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: --alt-preset standard and --r3mix (Read 12300 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

--alt-preset standard and --r3mix

The r3mix preset is tweaked to come down in bitrate and to sound better than
the old r3mix within Lame 3.88? This goal was reached. It is even very fast.
Adding nssafejoint would have made the new r3mix much bigger than the old r3mix.
You can imagine how people would have complaint when the new setting pruduced much
bigger files.

R3mix is very good to use with CDEX and is rational to explain to people that are
not to busy seeking for faults in their encoded material. I am still happy it exists.

I can´t explain to fellows - to go thru ripping with that - encoding with Razorlame -
tagging with that! They would call me it is a waste of time.

We have to thank this r3mix preset cause it somehow set a standard in easy to use mp3
encoding and kept many people from chosing totally stupid command lines.

The people that have found this r3mix compromise (included me) had only these set of
standard switches and Roels ATH, no hacking in the source.

So r3mix has its right to exist.

At the moment it is maybe time for another r3mix preset, but with standard switches it would
be nearly impossible to get much better quality at these low bitrates.

But you should know that...

Nobody can start a new thread at r3mix with finding the follower for the r3mix setting
without beeing compared to your presets. The only people who have the knowledge how
the new improvemnts work are on this board here and nobody can use a switch to make
them work.

I can´t speek for Roel but i think people are familar with using r3mix and there
should be a possibility to find a better preset together with people across the boards.
Letting the r3mix prest die may strength some egos but people are familar with that. So lets
give them what they want.

And now, i jump out of the window!

Wombat
Is troll-adiposity coming from feederism?
With 24bit music you can listen to silence much louder!

--alt-preset standard and --r3mix

Reply #1 –
What does that message has to do with the thread subject??

rev7 is in testing and development phase and already Dibrom should hurry to develope all new vbr-mtrh code level tweaks for Roel and his --r3mix which is already "perfect sound quality at optimal bitrate"?

uh oh...

Do you have any idea what you are talking about anyway, Wombat?? Nobody here is questioning r3mix setting's right to exists.

Personally I don't think there should be any personal "hack"-settings in Lame like --r3mix (or-dm-preset which was abadoned). It just gives newbies a picture that you always have to tweak settings to achieve optimal quality. So far this has been the case, but it shouldn't be so. Lame should provide highest possible official default profiles. This is what Dibrom is trying to do - implement best possible "official" profile set, alt-preset.

Anyway, if this discussion continues, better split the thread...
Juha Laaksonheimo

--alt-preset standard and --r3mix

Reply #2 –
Well I don't want to start a flame war here but I have a few comments on this...

Quote
Originally posted by Wombat
The r3mix preset is tweaked to come down in bitrate and to sound better than
the old r3mix within Lame 3.88? This goal was reached.
[/b]

Perhaps.  However, I'd say that there certainly wasn't a very extensive set of listening tests done when the new preset was made, same goes for Athtype 3.  I'm becoming skeptical of whether or not the new --r3mix is really better all around than the last --r3mix switch, aside from a few samples.  The fact is, I just don't know of much testing that supports all of this.  I know on some of the samples nspsytune offered improvements on, the new preset was a bit better.. but it also doesn't use --nspsytune correctly because of no max noise measuring.

So in short, I'm not necessarily convinced the newer --r3mix is better.  It certainly doesn't do so well on quite a few samples..

Quote
Adding nssafejoint would have made the new r3mix much bigger than the old r3mix.
You can imagine how people would have complaint when the new setting pruduced much 
bigger files.


True, but look at --alt-preset standard.  It uses quite a few more stereo frames than --r3mix and gains most of the advantages from using --nssafejoint without significantly increasing the bitrate.  So it can be done.

Quote
R3mix is very good to use with CDEX and is rational to explain to people that are
not to busy seeking for faults in their encoded material. I am still happy it exists.


I'm not sure I see the significance of this statement.  People can use --alt-preset standard with CDex just as easily as --r3mix, so where's the advantage?

