IPB

Welcome Guest ( Log In | Register )

> Hydrogenaudio Forum Rules

- No Warez. This includes warez links, cracks and/or requests for help in getting illegal software or copyrighted music tracks!


- No Spamming or Trolling on the boards, this includes useless posts, trying to only increase post count or trying to deliberately create a flame war.


- No Hateful or Disrespectful posts. This includes: bashing, name-calling or insults directed at a board member.


- Click here for complete Hydrogenaudio Terms of Service

4 Pages V  « < 2 3 4  
Closed TopicStart new topic
Which LAME VBR settings do you use?, Poll regarding LAME VBR preferences
LAME VBR Preferences
What VBR setting do you use?
-V0 (.XXX) (--[alt-]preset [fast] extreme) [ 125 ] ** [28.28%]
-V1 (.XXX) [ 17 ] ** [3.85%]
-V2 (.XXX) (--[alt-]preset [fast] standard) [ 171 ] ** [38.69%]
-V3 (.XXX) [ 43 ] ** [9.73%]
-V4 (.XXX) (--[alt-]preset [fast] medium) [ 23 ] ** [5.20%]
-V5 (.XXX) [ 52 ] ** [11.76%]
-V6 (.XXX) [ 7 ] ** [1.58%]
-V7 (.XXX) [ 0 ] ** [0.00%]
-V8 (.XXX) or -V9 (.XXX) [ 1 ] ** [0.23%]
Other [ 3 ] ** [0.68%]
What LAME version do you use?
3.98 (.2) [ 385 ] ** [87.10%]
3.97 [ 39 ] ** [8.82%]
3.90.3 [ 12 ] ** [2.71%]
Other [ 6 ] ** [1.36%]
If using 3.98, do you use floating point values?
Yes [ 57 ] ** [12.90%]
No [ 333 ] ** [75.34%]
I don't use 3.98 [ 52 ] ** [11.76%]
Total Votes: 603
  
Rei Murasame
post Jul 3 2009, 19:36
Post #76





Group: Members
Posts: 5
Joined: 3-July 09
From: Leicestershire
Member No.: 71177



I go with Lame 3.98, -V2. The distribution of the votes so far though, almost seems to suggest to me that V2 is migrating towards being 'standard' in name, but not actuality?
Go to the top of the page
+Quote Post
kwanbis
post Jul 3 2009, 20:22
Post #77





Group: Developer (Donating)
Posts: 2361
Joined: 28-June 02
From: Argentina
Member No.: 2425



What you mean? almost 40% use V2.


--------------------
MAREO: http://www.webearce.com.ar
Go to the top of the page
+Quote Post
Rei Murasame
post Jul 3 2009, 21:25
Post #78





Group: Members
Posts: 5
Joined: 3-July 09
From: Leicestershire
Member No.: 71177



QUOTE (kwanbis @ Jul 3 2009, 20:22) *
What you mean? almost 40% use V2.

I was looking mostly at the 25% that had shown up in V0; I had expected that there would've been like 70% on V2, and then most of the rest of it distributed somewhere below that.

Seeing these results is different than what I was anticipating. Basically it looks as if most people who didn't choose V2, tended to step up to V0, rather than step downwards. My question would be: Did those 25% of people that chose V0 do so "just to be safe", or was it because they were able to ABX artefacts with V2 that they weren't able to find in V0?
Go to the top of the page
+Quote Post
kwanbis
post Jul 3 2009, 22:21
Post #79





Group: Developer (Donating)
Posts: 2361
Joined: 28-June 02
From: Argentina
Member No.: 2425



I'm sure 95% just to be safe wink.gif

No abx to prove tongue.gif


--------------------
MAREO: http://www.webearce.com.ar
Go to the top of the page
+Quote Post
GeneV
post Jul 10 2009, 21:40
Post #80





Group: Members
Posts: 19
Joined: 3-July 09
Member No.: 71173



QUOTE (halb27 @ Jul 3 2009, 19:53) *
QUOTE (GeneV @ Jul 3 2009, 15:36) *
... I made ABX-Tests and the outcome is that ABR clearly sounds better. ...

Do you use 3.98?
I've been an ABR advocate for a long time before 3.98 came out (maybe I still am at very high bitrate ABR 250...300, but it's not very relevant to me any more).
With 3.98 however I can hardly imagine samples where ABR is clearly better than VBR in your bitrate range though ABR certainly is fine too.


OK, the tests were back in Jan.2008 and I used 3.98beta. Maybe things turn out different with 3.98.2 - I haven't tested that.
So THX for the hint, maybe I should repeat the tests based on the lastest stable version (I guess it's 3.98.2).


Go to the top of the page
+Quote Post
rpp3po
post Jul 11 2009, 05:04
Post #81





Group: Developer
Posts: 1126
Joined: 11-February 03
From: Germany
Member No.: 4961



QUOTE (Rei Murasame @ Jul 3 2009, 22:25) *
Did those 25% of people that chose V0 do so "just to be safe", or was it because they were able to ABX artefacts with V2 that they weren't able to find in V0?


