Talk:LAME - Revision history

95.90.251.128: /* Links pointing to actual sourcecode. */ new section

2017-07-24T16:42:40Z

Links pointing to actual sourcecode.: new section

← Older revision		Revision as of 16:42, 24 July 2017
Line 35:		Line 35:
	:The "4.0" name was chosen just to keep it from being confused with the trunk's official 3.93 alphas, and to reflect the fact that the branch is a testing ground for ideas of optimizations and architectural changes that ''may'' go into LAME 4.0 whenever they decide to start work on it in the trunk.		:The "4.0" name was chosen just to keep it from being confused with the trunk's official 3.93 alphas, and to reflect the fact that the branch is a testing ground for ideas of optimizations and architectural changes that ''may'' go into LAME 4.0 whenever they decide to start work on it in the trunk.
	:I attempted to add this info to the History section of the article. Hopefully it's phrased well enough. —[[User:Mjb\|Mjb]] ([[User talk:Mjb\|talk]]) 15:04, 12 October 2014 (BST)		:I attempted to add this info to the History section of the article. Hopefully it's phrased well enough. —[[User:Mjb\|Mjb]] ([[User talk:Mjb\|talk]]) 15:04, 12 October 2014 (BST)

			== Links pointing to actual sourcecode. ==

			Whenever I want to check out the source of LAME 3.100, I find myself looking for the code forever, because it is not at all evident where it is located.

			When following the link on the page "to the sourcecode", I end up on this page: https://sourceforge.net/projects/lame/files/lame/

			However what I actually am looking for, is the current development tree that I can check out.
			Of course it doesn't help, that SourceForge is a horrible site to navigate, but I think it would be quite a service, having clear linking to sources of versions beyond 3.99 as of right now. I.e. the beta versions of 3.100 and the current development tree.

			It is actually much easier to find compiled versions, because they respond to search tokens, like "LAME 3.100".

Mjb: /* LAME 4.0a14 */ mystery solved. the "4.0 alpha" builds that surfaced in 2003-2005 are unofficial, alternate versions of 3.92, for testing ideas that may end up in 4.0

2014-10-12T14:04:30Z

LAME 4.0a14: mystery solved. the "4.0 alpha" builds that surfaced in 2003-2005 are unofficial, alternate versions of 3.92, for testing ideas that may end up in 4.0

← Older revision		Revision as of 14:04, 12 October 2014
Line 30:		Line 30:
	There are versions of LAME numbered beyond 3.100 that have been published: most notably, 4.0a6 and 4.0a14. 4.0a14 apparently was released on June 28, 2005, so it is actually quite old - NOT a newer version than 3.100.		There are versions of LAME numbered beyond 3.100 that have been published: most notably, 4.0a6 and 4.0a14. 4.0a14 apparently was released on June 28, 2005, so it is actually quite old - NOT a newer version than 3.100.
	It might be worth noting this on the LAME page so that no one gets confused. [[Special:Contributions/98.240.142.9\|98.240.142.9]] 05:25, 8 October 2014 (BST)		It might be worth noting this on the LAME page so that no one gets confused. [[Special:Contributions/98.240.142.9\|98.240.142.9]] 05:25, 8 October 2014 (BST)
			:Thanks for pointing this out. I looked into it by digging through old forum posts and looking at the code in CVS.
			:The "4.0 alpha" builds that surfaced in 2003-2005 are based on one developer's experimental code branch that started in 2002. He was very unhappy that builds of this code were being distributed online by random people who thought it was official LAME 4.0 code.
			:If you were to build a working copy of LAME with code from his branch, as some people apparently did, the binary will indeed say it is LAME 4.0 alpha (4.0a1 through 4.0a14, depending on what snapshot you use). However, this doesn't mean it ''is'' the official LAME 4.0 code. It's just a semi-private, alternate version of 3.92, and is one of many such branches in the CVS repository. Official releases are only snapshots of the trunk (the "MAIN" branch).
			:The "4.0" name was chosen just to keep it from being confused with the trunk's official 3.93 alphas, and to reflect the fact that the branch is a testing ground for ideas of optimizations and architectural changes that ''may'' go into LAME 4.0 whenever they decide to start work on it in the trunk.
			:I attempted to add this info to the History section of the article. Hopefully it's phrased well enough. —[[User:Mjb\|Mjb]] ([[User talk:Mjb\|talk]]) 15:04, 12 October 2014 (BST)

