Welcome to Doom9's Forum, THE in-place to be for everyone interested in DVD conversion. Before you start posting please read the forum rules. By posting to this forum you agree to abide by the rules. |
22nd August 2011, 12:08 | #381 | Link | ||
Software Developer
Join Date: Jun 2005
Location: Last House on Slunk Street
Posts: 13,248
|
Quote:
(As you may have noticed, I linked 'libsndfile' in a static way in order to get rid of another DLL dependency, which explains the different size of the EXE file) Quote:
__________________
Go to https://standforukraine.com/ to find legitimate Ukrainian Charities 🇺🇦✊ |
||
22nd August 2011, 18:55 | #382 | Link | |
Registered User
Join Date: Aug 2008
Location: The Land Of Dracula (Romania - EU)
Posts: 934
|
Quote:
hi, by any chance did you modify the sources to add support for stdout? (for adts CBR)? _
__________________
if you ask a question and somebody give you the correct answer don't forget to leave a "thank you" note... Visit The Land Of Dracula (Romania - EU)! |
|
22nd August 2011, 19:01 | #383 | Link | |
Software Developer
Join Date: Jun 2005
Location: Last House on Slunk Street
Posts: 13,248
|
Quote:
Output as MP4 probably requires random access to the output file, which isn't possible when writing to stdout. ADTS might work though (in theory). Also I think the CBR limitation of ADTS mode is a limitation of the encoder library, not of the CLI front-end...
__________________
Go to https://standforukraine.com/ to find legitimate Ukrainian Charities 🇺🇦✊ |
|
24th August 2011, 14:21 | #384 | Link |
Registered User
Join Date: Apr 2011
Posts: 7
|
Speed of LameXP
Hi I have an old Athlon 64 3200+ running WinXP 32bit, and find the old version of LameXP (LameXP v3.18 Hotfix-2 Build 88) to be twice as fast when encoding 320kb constant bitrate mp3s, vs the 4.02 build 578. My only complaint is that the old version keeps bugging me that the code is more than a year old and I have to update (by the way, the check for updates is disabled). If I choose cancel, the program closes, and if I choose "yes", it goes online and checks for updates. Please tell me there's a way to disable this, because it's driving me nuts. I want to stick with the old version. Thanks
|
24th August 2011, 19:51 | #385 | Link | ||||
Software Developer
Join Date: Jun 2005
Location: Last House on Slunk Street
Posts: 13,248
|
Hello, nik33134.
Quote:
Quote:
Quote:
Quote:
There are just too many nasty restrictions in the old version (no support of Unicode file names or tags !!!) and the audio tools that were used in the old version are outdated too. So would you please run the following benchmark on your system and post the results, then we might be able to find out what is going wrong: http://www.mediafire.com/file/ct5yk2...2011-08-24.rar (This includes the LAME build that was used in LameXP v3.18 Hotfix-2 Build 88, the build that is used in the current version as well as another build for comparison). Here are my results: Code:
vendor_id : GenuineIntel cpu family : 6 model : 15 stepping : 7 flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe pni monitor ds_cpl vmx est tm2 ssse3 cx16 xtpr pdcm nx lm lahf_lm cpu MHz : 2398.008 model name : Intel(R) Core(TM)2 Quad CPU @ 2.40GHz [VBR MODE] lame-3.98.4-GEN.exe 8 2.113739 0.006455 D:\SVN\Tools\LAME\benchmark\\bin\lame-3.98.4-GEN.exe -h -V2 --nohist D:\SVN\Tools\LAME\benchmark\\etc\Input.wav NUL lame-3.99.0-ICC.exe 8 2.151387 0.012290 D:\SVN\Tools\LAME\benchmark\\bin\lame-3.99.0-ICC.exe -h -V2 --nohist D:\SVN\Tools\LAME\benchmark\\etc\Input.wav NUL lame-3.99.0-MSC.exe 8 2.318193 0.011178 D:\SVN\Tools\LAME\benchmark\\bin\lame-3.99.0-MSC.exe -h -V2 --nohist D:\SVN\Tools\LAME\benchmark\\etc\Input.wav NUL [CBR MODE] lame-3.98.4-GEN.exe 8 3.770149 0.076191 D:\SVN\Tools\LAME\benchmark\\bin\lame-3.98.4-GEN.exe -h --cbr -b 320 --nohist D:\SVN\Tools\LAME\benchmark\\etc\Input.wav NUL lame-3.99.0-ICC.exe 8 3.257647 0.038184 D:\SVN\Tools\LAME\benchmark\\bin\lame-3.99.0-ICC.exe -h --cbr -b 320 --nohist D:\SVN\Tools\LAME\benchmark\\etc\Input.wav NUL lame-3.99.0-MSC.exe 8 3.556128 0.003099 D:\SVN\Tools\LAME\benchmark\\bin\lame-3.99.0-MSC.exe -h --cbr -b 320 --nohist D:\SVN\Tools\LAME\benchmark\\etc\Input.wav NUL The "old" build was slightly faster in VBR mode and significant slower in CBR mode, compared to the "up-to-date" build.
