.99.4 186 kbps RareWares LAME 3.99.3 202 kbps lvcql's LAME 3.99.4 185 kbps lvcql's LAME 3.99.3 185 kbps Yngwie Malmsteen - Leviathan.mp3 RareWares LAME 3.99.4 189 kbps RareWares LAME 3.99.3 187 kbps lvcql's LAME 3.99.4 185 kbps lvcql's LAME 3.99.3 185 kbps |
|
|
|
Alex B LAME: different compilers producing significantly different bitstreams Jan 28 2012, 15:16
halb27 With my test set 3.99.4 produces an average bitrat... Jan 28 2012, 22:44
Alex B I uploaded a sample: http://www.hydrogenaudio.org/... Jan 28 2012, 23:31
JJZolx QUOTE (Alex B @ Jan 28 2012, 15:31) In so... Jan 29 2012, 00:18
Alex B QUOTE (JJZolx @ Jan 29 2012, 01:18) Some ... Jan 29 2012, 00:43
lvqcl Different compilers/options IMHO. Jan 28 2012, 23:58
Alex B At least John did not change the announced compile... Jan 29 2012, 00:12
lvqcl From RareWares main page: "Compiles are now M... Jan 29 2012, 00:23
bbrabant QUOTE (lvqcl @ Jan 29 2012, 02:23) Also, ... Jan 29 2012, 00:40
JJZolx Ok, then back to lvqcl: Why would a different comp... Jan 29 2012, 00:53
Alex B QUOTE (JJZolx @ Jan 29 2012, 01:53) Why w... Jan 29 2012, 01:06
C.R.Helmrich Indeed. How was the following done, if I may ask?
... Jan 29 2012, 02:10
john33 Sorry, guys, there was a compiler option error on ... Jan 30 2012, 12:42
halb27 Everything's fine now. Exactly the same averag... Jan 30 2012, 14:54
john33 QUOTE (halb27 @ Jan 30 2012, 13:54) Every... Jan 30 2012, 15:11
godrick Does the nature of the compiler error require user... Jan 30 2012, 15:56
john33 QUOTE (godrick @ Jan 30 2012, 14:56) Does... Jan 30 2012, 16:01
Alex B Thanks John.
The new 32-bit exe compile seems to ... Jan 30 2012, 17:16
john33 QUOTE (Alex B @ Jan 30 2012, 16:16) Thank... Jan 30 2012, 18:01
robert QUOTE (Alex B @ Jan 30 2012, 17:16) Thank... Jan 30 2012, 18:16
Alex B I don't think libsndfile is involved. I used f... Jan 30 2012, 18:36
mixminus1 I can confirm AlexB's findings vis a vis the b... Jan 30 2012, 21:49
DigitalMan Assume 64 bit compile is unaffected? Not at machi... Jan 31 2012, 00:17
mixminus1 No, latest 64-bit version of 3.99.4 (20120125) fro... Jan 31 2012, 00:47
john33 QUOTE (mixminus1 @ Jan 30 2012, 23:47) No... Jan 31 2012, 09:35
john33 OK, done some testing comparing the Rarewares 32 b... Jan 31 2012, 17:43
Pretentiousman Is it possible for an executable from one compiler... Jan 31 2012, 17:58
john33 QUOTE (Pretentiousman @ Jan 31 2012, 16:5... Jan 31 2012, 18:12
JJZolx QUOTE (Pretentiousman @ Jan 31 2012, 09:5... Jan 31 2012, 18:12
lvqcl QUOTE (JJZolx @ Jan 31 2012, 21:12) Perha... Jan 31 2012, 18:38
apodtele QUOTE (lvqcl @ Jan 31 2012, 12:38) QUOTE ... Jan 31 2012, 21:52
C.R.Helmrich QUOTE (apodtele @ Jan 31 2012, 22:52) Ple... Jan 31 2012, 23:01
apodtele Order is defined by precedence and compiler rarely... Jan 31 2012, 23:36
saratoga QUOTE (apodtele @ Jan 31 2012, 15:52) QUO... Feb 1 2012, 00:44