98.240.142.9: /* LAME 4.0a14 */ new section

2014-10-08T04:25:38Z

LAME 4.0a14: new section

← Older revision		Revision as of 04:25, 8 October 2014
Line 25:		Line 25:

	This issue was resolved in October 2006. Since then, the wiki only recommends and links to the builds at Rarewares. —[[User:Mjb\|Mjb]] ([[User talk:Mjb\|talk]]) 00:21, 11 September 2014 (BST)		This issue was resolved in October 2006. Since then, the wiki only recommends and links to the builds at Rarewares. —[[User:Mjb\|Mjb]] ([[User talk:Mjb\|talk]]) 00:21, 11 September 2014 (BST)

			== LAME 4.0a14 ==

			There are versions of LAME numbered beyond 3.100 that have been published: most notably, 4.0a6 and 4.0a14. 4.0a14 apparently was released on June 28, 2005, so it is actually quite old - NOT a newer version than 3.100.
			It might be worth noting this on the LAME page so that no one gets confused. [[Special:Contributions/98.240.142.9\|98.240.142.9]] 05:25, 8 October 2014 (BST)

Mjb: /* Forum discussions (old) */ more links, better titles

2014-09-10T23:46:00Z

Forum discussions (old): more links, better titles

← Older revision		Revision as of 23:46, 10 September 2014
Line 5:		Line 5:

	Here are links to some of the old recommendations and other posts of historical interest:		Here are links to some of the old recommendations and other posts of historical interest:
	* [http://www.hydrogenaud.io/forums/index.php?showtopic=32288&st=0&p=327262&#entry327262 Recommended settings for ~~Lame version~~ 3.90.x] ~~(includes acknowledgements)~~		* [http://www.hydrogenaud.io/forums/index.php?showtopic=32288&st=0&p=327262&#entry327262 Recommended settings for LAME 3.90.x]
	* [http://www.hydrogenaudio.org/forums/index.php?act=ST&f=15&t=203 ~~Forum~~ discussion of ~~the~~ recommended settings ~~(older)~~]		* [http://www.hydrogenaudio.org/forums/index.php?act=ST&f=15&t=203 Discussion of the recommended settings for LAME 3.89 to 3.97]
	* [http://www.hydrogenaud.io/forums/index.php?showtopic=18091 VBR -V n settings and presets]		* [http://www.hydrogenaud.io/forums/index.php?act=ST&f=16&t=32288 Further discussion of recommended settings for LAME 3.95 to 3.97]
			* [http://www.hydrogenaud.io/forums/index.php?showtopic=18091 VBR -V n settings and presets for LAME 3.95.1]
			* [http://www.hydrogenaud.io/forums/index.php?showtopic=90464 Discussion that led to removing a "quality vs. file size" graph]

	==Page discussion==		==Page discussion==

Mjb: /* LAME compiles Win32 */ added a note that this was resolved long ago

2014-09-10T23:21:38Z

LAME compiles Win32: added a note that this was resolved long ago

← Older revision		Revision as of 23:21, 10 September 2014
Line 21:		Line 21:
	:Agree. If noone oposes we should make the change.--[[User:Beto\|Beto]] <small>'''([[User_talk:Beto\|talk]] \| [[Special:Contributions/Beto\|contribs]])'''</small> 15:48, 27 September 2006 (CDT)		:Agree. If noone oposes we should make the change.--[[User:Beto\|Beto]] <small>'''([[User_talk:Beto\|talk]] \| [[Special:Contributions/Beto\|contribs]])'''</small> 15:48, 27 September 2006 (CDT)
	:PLUS, LAME binaries are illegal in many countries. It's easy to use google. {{User:Elliottmobile/sig2}} 02:10, 28 September 2006 (CDT)		:PLUS, LAME binaries are illegal in many countries. It's easy to use google. {{User:Elliottmobile/sig2}} 02:10, 28 September 2006 (CDT)

			This issue was resolved in October 2006. Since then, the wiki only recommends and links to the builds at Rarewares. —[[User:Mjb\|Mjb]] ([[User talk:Mjb\|talk]]) 00:21, 11 September 2014 (BST)

Mjb: Acknowledgements really only apply to the old info; historical recommendations moved here from external links section

2014-09-10T23:07:28Z

Acknowledgements really only apply to the old info; historical recommendations moved here from external links section

← Older revision		Revision as of 23:07, 10 September 2014
Line 1:		Line 1:
	{{talkheader}}		{{talkheader}}

	==~~Acknowledgements~~ (old)==		==Forum discussions (old)==
	~~The following apply mostly to previous versions~~ of the ~~LAME article and therefore may refer to outdated features~~, ~~such as~~ the ~~famous --alt-presets~~, ~~whose improvements have long since been folded into the main routines and default settings~~ of LAME.		This article began as a "List of recommended settings" post in the HA forum. Rather than updating the posts there, the information is now only updated here in the wiki. To reduce confusion, recommendations applicable to old versions of LAME are not kept in the wiki.

