Hey peeps, just registered today...

I'm using, and have used for ages, EAC 0.9b4 and LAME 3.92, with no problems. My rips (Plextor 16/10/40A, old but rock-fokken-solid) and encodes (192-320kbits js vbr 0 no filtering), have always come out pretty flawless as far as I can tell, so I'm still happy using them.

But they are both pretty old now. I've been semi-casually following the progress of both apps, and EAC seems to have only improved as far as I can tell, but LAME has been plagued with various quality problems from 3.93 onwards. There are so many threads on the latter, especially, that it's difficult for me to make a judgement on it. But since I'm here I thought I might as well ask... should I bother updating either yet, considering what I'm currently using is producing excellent results already?

All the best,

-Chis
www.chis.nildram.co.uk/chis_index.html

I can't see any reason that holds one from upgrading eac, since there are no major bugs (at least known to me) in versions following 0.9b4.
things like accurate-rip have been added, and I guess some small bug-fixes etc. have taken place.
you might want to read the changelogs in order to make up your mind.
one thing though, if you worried about the toc-'issue', stick to 0.95pb3.

coming to lame, that's not so clear.
you could safly upgrade to 3.93.1. if you want further information about the pros and cons of upgr. to a later version browse the forum, there are LOTS of infos there
if you want to take part in the evaluation of the newer builds (3.96), read this.
you could of course stick to your strusted setup as it is now, no harm done if you are happy with usability and quality. but then, if no one tests out newer things, nothing is ever going to improve...

edit: spelling etc

http://www.hydrogenaudio.org/show.php/showtopic/203

BTW. we are not impressed by people who think that VBR 0 + no filtering is better than the pre-defined alt-presets which are made and testen by the people who know most about lame and audio quality.

Actually, Mark, I chose that setup on purpose, because every time i've encoded mp3s with filtering, and then tried some without to compare, the latter ALWAYS had a more open top end, which shouldn't be surprising! (I don't want to get all elite and pompous about this, simply: I engineer and produce music, I know the theory, and I bought these Beyer DT-100s for a reason). Anyway, the idea with those settings is that LAME should encode at the highest bitrate it thinks it needs to, because I don't care about filesize (but of course CBR 320 is a little OTT), I care more about preserving the source material as much as poss.

If you can point me to posts or test which point out otherwise, or at least attempt to, please do...

Digga: thanks. Perhaps I will give those changelogs a look at least.

And generally we don't care how good your equipment is. Humans like to they hear things when they try to compare to different encodes of the same song. As you probably know this is called the placebo effect. Unless you've done double blind (or ABX) listening tests (use search) then no one here is going to believe you can actually hear a diffeence between your command line and --APS. Makeing statements about quality without ABX testing violates HA.org's Terms of Service. It is your responsibility to provide evidence to support your claims. You may think this is a little harsh (I know you said you didn't want to go all elite and everything), but this is a fundamental value that this board holds that doesn't go away when it's convieniant.. In future keep this in mind.

I would highly recommend you use --APS with LAME 3.90.3 as per this boards official recomendation.

I posted my settings really so that you all know I'm using VBR, and therefore could tell whether or not updating lame would improve or worsen things based on newer versions' either improved or inferior VBR encoding (again, I don't know which is the case, only that 3.90/3.92 WERE recommended... are they still?).

Did I say I wanted to discuss the command line I use? No, that's off topic, and EDIT: tried --alt-preset standard and yes, lame tells me exactly what settings it's using. Filtering is unacceptable to me, okay? So, no, I shan't use it. (leaving it at that)

Would anyone else care to comment on the question I DID ask? i.e. Any point in updating EAC or LAME yet?

EAC: It works for you - so the answer is NO

Lame: Since the encoded files sound good to your ears I wonder for what reason you could want to update... Faster encoding speed? Smaller filesize at equal quality?

I'm sure you can check encoding speed yourself easily.
About smaller filesize at equal quality: In case you want comments related to this: I don't think anyone here has performed extensive listening tests with the settings you use that would justify making a general statement about this. Besides, depending on the music you encode mostly, your hearing and preferences, your results might be different anyway. The fact that you don't like the settings that sound best to most people here, shows that either your hearing is special or you haven't found out yet how powerful your immagination is. In both cases quality-related recommendations won't be useful for you .

Besides this, there's one change in latest lame versions that might justify updating: During encoding a replaygain scan is performed and the values are written to the encoded files. Some audio players (e.g. foobar2000) can use this information to avoid clipping on decoding and to get the same perceived volume on playback from all tracks.

edit: grammar

It is very possible that your commandline is actually preserving *less* of the audible signal than than --alt-preset standard. There are code level tweaks in the presets that cannot be replicated by any command line. But you are right, you are free to use whatever you want. I just plea that you do try some ABX testing.

Tigre:
> EAC: It works for you - so the answer is NO

Fair enough!


> In both cases quality-related recommendations won't be useful for you

See my reply re: Dreamliner's comment...


> Since the encoded files sound good... what reason you could want to update

I believe I did say that I'm happy with my current set up, just that it's OLD. And despite being very much in the "if it aint broke don't fix it" camp, with all the updates happening to both apps - and considering that they're both crucial to making my CD rips - I'm concerned that I might be missing out on even better quality ripping or encoding.


> Besides this, there's one change in latest lame versions that might justify updating: During encoding a replaygain

What I was really after, differences/improvements! Replaygain isn't something I need or want (I'll just use some outboard compression ), but it's handy to know LAME has it.