Quote
I can´t explain to fellows - to go thru ripping with that - encoding with Razorlame -
tagging with that! They would call me it is a waste of time.


So just explain to them how to configure CDex to use --alt-preset standard then.

Quote
We have to thank this r3mix preset cause it somehow set a standard in easy to use mp3
encoding and kept many people from chosing totally stupid command lines.


Well this is certainly a touchy subject, but I'm going to go out on a limb here and also say that it has hampered progress some.  The original reason I was motivated enough to form this site is because it was impossible to try and work on improving things on the r3mix.net board because of constant opposition towards experimentation or even trying to explain why certain things offer higher quality than --r3mix.

So yes, it may help the newbies, but at this point I don't see it doing much to really break new ground.  In fact what I see instead are constant statements about "perfect CD Quality", "guaranteed perfect", "absolutely transparent", etc.  Most of the people around here now know that such a thing with any preset is pretty much a joke, but this certainly isn't made clear with --r3mix.  In addition, it gives the impression that further improvements are unnecessary and unimportant.

Quote
The people that have found this r3mix compromise (included me) had only these set of
standard switches and Roels ATH, no hacking in the source.

So r3mix has its right to exist.


I don't quite understand what you mean here.  Sure, --r3mix has a right to exist, but as JohnV says, I don't think these kind of preset hacks should be promoted.  There should be a consolidated and uniform effort to provide a "standard" set of presets that "just work".  MPC has this, PsyTEL has this, Vorbis has this in a way... LAME should also.

Quote
At the moment it is maybe time for another r3mix preset,


Good luck convincing the necessary people that this is needed

Quote
Nobody can start a new thread at r3mix with finding the follower for the r3mix setting
without beeing compared to your presets. The only people who have the knowledge how
the new improvemnts work are on this board here and nobody can use a switch to make
them work.


I'm not sure if this is a negative comment or not.

If you are blaming me for not being at r3mix.net trying to improve things over there then I think you should know better than that.

I'm also getting tired of the reaction I'm getting from a few people in the community over my work now.

So far, I've seen you get mad, mitiok, cd-rw.org, not to mention others.  People accusing me of trying to build some sort of LAME "empire" and keep everything to myself... people accusing me of trying to hijack the LAME project.  Heh.. it's really starting to get old now.  I've gotten so much flak from a lot of people in the community that quite often I just consider dropping the whole damn thing and doing something else.  I may seem callused sometimes, and a bit curt these days, but that's not because that's the way I want things to be... rather its a side effect of having to constantly deal with these types of reactions.

At any rate, the fact of the matter is that nobody else is really doing the work to make these presets better and until someone else does and actually has some grounds to stand on as far as complaining goes, they should just lay off.

I'm trying to provide 1 setup that works, without all the experimental stuff being necessary.  Considering your comments and the comments of others, I would think you would like this.  Didn't you just mention a bunch of things about newbies and how if something is too complicated they won't want to use it?  Heh.  I am constantly seeing this inconsistency now... people want simplicity but at they same time they want all kinds of new experimental switches.  Hrmm..

Quote
I can´t speek for Roel but i think people are familar with using r3mix and there
should be a possibility to find a better preset together with people across the boards.


Lol...

Have you forgotten everything that has happened in the past at r3mix.net? 

For a quick refresher, when I first started posting my findings to the board I was trying to improve --r3mix.  Go back and look at my first posts if you don't believe me.  The original goal was actually to make --r3mix better.  After some time, it became clear that Roel was uninterested in that.  Then came opposition from people like Julius and others.  Well I stuck around and still kept providing test results and tried to find ways to improve things.  That's when I decided to create a new preset (because working together wasn't going to happen).  Even after that point and after listening tests were constantly confirming benefits (even his own test!) he was uninterested so I eventually just decided to go elsewhere and work my own way.

Really, I have no responsibility to anyone.  I don't owe anything to any of the r3mix.net users or Roel.  I provide a service here and do what I can to make things work and I'm more than willing to cooperate with people who are interested, but I'm not going to go out of my way, against opposition, to try reach some sort of compromise anymore.  I don't need to do that, I don't have to do that, and unless I see some sort of effort from the other side so to speak, I'm not even going to worry about it.