	~~;People who took part in suggesting~~ the ~~different settings~~:		Here are links to some of the old recommendations and other posts of historical interest:
	:~~Dibrom, r3mix, ff123, Hans Heijden, kjempen, Benjamin Lebsanft, GeSomeone, Wombat~~ & ~~GuruBoolez~~ for ~~his immense testing~~.		* [http://www.hydrogenaud.io/forums/index.php?showtopic=32288&st=0&p=327262&#entry327262 Recommended settings for Lame version 3.90.x] (includes acknowledgements)
			* [http://www.hydrogenaudio.org/forums/index.php?act=ST&f=15&t=203 Forum discussion of the recommended settings (older)]
	~~;Creation of the alt preset system and related special code level quality enhancements:~~		* [http://www.hydrogenaud.io/forums/index.php?showtopic=18091 VBR -V n settings and presets]
	:Dibrom, with technical assistance from Robert Hegemann and Naoki Shibata, Gabriel and extensive tuning help and quality verification via listening tests from JohnV and also initial help (--dm-preset era) from Hans Heijden, ff123, Wombat, and others. Test clips, bitrate information, and further listening tests provided by TheBashar, zbutsam, Pio2001, BadDuDeX, r3mix, h, TarX, Hans Heijden, ff123, Wombat, Filburt, Volcano, Garf, MrDrew, TrNSZ, nyaochi, Amadeus93, in no particular order, and many, many others we (Dibrom, user) probably forgot to mention.

	~~;Idea~~ (~~also exposing the need for a unified preset system~~)~~, Original post and list of original settings collected by:~~
	~~:user~~

	~~;Layout and additional work by~~:
	~~:dev0, CiTay, SNYder, Dibrom~~.

	~~And finally~~...

	~~Thank you ALL in the community for making it what it is, providing interest and discussion and helping to work towards the most concise, well tuned,~~ and ~~most thought out MP3 quality "paradigm" seen yet! —Dibrom~~

	==Page discussion==		==Page discussion==

Mjb: moved acknowledgements to here; removed my old post as I'm reformulating my thoughts on the matter; removed random recommendation

2014-09-10T18:39:32Z

moved acknowledgements to here; removed my old post as I'm reformulating my thoughts on the matter; removed random recommendation

@@ Line 1: / Line 1: @@
 {{talkheader}}
 ==Page discussion==
@@ Line 13: / Line 32: @@
 :Agree. If noone oposes we should make the change.--[[User:Beto|Beto]] <small>'''([[User_talk:Beto|talk]] | [[Special:Contributions/Beto|contribs]])'''</small> 15:48, 27 September 2006 (CDT)
 :PLUS, LAME binaries are illegal in many countries. It's easy to use google. {{User:Elliottmobile/sig2}} 02:10, 28 September 2006 (CDT)

79.133.131.255 at 20:11, 10 May 2013

2013-05-10T20:11:13Z

← Older revision		Revision as of 20:11, 10 May 2013
Line 21:		Line 21:

	== My usual (best) settings: ==		== My usual (best) settings: ==
			lame version 3.93.1 - only this version work ok for -k.
	lame.exe -q0 -k -v -V3 -b32 -B256 my.wav		lame.exe -q0 -k --vbr-new -v -V3 -b32 -B256 my.wav

79.133.131.255: /* My usual (best) settings: */ new section

2013-05-10T20:06:43Z

My usual (best) settings:: new section

← Older revision		Revision as of 20:06, 10 May 2013
Line 19:		Line 19:

