Skip to main content

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: 3.90.3 versus 3.96.1 (Read 90824 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

3.90.3 versus 3.96.1

Reply #25
Right now 3.97a7 it's faster and have noticed nothing bad or worse yet.

3.90.3 versus 3.96.1

Reply #26
Thanks to the wonderful dev0 we have a 3.90.3 vs. 3..96 ABX test thread.

I personally think that it's best to use 3.90.3 only for --alt-preset standard encodings.

3.90.3 versus 3.96.1

Reply #27
Quote
Thanks to the wonderful dev0 we have a 3.90.3 vs. 3..96 ABX test thread.[a href="index.php?act=findpost&pid=279907"][{POST_SNAPBACK}][/a]
That was 3.90.3 vs. 3.96b2. Are we sure it still applies to 3.96.1?
Sergio
Sergio
M-Audio Delta AP + Revox B150 + (JBL 4301B | Sennheiser Amperior | Sennheiser HD598)

3.90.3 versus 3.96.1

Reply #28
Been using 3.96.1 to encode about 50% of my library, but taking a pause until 3.97 gets worked out, as the --vbr-new stuff is interesting.

3.90.3 versus 3.96.1

Reply #29
Call me an unreconstructed, hidebound traditionalist, but I'm sticking to 3.90.3 -- at least until 3.97 comes out. 
WavPack 4.31 / LAME 3.98 alpha 3 -V9 -vbr-new

3.90.3 versus 3.96.1

Reply #30
Quote
Call me an unreconstructed, hidebound traditionalist, but I'm sticking to 3.90.3 -- at least until 3.97 comes out. 
[a href="index.php?act=findpost&pid=279925"][{POST_SNAPBACK}][/a]


What will stop you from being a "traditionalist" then?

3.97 won't be tested as much as 3.90.2 was. At some point, as many of us did, you will have to take a leap of faith.

Although I do undesrtand your point, and if you are archiving, it would be ill-advised to use anything but a stable version. But if you are only "temporarly" encoding (for example, to listen to music on your DAP, or something like that) you could enjoy the benefits on the newer versions, and I would bet you wouldn't notice the quality difference (if there is indeed one of significant importance to speak of)
I'm the one in the picture, sitting on a giant cabbage in Mexico, circa 1978.
Reseñas de Rock en Español: www.estadogeneral.com

3.90.3 versus 3.96.1

Reply #31
Quote
3.96.1. 3.90.3 is effete.
[a href="index.php?act=findpost&pid=279658"][{POST_SNAPBACK}][/a]

I postulate that Roberto invokes a thesaurus for such exuberant articulations.

edit:  I use 3.96.1 because it is faster and produces better results for my portable (-V5 --athaa-sensitivity 1) then 3.90.3.
gentoo ~amd64 + layman | ncmpcpp/mpd | wavpack + vorbis + lame

3.90.3 versus 3.96.1

Reply #32
I use 3.96.1 when I need to make mp3s because I honestly can't hear a difference between it and 3.90.3 for --alt-preset standard.  3.96.1 is also a lot smaller which is kind of the point of using lossy compression IMO
Nero AAC 1.5.1.0: -q0.45

3.90.3 versus 3.96.1

Reply #33
I use 3.96.1 -V5 --athaa-sensitivity 1, as per music_man_mpc (and the setting used in rjamorim's recent 128kbps test).

I think this thread should be a poll, and should have some bearing on the HA recommended version.  As it appears thus far that the majority are using 3.96.1, isn't it a little wrong to be recommending 3.90.3?!

Whether it is less tested or not, if the majority of the community are using it then surely that is the implied recommendation of that community?

Edit: Sorry, just re-read the initial post. My reason is thus: to support the continued development of LAME.  If everyone stuck with 3.90.3 no improvements would ever be made.  Unfortunately my support is of little benefit, compared to the those great souls who actually conduct the ABX tests which let me make my informed decisions, but it's a camaraderie thing.
I'm on a horse.