Quote
Letting the r3mix prest die may strength some egos but people are familar with that. So lets 
give them what they want.


Pinning the responsibility for what happens to --r3mix on me is a pretty pathetic thing to do.  Sorry for saying so, but as I just explained, I've had more than my share of negative feedback when I was already trying to help in the past.  As I said, I owe nothing to anyone on r3mix.net or to Roel.

All I care about are the people in the community here or at least the friends I have at r3mix.net that I do talk to.  Judging by what they are saying I have been giving them what they want.  Why the hell should I bother to do all of this under the --r3mix banner?  Roel has made it clear that it is his preset, a quality level defined by him alone and not the community (or else he wouldn't constantly disregard test clips and quality improvements) and that he is uninterested (or at least thinks it is marginal at best) in my work for the most part.

This has nothing to do with ego, it has to do with practicality.  I'm more interested in results, I'm tired of the politics and I'm not going to argue with someone else about what should or shouldn't be done.  I just do it.... and it works.  That's enough for me, and apparently that's enough for the users (just look around at some of the positive feedback).  Everything else is unimportant.

Anyway, none of what I'm saying here should be taken personally by anyone.. I want to make that really clear because I'm not trying to start a fight or anything like that.

--alt-preset standard and --r3mix

Reply #3 –
Quote
In fact what I see instead are constant statements about "perfect CD Quality", "guaranteed perfect", "absolutely transparent", etc. Most of the people around here now know that such a thing with any preset is pretty much a joke, but this certainly isn't made clear with --r3mix. In addition, it gives the impression that further improvements are unnecessary and unimportant.


I endorse this comment. When I first started researching audio compression/data reduction codecs, I ended up at r3mix very early in the process via links. Reading that site led me to believe that mp3 was CD quality at 250kbps, but that using r3mix, this same level of quality could be achieved at a lower bitrate.

It wasn't until I started reading the r3mix board, and then came here, that reality bit. The work that Dibrom has done has been addressed at improving the known problems with mp3, which seem to be swept under the carpet at r3mox - God knows why they do it - blind faith?

Of couse, the other thing I have learnt is that MusePack appears to be superior to mp3 on most ocassions anyway, so I don't use mp3 for archiving now, only for sharing.
Ruse
____________________________
Don't let the uncertainty turn you around,
Go out and make a joyful sound.

--alt-preset standard and --r3mix

Reply #4 –
Well, i try to explain.

quote: (btw. my 1st quote)
Perhaps. However, I'd say that there certainly wasn't a very extensive set of listening tests done
when the new preset was made, same goes for Athtype 3

Thats right, people were happy enough to have very much improvement over the last months before
this release. To be honest it has become more scientistic over the last time and today it would
sure be different in the result.

quote:
I'm not sure I see the significance of this statement. People can use --alt-preset standard with
CDex just as easily as --r3mix, so where's the advantage?

People i talk on the telephone or just give a short advice is - download CDEX -Choose Lame r3mix -
thats it. This can change when alt-presets can find their way there.

quote:
There should be a consolidated and uniform effort to provide a "standard" set of presets that "just work".
MPC has this, PsyTEL has this, Vorbis has this in a way... LAME should also.

This is the main part why my message comes across a bit reproachful. I am a bit dissapointed that
all i can do is wait for another suggestion of you and listen to it. So i have to take anything as
given and doesn´t have a real plan what is happening. This is everbodys but your fault cause i am
not as deep gone into Lame as you.

qute:
To much to quote - just the rest about you and the r3mix forum..

Your explanation of how this all comes together is understandable.

After all, it seems like my ego is this what made me write the above post.

To say all this in public shouldn´t happen to often.


Wombat
Is troll-adiposity coming from feederism?
With 24bit music you can listen to silence much louder!