	There is also no mention of the tuning that has been done for LAME's presets and default settings at certain bitrates. Much attention has been given to optimizing VBR, ABR, and the common (128 kbps and up) bitrates, but not so much for the CBR lower bitrates. For example, as I mentioned at the bottom of the [[CBR]] article, a certain other codec's default setting for 96 kbps CBR stereo output uses about an 11.5 kHz lowpass filter on the input, as compared to LAME which defaults to about a 12.5 kHz filter. LAME allows for more high end in the audio, but encodes it with more noise (bleepy artifacts), too. For some, it's an even trade, and LAME's definition of "quality" is equivalent to that of the other encoder. But for others, one encoder or the other produces qualitatively inferior output. In other words, each encoder has its own definition of ''measurable'' quality, and its own standard of quality for its presets, and each listener has his/her own subjective tolerances and criteria for ''perceived'' quality. So this is why I'd like to see an explanation of exactly which factors are used in LAME's quality settings. -- [[User:Mjb\|Mjb]] 04:10, 19 January 2008 (CST)		There is also no mention of the tuning that has been done for LAME's presets and default settings at certain bitrates. Much attention has been given to optimizing VBR, ABR, and the common (128 kbps and up) bitrates, but not so much for the CBR lower bitrates. For example, as I mentioned at the bottom of the [[CBR]] article, a certain other codec's default setting for 96 kbps CBR stereo output uses about an 11.5 kHz lowpass filter on the input, as compared to LAME which defaults to about a 12.5 kHz filter. LAME allows for more high end in the audio, but encodes it with more noise (bleepy artifacts), too. For some, it's an even trade, and LAME's definition of "quality" is equivalent to that of the other encoder. But for others, one encoder or the other produces qualitatively inferior output. In other words, each encoder has its own definition of ''measurable'' quality, and its own standard of quality for its presets, and each listener has his/her own subjective tolerances and criteria for ''perceived'' quality. So this is why I'd like to see an explanation of exactly which factors are used in LAME's quality settings. -- [[User:Mjb\|Mjb]] 04:10, 19 January 2008 (CST)

			== My usual (best) settings: ==

			lame.exe -q0 -k -v -V3 -b32 -B256 my.wav

Mjb: /* Define quality */ minor edits to previous

2008-01-19T10:19:30Z

Define quality: minor edits to previous

← Older revision		Revision as of 10:19, 19 January 2008
Line 18:		Line 18:
	Much is made of LAME's "quality" settings, but this term is not really defined. I believe the main definition of quality is inverse of quantization noise (more noise = lower quality). Are there other factors? And it'd be worth remarking on how higher quality is often achieved: by using higher bitrates, using different psychoacoustic models, or simply chopping off the high end.		Much is made of LAME's "quality" settings, but this term is not really defined. I believe the main definition of quality is inverse of quantization noise (more noise = lower quality). Are there other factors? And it'd be worth remarking on how higher quality is often achieved: by using higher bitrates, using different psychoacoustic models, or simply chopping off the high end.

	There is also no mention of the tuning that has been done for LAME's presets and default settings at certain bitrates. Much attention has been given to optimizing VBR, ABR, and the common (128 kbps and up) bitrates, but not so much for the CBR lower bitrates. For example, a certain other codec's default setting for 96 kbps CBR stereo output uses about an 11 kHz lowpass filter on the input, as compared to LAME which defaults to about a 12.5 kHz filter. LAME allows for more high end in the audio, but encodes it with more noise (bleepy artifacts), too. For some, it's an even trade, and LAME's definition of "quality" is equivalent to that of the other encoder. But for others, one encoder or the other produces qualitatively inferior output. ~~Each~~ encoder has its own definition of quality and ~~standards~~ of ~~measurable~~ quality for its presets, and each listener has his/her own subjective tolerances and criteria for perceived quality. So this is why I'd like to see an explanation of factors are used in LAME's quality settings. -- [[User:Mjb\|Mjb]] 04:10, 19 January 2008 (CST)		There is also no mention of the tuning that has been done for LAME's presets and default settings at certain bitrates. Much attention has been given to optimizing VBR, ABR, and the common (128 kbps and up) bitrates, but not so much for the CBR lower bitrates. For example, as I mentioned at the bottom of the [[CBR]] article, a certain other codec's default setting for 96 kbps CBR stereo output uses about an 11.5 kHz lowpass filter on the input, as compared to LAME which defaults to about a 12.5 kHz filter. LAME allows for more high end in the audio, but encodes it with more noise (bleepy artifacts), too. For some, it's an even trade, and LAME's definition of "quality" is equivalent to that of the other encoder. But for others, one encoder or the other produces qualitatively inferior output. In other words, each encoder has its own definition of ''measurable'' quality, and its own standard of quality for its presets, and each listener has his/her own subjective tolerances and criteria for ''perceived'' quality. So this is why I'd like to see an explanation of exactly which factors are used in LAME's quality settings. -- [[User:Mjb\|Mjb]] 04:10, 19 January 2008 (CST)