Usually when LAME fails miserably at ~192kbit/s it will fail up to 320kbit/s. That's why I wouldn't go above Quicktime TVBR 127 (~192kbit/s) or Nero q 5.5. The codec either gets it (most of the time) or it fails anywhere up to 320 kbit/s CBR. This "just to be safe" nest egg is often worthless.

This post has been edited by rpp3po: Jul 11 2009, 05:04
Go to the top of the page
+Quote Post
GeneV
post Oct 11 2009, 19:02
Post #82





Group: Members
Posts: 19
Joined: 3-July 09
Member No.: 71173



QUOTE (GeneV @ Jul 10 2009, 21:40) *
QUOTE (halb27 @ Jul 3 2009, 19:53) *
QUOTE (GeneV @ Jul 3 2009, 15:36) *
... I made ABX-Tests and the outcome is that ABR clearly sounds better. ...

Do you use 3.98?
I've been an ABR advocate for a long time before 3.98 came out (maybe I still am at very high bitrate ABR 250...300, but it's not very relevant to me any more).
With 3.98 however I can hardly imagine samples where ABR is clearly better than VBR in your bitrate range though ABR certainly is fine too.


OK, the tests were back in Jan.2008 and I used 3.98beta. Maybe things turn out different with 3.98.2 - I haven't tested that.
So THX for the hint, maybe I should repeat the tests based on the lastest stable version (I guess it's 3.98.2).


Coming back to my own posting I'd like to add this:
I finally made some new ABX-tests with lame 3.98.2. The outcome again is that with the same bitrate ABR sounds slightly better than VBR. Possible explanation is: With ABR I can effectively use the -q0 switch, which is ignored(!) when using VBR (-V2 is always q=2, -V3 is always q=3 etc., even if -q0 is specified sad.gif ). Although the encoding time significantly increases with -q0 compared to -q2 or less, I don't care, because on my machine encoding is still fast enough. What counts is that I get better quality at the same bitrate resp. more qualtiy with lower filesize.

Cheers
Gene
Go to the top of the page
+Quote Post
lvqcl
post Oct 12 2009, 15:54
Post #83





Group: Developer
Posts: 3327
Joined: 2-December 07
Member No.: 49183



QUOTE (GeneV @ Oct 11 2009, 22:02) *
Possible explanation is: With ABR I can effectively use the -q0 switch, which is ignored(!) when using VBR (-V2 is always q=2, -V3 is always q=3 etc., even if -q0 is specified sad.gif )


In fact, it's the other way round: if you specify "-V2 -q2", LAME 3.98 will use "-V2 -q0" settings.
Go to the top of the page
+Quote Post
GeneV
post Oct 14 2009, 15:58
Post #84





Group: Members
Posts: 19
Joined: 3-July 09
Member No.: 71173



QUOTE (lvqcl @ Oct 12 2009, 15:54) *
QUOTE (GeneV @ Oct 11 2009, 22:02) *
Possible explanation is: With ABR I can effectively use the -q0 switch, which is ignored(!) when using VBR (-V2 is always q=2, -V3 is always q=3 etc., even if -q0 is specified sad.gif )


In fact, it's the other way round: if you specify "-V2 -q2", LAME 3.98 will use "-V2 -q0" settings.


This I can not verify. When encoding with -V2 lame always reports VBR(q=2), no matter what the switch -q specifies. I've tried this with different lame-compiles (from bakerweb and from Rarewares) and it always turns out this way. Then, which lame.exe are you using?
Go to the top of the page
+Quote Post
jmcguckin
post Oct 14 2009, 17:19
Post #85





Group: Members
Posts: 52
Joined: 2-January 08
Member No.: 50051



while I stick to AAC for my primary lossy format, I do encode to LAME mp3 from time to time, and when I do I use v. 3.99a (since after testing, I haven't been able to distinguish a difference between it and v. 3.98.2, and it encodes approximately 25-30% faster on my MacBook)... as for the quality settings, I'm a pretty die-hard -V2 fan, though I'll occasionally drop down to -V3 for stuff in my library that I won't need nearly as high a quality setting for (less complex stuff). and no, I don't use floating point (though that's only because I've never looked into it).


--------------------
Archive- FLAC (-v 8)
Portable- QuickTime AAC (True VBR/-q 77)
Go to the top of the page
+Quote Post
vvill
post Nov 12 2009, 17:08
Post #86





Group: Members
Posts: 5
Joined: 25-March 07
Member No.: 41833



I switch between v0 and v2 for music, depending on the style and how much I like the music.

Voice, books, etc. v2 at best.