--alt-preset standard and --r3mix

Reply #5 –
Quote
Originally posted by Wombat
Thats right, people were happy enough to have very much improvement over the last months before
this release. To be honest it has become more scientistic over the last time and today it would
sure be different in the result.


I'm sure that's true to an extent, but also even when the very short testing period of the new preset was going on, there were known issues that where largely ignored.  It seems to me that it's just a matter of not really being interested in pushing for that last extra bit of quality.  The problem is that just doesn't make sense to me, especially considering much of what is written on r3mix.net about "CD Quality" and all that.

On top of that, it's one thing to not be able to do much about it if you don't dabble in the source, but its certainly an entirely different issue when you begin to downplay and marginalize the work other people are trying their hardest to apply for the benefit of everyone.  This is why collaboration is pretty much impossible I think.

Quote
People i talk on the telephone or just give a short advice is - download CDEX -Choose Lame r3mix -
thats it. This can change when alt-presets can find their way there.


I do see your point here, but I don't think it is quite so hard to apply a few extra steps.  Maybe that's just me though.  Anyway, I eventually plan on trying to modify the .dll to work with these presets so that should solve this problem.

Quote
This is the main part why my message comes across a bit reproachful. I am a bit dissapointed that
all i can do is wait for another suggestion of you and listen to it. So i have to take anything as
given and doesn´t have a real plan what is happening. This is everbodys but your fault cause i am 
not as deep gone into Lame as you.


Yes.. I'd have to agree and say that this isn't particularly my fault or problem.  The work I have done has been made available for all to see and those with the time and desire are more than welcome to try and work on this stuff with me.  I just think it would be irresponsible to expose even more internal behavior on the command line.  You talked some about users picking stupid command lines and not knowing with many of these options do... just imagine how much worse that could get if every new tweak brought a new switch along with it.

Quote
To much to quote - just the rest about you and the r3mix forum..

Your explanation of how this all comes together is understandable.

After all, it seems like my ego is this what made me write the above post.

To say all this in public shouldn´t happen to often.


Perhaps I did come across a little bit harshly, but I really wanted to emphasize the point that I did try to work together, and that I am working for the good of everyone here.  That's all.

--alt-preset standard and --r3mix

Reply #6 –
Over the years, I've grown accustomed to MP3 compression and almost emjoy the occasional relic on Lame encoders in much the same way as vinyl lovers like the crackling sound that, well... makes vinyl... vinyl! So, I would say that, MP3 being far from perfect, will always have a place in my heart, no matter what "better" formats are developed.

--alt-preset standard and --r3mix

Reply #7 –
Quote
emjoy the occasional relic on Lame encoders in much the same way as vinyl lovers like the crackling sound that, well... makes vinyl... vinyl!


You should get an old scratched Beatles 45rpm single, play it on a Garrard stacker, encode it using an Xing mp3 encoder: you'd be in pig heaven
Ruse
____________________________
Don't let the uncertainty turn you around,
Go out and make a joyful sound.

--alt-preset standard and --r3mix

Reply #8 –
Quote
People i talk on the telephone or just give a short advice is - download CDEX -Choose Lame r3mix -


As I remember, the dll r3mix is not the same what exe r3mix is. The files were much higher in bitrate because not every switches could be set for .dll. Am I wrong?

If not, is it really r3mix setting so easy to use with cdex?


If we were talking about Roel himself - I think HE tried to build some sort of LAME "empire" and keep everything to himself.

I first visited his site when I was looking for some improvement of stereo separation. I had problem with sound on surrond system - I was using xing at that time. I found r3mix setting for lame, it was supposed to be perfect so I was so happy! But it turned out that it wasn't. I wrote about it - was a complete newbie at r3mix then - and Roel was so mad at me! He wrote that is impossible to solve all problems and my problem doesn't make sense, and, and... I thought I was stupid or something. Maybe that problem wasn't so important - I'm not sure.

Anyway, when mtrh was improved recently, Roel made his new r3mix setting and said that now it is fast but also it is perfect quality so I started using it. Then I found nasty problem (nasty for me) with low volume clips. Roel tried to convince me it is impossible to hear that problem! That was too much for me. He decided on his own what is perfect quality assuming that no one has better hearing than he does.