__________________
Go to https://standforukraine.com/ to find legitimate Ukrainian Charities 🇺🇦✊ Last edited by LoRd_MuldeR; 24th August 2011 at 20:23. |
||||
24th August 2011, 20:24 | #386 | Link |
Registered User
Join Date: Aug 2008
Location: The Land Of Dracula (Romania - EU)
Posts: 934
|
Here are my results:
Code:
vendor_id : AuthenticAMD cpu family : 15 model : 47 stepping : 2 flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 pni syscall nx mmxext fxsr_opt lm 3dnowext 3dnow lahf_lm cpu MHz : 2380.318 model name : AMD Athlon(tm) 64 Processor 3800+ [VBR MODE] lame-3.98.4-GEN.exe 8 2.653358 0.005877 F:\benchmark.2011-08-24\\bin\lame-3.98.4-GEN.exe -h -V2 --nohist F:\benchmark.2011-08-24\\etc\Input.wav NUL lame-3.99.0-ICC.exe 8 2.942623 0.006475 F:\benchmark.2011-08-24\\bin\lame-3.99.0-ICC.exe -h -V2 --nohist F:\benchmark.2011-08-24\\etc\Input.wav NUL lame-3.99.0-MSC.exe 8 3.315548 0.007737 F:\benchmark.2011-08-24\\bin\lame-3.99.0-MSC.exe -h -V2 --nohist F:\benchmark.2011-08-24\\etc\Input.wav NUL [CBR MODE] lame-3.98.4-GEN.exe 8 5.171427 0.012839 F:\benchmark.2011-08-24\\bin\lame-3.98.4-GEN.exe -h --cbr -b 320 --nohist F:\benchmark.2011-08-24\\etc\Input.wav NUL lame-3.99.0-ICC.exe 8 4.843703 0.006763 F:\benchmark.2011-08-24\\bin\lame-3.99.0-ICC.exe -h --cbr -b 320 --nohist F:\benchmark.2011-08-24\\etc\Input.wav NUL lame-3.99.0-MSC.exe 8 5.425726 0.004313 F:\benchmark.2011-08-24\\bin\lame-3.99.0-MSC.exe -h --cbr -b 320 --nohist F:\benchmark.2011-08-24\\etc\Input.wav NUL Wed 08/24/2011 - 22:23:11.17
__________________
if you ask a question and somebody give you the correct answer don't forget to leave a "thank you" note... Visit The Land Of Dracula (Romania - EU)! |
24th August 2011, 20:31 | #387 | Link |
Software Developer
Join Date: Jun 2005
Location: Last House on Slunk Street
Posts: 13,248
|
b66pak, interesting results. Even though you have an AMD processor, the ICL build (as used in current LameXP) still is significant faster than the MSVC build.
Also your results confirm that the "old" build was a bit faster in VBR mode (although the difference is bigger than on my system) and slower in CBR mode, compared to the "up-to-date" build. However this doesn't explain what nik33134 is reporting. He claims that the "old" build was faster in CBR mode. And he uses a very similar CPU
__________________
Go to https://standforukraine.com/ to find legitimate Ukrainian Charities 🇺🇦✊ Last edited by LoRd_MuldeR; 24th August 2011 at 22:29. |
24th August 2011, 23:08 | #388 | Link |
Registered User
Join Date: Apr 2011
Posts: 7
|
So here are the results I got:
Code:
vendor_id : AuthenticAMD cpu family : 15 model : 47 stepping : 0 flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 pni syscall nx mmxext fxsr_opt lm 3dnowext 3dnow lahf_lm cpu MHz : 1995.039 model name : AMD Athlon(tm) 64 Processor 3200+ [VBR MODE] lame-3.98.4-GEN.exe 8 3.203830 0.031366 "C:\Documents and Settings\n\Desktop\benchmark.2011-08-24\\bin\lame-3.98.4-GEN.exe" -h -V2 --nohist "C:\Documents and Settings\n\Desktop\benchmark.2011-08-24\\etc\Input.wav" NUL lame-3.99.0-ICC.exe 8 3.548249 0.112188 "C:\Documents and Settings\n\Desktop\benchmark.2011-08-24\\bin\lame-3.99.0-ICC.exe" -h -V2 --nohist "C:\Documents and Settings\n\Desktop\benchmark.2011-08-24\\etc\Input.wav" NUL lame-3.99.0-MSC.exe 8 4.010275 0.228478 "C:\Documents and Settings\n\Desktop\benchmark.2011-08-24\\bin\lame-3.99.0-MSC.exe" -h -V2 --nohist "C:\Documents and Settings\n\Desktop\benchmark.2011-08-24\\etc\Input.wav" NUL [CBR MODE] lame-3.98.4-GEN.exe 8 6.203974 0.026840 "C:\Documents and Settings\n\Desktop\benchmark.2011-08-24\\bin\lame-3.98.4-GEN.exe" -h --cbr -b 320 --nohist "C:\Documents and Settings\n\Desktop\benchmark.2011-08-24\\etc\Input.wav" NUL lame-3.99.0-ICC.exe 8 5.810394 0.007591 "C:\Documents and Settings\n\Desktop\benchmark.2011-08-24\\bin\lame-3.99.0-ICC.exe" -h --cbr -b 320 --nohist "C:\Documents and Settings\n\Desktop\benchmark.2011-08-24\\etc\Input.wav" NUL lame-3.99.0-MSC.exe 8 6.503573 0.004490 "C:\Documents and Settings\n\Desktop\benchmark.2011-08-24\\bin\lame-3.99.0-MSC.exe" -h --cbr -b 320 --nohist "C:\Documents and Settings\n\Desktop\benchmark.2011-08-24\\etc\Input.wav" NUL Thu 08/25/2011 - 1:00:20.28 Last edited by LoRd_MuldeR; 24th August 2011 at 23:17. |