3.90.3 versus 3.96.1

Reply #34
Quote
I postulate that Roberto invokes a thesaurus for such exuberant articulations.

Nah, its Mac's word of the month.

This discussion can go on for ever.. In reality, I dont expect anyone is disappointed / annoyed at the quality of either 3.90.3 or 3.96.1

3.90.3 versus 3.96.1

Reply #35
Quote
This discussion can go on for ever.. In reality, I dont expect anyone is disappointed / annoyed at the quality of either 3.90.3 or 3.96.1
[a href="index.php?act=findpost&pid=280046"][{POST_SNAPBACK}][/a]

Same goes for many formats. MPC encoders made by Buschmann are as transparent as Frank Klemm's one for most people (if not all...). But the old one is not recommanded.

The problem is HA.org current recommandations, and not LAME quality.

3.90.3 versus 3.96.1

Reply #36
Quote
The problem is HA.org current recommandations, and not LAME quality.
[a href="index.php?act=findpost&pid=280048"][{POST_SNAPBACK}][/a]


I wholeheartedly agree.


@Lev: actually, I mentioned that word to Iain first, as well as other oddities like ennui and some slightly unmentionable ones.

3.90.3 versus 3.96.1

Reply #37
as said earlier, get a poll for perfered lame encoder version.  that might have more impact, and give a better representation of everyone here.  careful how you word stuff tho.  dont want to step on any toes.  i know alot of work was put into 3.90.3, and it is undisputedly amazing what dibrom and others have done with lame.
a windows-free, linux user since 1/31/06.

3.90.3 versus 3.96.1

Reply #38
I understand why HA can only recommend 3.90.3 because that is the only one that has been properly tested.

What I don't understand is the comment that 3.96.x is better at lower bitrates than 3.90.3 (Lame Versions on HA).  This advice is repeated often on these forums.  If it has not been tested then why should there even be a recommendation?

Anyone know of the background to this recommendation?  Is there some thread I have missed?

3.90.3 versus 3.96.1

Reply #39
3.96.1 -V1

For me, best compromise between quality and bit rate based on my personal ABX tests...

3.90.3 versus 3.96.1

Reply #40
Polls - suprisingly - only show whats popular. It doesn't tell you anything about the reasons why people did choose something nor does it tell you which one overally is better. Polls are just a popularity-meter... not more, not less.

- Lyx
I am arrogant and I can afford it because I deliver.

3.90.3 versus 3.96.1

Reply #41
Quote
Polls - suprisingly - only show whats popular. It doesn't tell you anything about the reasons why people did choose something nor does it tell you which one overally is better. Polls are just a popularity-meter... not more, not less.

I basically agree - but what I'm saying is, rightly or wrongly, if the majority of the members of this forum are using 3.96.1 how we can credibly tell new members that 3.90.3 is the codec we'd recommend?

This forum is its members, and should reflect their choices (whether they are right or wrong).

If I were asked personally what version someone should use I would say 3.96.1, because that's what I use, and I feel that I have made an informed decision in choosing my encoder.

Someone may say, quite rightly, that I am not as informed as the senior members of this forum. However, when the majority of the members say the same thing, that person can't say that we are all imbeciles and voting against the ideals of Hydrogen Audio... we are Hydrogen Audio.

Edit: I am very aware that there have been 1849 views on this thread, and only 38 replies.  A formal poll, in contradiction to the minor response so far,  may reflect a huge majority for 3.90.3 - in which case I would gladly see it remain as the recommended codec (well, I could understand it anyway).  I am also aware that 3.97 may not be too far off, and perhaps that version will be the one to finally take the throne.  I just think we've raised some interesting questions here.
I'm on a horse.

3.90.3 versus 3.96.1

Reply #42
Most of you will have noticed by now, but I'd like to invite you to place your votes in this preferred MP3 codec poll.

3.90.3 versus 3.96.1

Reply #43
Quote
What will stop you from being a "traditionalist" then?