Then Dibrom wrote he hears the problem with low volume, then there were some other clips with similar problem - that were some of the last posts by Dibrom at r3mix forum, later it was only war. Then Dibrom made his own forum what made me very happy. He improves Lame as r3mix never tried to do, and this forum is very "user-friendly". If you show a problem to Dibrom, he tries to solve it, Roel tries to convince you that you don't really hear that problem  Roel does not want to change anything. So I don't see point that Dibrom should try to cooperate with him.

Hmm... never wrote longer post...

--alt-preset standard and --r3mix

Reply #9 –
Dib: I don't know what impression I gave, but I never thinked you were trying to appropriate Lame to yourself.
My suggestion about a fork was a real suggestion (perhaps not the best, I don't really know).

I've got a new "today's suggestion": keeping in mind that we must keep all the switches into the Lame main branch a possible solution could be: use by default your presets (at least for cbr/abr) and provide a --nopreset switch for people willing to disable it.

--alt-preset standard and --r3mix

Reply #10 –
Quote
Originally posted by HomeK
 
If we were talking about Roel himself - I think HE tried to build some sort of LAME "empire" and keep everything to himself....

Roel tried to convince me it is impossible to hear that problem!


Like Dibrom, I have not forgotten Roel's duplicity.

Roel talks about "community" but rules it like a dictator--this coming from a man who politically says he believes in communism... can we get any more contradictory?

I tried to private message him about certain issues which he would just ignore and/or marginalize them.

Contrary to wombat, I see Roel, his board, and his beloved --r3mix setting as a hinderance to
progress in lame. 

I have personally witnessed new insights and issues squashed on his site because Roel did not agree with them...

I have personally witnessed censorship there too (anyone remember ww.h y d r o g e n a u d i o.org?) ....

On a daily basis he feeds thousands a lie about the "perfect" and "cd quality" of the --r3mix setting.

Don't believe me? Still to this very day he has on his quality page concerning --r3mix:

"Remarks: Perfect(x) sound quality at optimal bitrate."

And if you click on perfect it says:

"...My main page clearly states "produce 'cd-quality' mp3's".  For the sake of objectivity
I will yield and admit that there is no such thing as "perfect" in real life, but the vast majority of people who find cd-audio to satisfy their needs will find use for these pages."  (Note how he accepts on technicallity --r3mix is not pefect, yet still maintains its close to perfect... having his cake and eating it too...)

As a result many uncritically accept and use --r3mix to encode their mp3s... and I do not see this changing anytime soon, unless with 3.90 stable, --alt-preset standard, and a good advertising campaign back-uped with user feedback and listening tests, most people finally realize the many problems that plague "--r3mix".

There is no doubt that in his early career Roel helped promote lame and quality, and we should be thankful for this.  But his later career, ruled by an anti-intellectual fervor to maintain control and dominance, has DONE NOTHING BUT HARM to the lame community and its progress. 

And I am not one who loves to flame.  I had over 100 "friendly" posts there--trying to give him the benefit of the doubt, until it became very clear to me what sort of man he is...

Anyway, these are my last comments about "that man"; and were written lest others FORGET the past.

PS: Greets to Neo Renegade (audiograbber... )

And once again thanks to Dibrom for his remarkable improvements to Lame and for making a friendly, open community we can all feel proud of...

--alt-preset standard and --r3mix

Reply #11 –
Well I believe it is clear where most people stand on this issue.  I'd rather not turn this thread into a Roel/r3mix.net bashing thing, but it does appear to be noted now and understood by many people that despite the claims of a few regarding my work, and how they might perceive it as a negative thing, the majority do not seem to view it this way at all.

Quote
Originally posted by Gabriel
Dib: I don't know what impression I gave, but I never thinked you were trying to appropriate Lame to yourself.


It is good to know this.  Sometimes it is difficult to gauge community reaction for various reasons, and sometimes I probably pay more attention to the negative comments than I should but anyway..