24th August 2011, 23:22 | #389 | Link |
Software Developer
Join Date: Jun 2005
Location: Last House on Slunk Street
Posts: 13,248
|
This shows that the "new" version encodes a bit slower (but certainly not two times!) in VBR mode and encodes quite a bit faster in CBR mode, compared to the "old" version.
I can't see how this matches your previous report How exactly did you measure the speed of LameXP v4.03 against v3.18 on your system before? And are you absolutely sure that you got the latest build of LameXP v4.03 (build #664 at this moment) ??? Also: Is your processor single or dual core? And how many encoder instances are you running in parallel? How is the CPU usage - encoder process(es) -vs- LameXP main process - while encoding?
__________________
Go to https://standforukraine.com/ to find legitimate Ukrainian Charities 🇺🇦✊ Last edited by LoRd_MuldeR; 24th August 2011 at 23:26. |
24th August 2011, 23:53 | #390 | Link |
Registered User
Join Date: Apr 2011
Posts: 7
|
Ok, I just encoded a 7:26 flac to mp3 with the old version (constant bitrate 320, max quality, at 44100Hz, auto select channel mode) and the encoding took 1:12:34. Then I used the new version 4.02 build 578 (320 CBR, best quality, 44100Hz, auto select), and got 2:18:36. The new version created an mp3 of 17,886,730 bytes, and the old version a 17,858,745 bytes mp3 file.
I have a single core and use only one instance at a time. My observations have been empirical, and this is the first time that I used a stopwatch while encoding. But it's always been like this for me when I compared with the eye (how fast he percentage numbers change). Now either the benchmark does not fit the situation, or I don't know what... |
25th August 2011, 00:04 | #391 | Link | ||
Software Developer
Join Date: Jun 2005
Location: Last House on Slunk Street
Posts: 13,248
|
Quote:
Oh, and can you please update to the latest v4.03 build? Enable "Tools" -> "Configuration" -> "Check for Beta Updates" and then check for updates! Quote:
And your results clearly show that the difference in pure encoding speed is rather small. Also it shows that the "new" version encodes even faster in CBR mode. There are still two scenarios I can think of: (a) The encoding parameters are different between v4.03 and v3.18 in some way. If slower/faster settings were used, this would explain the difference. (b) The LameXP v4.03 front-end process uses significantly more CPU time than v3.18 on your system and this way slows down the encoder process. Can you please check both things with the help of ProcessExplorer ???
__________________
Go to https://standforukraine.com/ to find legitimate Ukrainian Charities 🇺🇦✊ Last edited by LoRd_MuldeR; 25th August 2011 at 00:25. |
||
25th August 2011, 00:29 | #392 | Link |
Registered User
Join Date: Apr 2011
Posts: 7
|
My original file is a FLAC, and the times that I state are only the encoding time, without the initial decoding time. I have not updated to the latest beta yet. I just run another test with the same file with ABR 192kbps. Results are 2:06:43 with the new stable version, and 1:16:54 with the old version. I will try to use the process explorer and see if I can see anything. My XP has 2 Gigs of memory and is pretty slimmed down (few services, and no unnecessary programs running, only Avira and firefox).
|
25th August 2011, 00:33 | #393 | Link | ||
Software Developer
Join Date: Jun 2005
Location: Last House on Slunk Street
Posts: 13,248
|
Quote:
Quote:
Also in the properties of the "lame.exe" process check the "command line" field - again for both versions. If there is a difference in the command-line arguments, it may explain the difference.