nu774 CODEmain()
{
volatile double x = 53, y = ... Feb 1 2012, 02:53
apodtele QUOTE (nu774 @ Jan 31 2012, 20:53) 54 52 ... Feb 1 2012, 04:23
nu774 QUOTE (apodtele @ Feb 1 2012, 12:23) You ... Feb 1 2012, 04:45

nu774 QUOTE (nu774 @ Feb 1 2012, 12:45) QUOTE (... Feb 1 2012, 11:42
timcupery QUOTE (apodtele @ Jan 31 2012, 22:23) We ... Feb 2 2012, 03:56
lvqcl 2 seconds of Liszt.wav (LAME -V1):
original track... Jan 31 2012, 18:05
JJZolx But to what degree in terms of overall bitrate? Th... Jan 31 2012, 18:45
Alex B Here's how EncSpot Pro sees the differences in... Jan 31 2012, 18:58
JJZolx Here's some more data at different VBR quality... Feb 1 2012, 00:21
lvqcl Test signal: sine (-12 dBFS) + white noise (-N dBF... Feb 2 2012, 16:56
halb27 Looks like your MSVC10 compile behaves more reason... Feb 2 2012, 17:46
lvqcl Spectrograms of sine+noise (3 seconds from an mp3 ... Feb 2 2012, 18:13
apodtele QUOTE (lvqcl @ Feb 2 2012, 12:13) Spectro... Feb 2 2012, 19:53
halb27 Looks like it's the decision process for decid... Feb 2 2012, 20:27
john33 Just for the record, Robert is looking at this iss... Feb 2 2012, 20:35
lvqcl QUOTE (john33 @ Feb 2 2012, 23:35) Just f... Feb 2 2012, 21:10
john33 QUOTE (lvqcl @ Feb 2 2012, 20:10) ...
It... Feb 2 2012, 23:09
Destroid From the sidelines, I found this discussion of the... Feb 5 2012, 11:49
antman Agreed Destroid. This has been an interesting rea... Feb 7 2012, 04:08
JJZolx I'm surprised this thread wasn't split a l... Feb 7 2012, 04:28
DARcode The "minutiae" in this thread are actual... Feb 7 2012, 14:15
john33 Robert is currently working on a solution that wou... Feb 7 2012, 14:27
apodtele QUOTE (john33 @ Feb 7 2012, 08:27) Robert... Feb 7 2012, 16:24
john33 QUOTE (apodtele @ Feb 7 2012, 15:24) This... Feb 7 2012, 17:21
robert QUOTE (apodtele @ Feb 7 2012, 16:24) This... Feb 7 2012, 17:38
apodtele QUOTE (robert @ Feb 7 2012, 11:38) The co... Feb 7 2012, 21:48
john33 There is a test compile (32 bit) now available at:... Feb 7 2012, 17:13
JJZolx Here's that set of files again, encoded using ... Feb 8 2012, 00:24
halb27 I checked Alex B's Liszt sample, and 3.99.5 be... Feb 9 2012, 10:56
Kilu This thread was an interesting read.
So just to b... Feb 10 2012, 17:46
john33 QUOTE (Kilu @ Feb 10 2012, 16:46) This th... Feb 10 2012, 18:27
antman Kilu, in summary, the encoding speed differences w... Feb 11 2012, 02:37
Kilu Alrighty, thanks guys.
I think I'll wait for t... Feb 11 2012, 03:25
john33 3.99.5 test 2 on the link below fixes the decoding... Feb 13 2012, 10:21
psycho 3.99.5 is out, compiles anyone? Feb 15 2012, 21:54
john33 QUOTE (psycho @ Feb 15 2012, 20:54) 3.99.... Feb 15 2012, 23:28
lvqcl There's still no mention of 3.99.5 at http://s... Feb 15 2012, 22:06![]() ![]() |
|
Lo-Fi Version | Time is now: 19th June 2013 - 09:06 |