Quote
I've got a new "today's suggestion": keeping in mind that we must keep all the switches into the Lame main branch a possible solution could be: use by default your presets (at least for cbr/abr) and provide a --nopreset switch for people willing to disable it.


I think this would be a very good idea, and could actually solve many of the functionality problems (from an end user's point of view) in LAME.  It would certainly allow for LAME to be much more user friendly.  And I think if people really want to experiment, having them use a --nopreset switch to turn all experimental stuff back on would make sense.  Perhaps this along with a warning that overriding default settings may not lead to optimal quality.

Anyway though, I think the vbr presets should be defaulted as well, or promoted as the "official" settings (overriding the older and less tuned presets).  To those who use --r3mix, this shouldn't really be an issue anyway because --r3mix has always been a sort of "external" setting.  It is a setting defined only by Roel on r3mix.net and he retains complete control over what goes in this.  What I am trying to do with the vbr settings I have created is different.  I am trying to build a setting according to the specifications of the people in the community (with more than 1 setting for different purposes) specifically to make LAME easier to use and to make it easier to get high quality encodings by default.  I hope that ego won't end up playing a role here and that these settings will not be defaulted because it might hurt some people's feelings.  The fact of the matter is that the vbr presets I have been working on are where the quality is at.. this can be proven, and just about everyone who is participating in testing knows this to be true as well.

Also, I think that by defaulting high quality presets, it would make this LAME tag stuff work much better.  Instead of storing all of these experimental settings which can change at any given moment, we could just store the preset used.  This is what MPC does and I think it works really well.

--alt-preset standard and --r3mix

Reply #12 –
Quote
Originally posted by Ruse


You should get an old scratched Beatles 45rpm single, play it on a Garrard stacker, encode it using an Xing mp3 encoder: you'd be in pig heaven
lmfao!

--alt-preset standard and --r3mix

Reply #13 –
first point: I'm not against (ok, perhaps a little) defaulting vbr presets, but practically, how would you fit your presets into the current -V scale?

second thing: I think that the current "Lame tag" (which is not really Lame specific as it could easily be described in a non lame-centric way) is a low level info stuff, with a lot of info that are usefull or could be usefull. The preset used is more an high level thing. I personnaly think that the place of the preset name should better be into an id3tag than into the Info Tag.

--alt-preset standard and --r3mix

Reply #14 –
Quote
Originally posted by Gabriel
first point: I'm not against (ok, perhaps a little) defaulting vbr presets, but practically, how would you fit your presets into the current -V scale?


Hrmm.. well for one I think that named presets is still a good idea to include for simplicity.

However, I don't think that fitting the VBR presets into the V scale would be so difficult.  I could build a system similar to how the ABR presets work, using --alt-preset standard as a template.  Then according to the V setting specified it would change a set of internal options, not just adjust the masking (or the ath with athtype 4) as it does now.  Thinking about it, that actually could work really well.  I think this is a good idea so I'll start working on it some and see what I come up with.

Quote
second thing: I think that the current "Lame tag" (which is not really Lame specific as it could easily be described in a non lame-centric way) is a low level info stuff, with a lot of info that are usefull or could be usefull. The preset used is more an high level thing. I personnaly think that the place of the preset name should better be into an id3tag than into the Info Tag.


My only issue with the LAME tag is that it's going to be used in a more high level manner via Encspot.  I'm concerned that as internal behavior in LAME changes, that will break compatibility with the tag, or at least will make all tags created before the changes somewhat useless.  This is why I think storing presets themselves could be a better idea, because then you know what particular settings where used at a given point in time and if internal behavior changes it's not an issue because its still all encapsulated within the preset.

I just don't see much of a practical way to store all of the really important stuff in the tag and continue to improve lame without causing problems each time something major is added/changed/defaulted/removed, or whatever.

--alt-preset standard and --r3mix

Reply #15 –
I for one appreciate all the hard work you are putting into this Dibrom. Keep truckin sukka! Great job.