__________________
Go to https://standforukraine.com/ to find legitimate Ukrainian Charities 🇺🇦✊ Last edited by LoRd_MuldeR; 25th August 2011 at 00:36. |
||
25th August 2011, 00:58 | #394 | Link |
Registered User
Join Date: Apr 2011
Posts: 7
|
Ok, I updated to the latest beta. I run process explorer and got CPU usage at 98.44% to 100% for both the old, and the new versions (ool_lame.exe and ool_lameenc.exe). The lameXP.exe showed no CPU usage at all in either version. In the new version I had a console running that displayed "This OS doesn't support ItaskbarList3 interface".
The times for the encode this time were 1:12:39 for the old version, and 2:17:79 for the new beta, for the same 52.7MB Flac file. |
25th August 2011, 01:08 | #395 | Link | ||
Software Developer
Join Date: Jun 2005
Location: Last House on Slunk Street
Posts: 13,248
|
Quote:
Quote:
(The debug console is shown for all Beta versions) Did you check the "Command line" for 'tool_lame(enc).exe' for both versions? You find it by right-click and Properties in the ProcessExplorer.
__________________
Go to https://standforukraine.com/ to find legitimate Ukrainian Charities 🇺🇦✊ |
||
25th August 2011, 01:23 | #396 | Link |
Registered User
Join Date: Apr 2011
Posts: 7
|
Yes, the parameter for the new version is:
tool_lame.exe --nohist -q 0 --cbr -b 320 --resample 44100 --tt "Theme from 'Antarctica'" --ta Vangelis --tl "The Best Of Instrumental Works" --tg Instrumental --tc "Encoded with LameXP" --ty 2008 --tn 1 --ti For the old version, it's: tool_lameenc.exe" --nohist -q 0 -b 320 --resample 44.1 --add-id3v2 --tt "Theme from 'Antarctica'" --ta "Vangelis" --tl "The Best Of Instrumental Works" --ty 2008 --tc "Encoded with LameXP" --tn 1 --tg "Instrumental" Is the 100% CPU usage normal? |
25th August 2011, 01:42 | #397 | Link | |
Software Developer
Join Date: Jun 2005
Location: Last House on Slunk Street
Posts: 13,248
|
Quote:
I can re-produce that the "new" version indeed is noticeably slower than the "old" version when using these ultra slow placebo settings! Of course both versions are SLOW with such settings. So try to use something more sane, like "-q 2", which is equivalent to "-h". (It is the "LAME Algorithm Quality" slider that you want to adjust) Yes, normal and desirable CPU usage is the percentage of time that your CPU is working, i.e. not sleeping (waiting for I/O). When encoding, you want your CPU to sleep as few as possible, right? ^^
__________________
Go to https://standforukraine.com/ to find legitimate Ukrainian Charities 🇺🇦✊ Last edited by LoRd_MuldeR; 25th August 2011 at 01:49. |
|
25th August 2011, 01:54 | #398 | Link |
Registered User
Join Date: Apr 2011
Posts: 7
|
I see now. But how do I set the q to not be zero. Is it from the "Advanced Options" menu, the Lame Algorithm Quality? I've always set it to "best" in the past, should I set it to high quality instead? Anyways, thank you for your help.
|
25th August 2011, 02:02 | #399 | Link | ||
Software Developer
Join Date: Jun 2005
Location: Last House on Slunk Street
Posts: 13,248
|
Quote:
Quote:
So if you want to squish out more quality at the same bitrate, then you have to accept a slower encoding speed. That's live. However there always is a point where using even slower settings only gives a very minor additional improvement (if at all) for a significant additional speed cost. Obviously you would only use such "placebo" settings, if you don't care about speed at all The LAME manual says: -q 0: use slowest & best possible version of all algorithms. -q 0 and -q 1 are slow and may not produce higher quality. -q 2: recommended. Same as -h. -q 5: default value. Good speed, reasonable quality. -q 7: same as -f. Very fast, ok quality. No problem.
__________________
Go to https://standforukraine.com/ to find legitimate Ukrainian Charities 🇺🇦✊ Last edited by LoRd_MuldeR; 25th August 2011 at 02:07. |
||
27th August 2011, 19:46 | #400 | Link | |
Software Developer
Join Date: Jun 2005
Location: Last House on Slunk Street
Posts: 13,248
|
LameXP v4.03 Beta-1:
http://sourceforge.net/projects/lame...29/2011-08-27/ Quote:
__________________
Go to https://standforukraine.com/ to find legitimate Ukrainian Charities 🇺🇦✊ |
|
Tags |
aac, aotuv, flac, lame, lamexp, mp3, mp4, ogg, oggenc, opus, vorbis |
Thread Tools | Search this Thread |
Display Modes | |
|
|