I think I'm still using 3.97 *shrug*
Go to the top of the page
+Quote Post
Soviet Commissar
post Nov 12 2009, 22:28
Post #87





Group: Members
Posts: 43
Joined: 8-October 08
From: Virginia, USA
Member No.: 59765



I actually regularly use V0 and V2, one for classical or orchestral, the other for everything else, so I just put V2 since that's the one I use the most (a majority of my classical has already been ripped). I have 3.98.2, though I've been playing around with the 3.99.a1 LAME a bit. I don't use floating point, seeing as I don't really know what it is or what it does.
Go to the top of the page
+Quote Post
db1989
post May 27 2010, 13:58
Post #88





Group: Super Moderator
Posts: 5275
Joined: 23-June 06
Member No.: 32180



I don't want to bump for its own sake, and this information has been posted several times on the forum, but this exchange needs clarification:
QUOTE (GeneV @ Oct 14 2009, 15:58) *
QUOTE (lvqcl @ Oct 12 2009, 15:54) *
QUOTE (GeneV @ Oct 11 2009, 22:02) *
Possible explanation is: With ABR I can effectively use the -q0 switch, which is ignored(!) when using VBR (-V2 is always q=2, -V3 is always q=3 etc., even if -q0 is specified sad.gif )


In fact, it's the other way round: if you specify "-V2 -q2", LAME 3.98 will use "-V2 -q0" settings.


This I can not verify. When encoding with -V2 lame always reports VBR(q=2), no matter what the switch -q specifies. I've tried this with different lame-compiles (from bakerweb and from Rarewares) and it always turns out this way. Then, which lame.exe are you using?

--vbr-new effectively has only 3 -q settings: Values from 0 to 3 are equivalent, as are 4-6 and 7-9 respectively. Also, in all modes, -q3 is now default, with -h equivalent to -q2.
Go to the top of the page
+Quote Post
moozooh
post May 27 2010, 14:27
Post #89





Group: Members
Posts: 357
Joined: 22-September 04
From: Moscow
Member No.: 17192



LAME has become surprisingly good in the last few years. Its current pre-echo robustness is amazing (3.98.4 -V0 and -b 320 have caught up to Vorbis aoTuV b4 -q5 according to my experience, which is unbelievable considering the limitations), and the psy-model is really efficient at spending bits. Hats off to Gabriel, Robert and the rest of the team!

Sometime ago I decided to encode all music for my own portable needs at ~150–160 kbps, because I found it to be the perfect bitrate at which a high degree of transparency can be achieved, thus my codecs of choice nowadays are LAME 3.98.4 at -V 3.5 for artifact-robust music, and either Vorbis aoTuV post-b5.7 -q 5 or Musepack 1.3.0 SV8 -q 5 for the rest, depending on the resulting bitrate.

This post has been edited by moozooh: May 27 2010, 14:28


--------------------
Infrasonic Quartet + Sennheiser HD650 + Microlab Solo 2 mk3. 
Go to the top of the page
+Quote Post
а.п.т.
post May 28 2010, 12:09
Post #90





Group: Members
Posts: 36
Joined: 25-January 09
Member No.: 65946



QUOTE (moozooh @ May 27 2010, 16:27) *
Sometime ago I decided to encode all music for my own portable needs at ~150–160 kbps, because I found it to be the perfect bitrate at which a high degree of transparency can be achieved, thus my codecs of choice nowadays are LAME 3.98.4 at -V 3.5 for artifact-robust music, and either Vorbis aoTuV post-b5.7 -q 5 or Musepack 1.3.0 SV8 -q 5 for the rest, depending on the resulting bitrate.


If I get it right, you use a 3-phase process with your transcodes:
1. Encode to mp3 (LAME)
2. Check whether you can catch some artifacts in the result files, and if you can:
3. Encode to ogg and mpc and choose the filess with appropriate bitrate.

I wonder, why don't you skip the first two phases, though? You will save time and you will make the process more error-proof. Otherwise, you should check carefully every mp3 file and even then it is possible to omit some artifacts.
Go to the top of the page
+Quote Post
moozooh
post Jun 1 2010, 00:21
Post #91





Group: Members
Posts: 357
Joined: 22-September 04
From: Moscow
Member No.: 17192



Actually it's somewhat easier. If I know an album isn't prone to pre-echo (pretty much the only artifact I can easily distinguish at the bitrate range I've chosen), or I'm not sure, or I don't care much, I'll encode to all three simultaneously and see which one results in the lowest bitrate. If that happens to be LAME, I'll quickly check if I'm bothered by quality, and if I'm not (which is what happens most of the time), I'll take it.

However, if I know an album is prone to pre-echo — and a lot of the music I listen to is like that — I'll just encode to Vorbis and Musepack, skipping LAME, and see which one results in lower bitrate. Otherwise these two are pretty much identical for my purposes, and I can't distinguish between them.

It requires slightly more time, but the methodology is pretty simple and a satisfying result is ensured.



--------------------
Infrasonic Quartet + Sennheiser HD650 + Microlab Solo 2 mk3. 
Go to the top of the page
+Quote Post

4 Pages V  « < 2 3 4
Closed TopicStart new topic
1 User(s) are reading this topic (1 Guests and 0 Anonymous Users)
0 Members:

 



RSS Lo-Fi Version Time is now: 25th July 2014 - 15:45