Sorry, I meant to say I'm being conservative in my choice of lame version, at least for now.  I've actually been thinking of switching to 3.96.1 when I came across some posts here saying it had a few problems with highly tonal music. Yes, I know that most people can't tell between 3.96.1 and 3.90.3 in most cases,  but then about the same time excitement began bubbling over hereabouts with testing of the 3.97 alphas (not to mention the news about work being done on LAME 4).  I gathered from the alpha testing thread that 3.97 looks quite promising and even trumps 3.96.1 in some cases (and 3.90.3 IIRC), so I've decided to wait for 3.97 instead.  Actually, I can't wait to try it out, so  I guess I'm not as hidebound as my post implied after all, because I do want LAME to get even better. 

Quote
3.97 won't be tested as much as 3.90.2 was. At some point, as many of us did, you will have to take a leap of faith.


It looks like many have indeed. I seem to be the only one who voted for the recommended version so far.   

cheers
WavPack 4.31 / LAME 3.98 alpha 3 -V9 -vbr-new

3.90.3 versus 3.96.1

Reply #44
I use both for different purposes.

I understand that a lot of people use version 3.96.1 for one of the following reasons:

1) other people say it's good
2) they can't tell the difference (even with critical samples)
3) it's faster
4) they blind-tested 3.96.1 (or whatever the latest stable version) and found it better than 3.90.3
5) it's the latest stable version
6) other

If people want to change the recommended version for the reason 1) or 2) or 3), I don't think that's a good enough reason. That's lowering our standards. Rather, I would love to see 3.90.3 overthrown for the reason 4), which is the hardest and the most time-consuming (but true to the original HA principles) option.

5) seems like a tricky one. From the developers' standpoint, this is the way to go, and they have no reason not to recommend the latest stable version. But since HA is a separate entity from the LAME project, it should ideally hold onto its own standard (QA) that it's founded upon.

I think a compromise (copout) might be to recommend the latest stable version, and move 3.90.3 aside to a Legacy status (or an alternative recommendation) - something that we can fall back on if we want to, and to encourage users to report any quality problems. Or, do the right thing and test the hell out of the latest stable version (or the lastest alpha) in order to overthrow 3.90.3. But who would have the time and energy to volunteer except Guru?

Only my 2c.

 

3.90.3 versus 3.96.1

Reply #45
Quote
I've actually been thinking of switching to 3.96.1 when I came across some posts here saying it had a few problems with highly tonal music.
[a href="index.php?act=findpost&pid=280255"][{POST_SNAPBACK}][/a]

What posts? Which samples?

3.90.3 versus 3.96.1

Reply #46
Quote
But since HA is a separate entity from the LAME project, it should ideally hold onto its own standard (QA) that it's founded upon.

But let's be honest: nowadays, I can easily list the number of people testing Lame versions using fingers of a single hand.

3.90.3 versus 3.96.1

Reply #47
Quote
Quote
I've actually been thinking of switching to 3.96.1 when I came across some posts here saying it had a few problems with highly tonal music.
[{POST_SNAPBACK}][/a]

What posts? Which samples?
[a href="index.php?act=findpost&pid=280297"][{POST_SNAPBACK}][/a]


whoops... I think I may have been wrong about the "problems with highly tonal music" (my mistake, sorry  ), but I do remember looking at [a href="http://www.hydrogenaudio.org/forums/index.php?showtopic=19813&st=25]this thread[/url] and the regression samples reported here.    There was also some stuff about abr/cbr at here.
WavPack 4.31 / LAME 3.98 alpha 3 -V9 -vbr-new

3.90.3 versus 3.96.1

Reply #48
I'm no programmer, but could Lame 3.90.3 be further optimized
for speed?
Everyone universally likes 3.90.3, but its speed at preset standard can be painful.

3.90.3 versus 3.96.1

Reply #49
Quote
I'm no programmer, but could Lame 3.90.3 be further optimized
for speed?

Yes, but who would do it?