Dreamliner:

> There are code level tweaks in the presets..I just plea that you do try some ABX testing.

Okay, I've just tried A/B'ing a recording I know inside out: Model 500's "Milky Way". Nope, I can't tell the difference between APS and mine. Fantastic. But that isn't really the point - I know APS uses filtering and that isn't acceptable. Give me a high-quality alt-preset that DOESN'T use filtering (and perhaps contains some of these "tweaks"), and I'll be more interested...

Some things you might want to know:

Lame VBR has an upper bitrate limit of 320kbps. If you encode something at --alt-preset standard (with default lowpass) and watch the bitrate distribution you'll see that for some frames 320kbps is used. It's a known design flaw of mp3 as a format that > 16kHz content is very expensive to encode (= many bits needed). On some of those frames where --alt-preset standard with lowpass uses 320kbps, lame runs out of bits if you encode without lowpass, i.e. it is forced to use bits for 19-22kHz range that would be necessary to encode content in the audible frequency range. So removing the lowpass will lead to audible artifacts in some cases (this applys to other VBR/ABR/CBR settings as well). In most cases there won't be an audible difference though.

If it doesn't have to be mp3 (e.g. because of portable/standalone/DVD player support) I'd recommend looking into using some other modern lossy format that doesn't have these limitations, e.g. Musepack (MPC) or AAC or just use lossless (FLAC, Wavpack...)

Besides this, I'd really like to know why (lowpass) filtering isn't acceptable for you. Are you applying some DSPs after decoding (e.g. pitch shift) or similar?

Just my 2 cents: recent versions of EAC don't have the "manual TOC detection" feature, so upgrading to the latest version isn't recommended if you need to backup some types of copy-protected CDs.

...

Moneo: thanks, perhaps I'll steer clear of those particular versions, although the only CP CD I have, I have downloaded a rip of, since bypassing copyprotections is now illegal in my country (UK ). I take it that using apps like EAC to get around CP CDs IS technically breaking/bypassing the copy protection?

Tigre:

> Lame VBR has an upper bitrate limit of 320kbps

Yes and this concerns me as - again - whilst filesize isn't really a problem, because I don't have a DVD writer, or an OGG/FLAC-enabled hardware player of any sort (2 portable MP3 CD players though), lossless codecs are not really an option for the amount of music I like to have handy. I do use FLAC for keeping lossless copies of masters of music I've created etc. It consistently surprises me with its filesizes...

I have heard about issues WRT MP3 encoding of the top end of unfiltered audio, and I was rather hoping that setting it to use as high a bitrate as it needs (which is basically what my command line does, I believe), would mean this doesn't matter. Is this still an issue despite using joint stereo instead of real two-channel encoding? (in theory allowing more space for sample data)

The moment I get a hardware Vorbis player, this won't be a problem. Hopefully.

And WRT the lowpass filtering, it is purely a "personal requirement." I would prefer to keep as much of the top end in there as possible, not remove it entirely!

You know that you are unable to hear a 21kHz tone, don't you?

Yes, I'm quite aware of that, Gabriel. That is beside the point. Again, I chose that command line to make LAME go as high as it needs to when it feels it needs to, to keep those high frequencies. Many CDs are mastered with a LPF at the very top end to prevent problems during the final resampling procedure, so there should be no REASON to filter it any further... *sigh* This is waaay off-topic, but you should hopefully by now have some idea of my priorities WRT archiving audio.

Before you say it, I have long been considering other formats, and again I'm only a step away from ditching LAME altogether, once I have a hardware Vorbis player. Hopefully Vorbis encoding deals better with frequencies above 16kHz. And again, I don't have the kind of storage space required to keep everything in FLAC...

chis

What everybody is tring to say is that unless you go lossless you WILL loose something and that it is better to loose something that you can't hear compared to loosing something you would hear...

Scores of persons in this community have invested HUGE amount of time listening to diferent samples of music helping others to tweak LAME at its maximum capabilities (that is not to loose musical content, not KHz of bandwidth). Give a look at the sticky posts about adviced LAME versions and settings...

I personally don't think to be able to do any better, so I stick to their advices. If you think you can do better, that's your privilege...

Best regards

Sergio

That doesn't make any sense. You can't hear the lowpassing, yet because it actually does lowpass it isn't good enough. This is completely illogical. What purpose do you have for keeping this extra material. tigre has already warned you of what can happen in such circumstances. If you can't hear the lowpassing then you are potentially making your encodings worse by using your command line. Forgive our continual discussion of this topic, we're merely trying to get you to encode with the best possible LAME settings.

But remember, because of the inherent 320Kb limit, Lame can't always go as high as it needs to.

I did initially wonder if anyone would comment on the performance of that command line... let's just say I suppose I'm not surprised.

smz: Never in doubt about that - the removal of masked material etc I'm quite happy with, again it's just the LPF I'd prefer to steer clear of.

If I get the time, I'll try some more A/B testing between the two settings. I understand what APS etc are for, and why they are used... but I suppose I should move over to another codec entirely, the rate this is going!

Of course. It's like asking for trouble starting out with home made command lines here.

One piece of information which could be interesting for you: Since lame 3.96 the commandline you now use would map into a derivation of alt-preset (extreme or so I think). Upgrade or not is up to you - but you already heard that recommended version and setting is: lame 3.90.3 --alt-preset standard. Can't really say more than that wrt to LAME, right?

Oh, btw, you do have some arguments which contradict themselves. E.g. "don't care about filesize" <-> "320 kbps too high". I suggest you sit down and sort out your goals first, then it might get clearer.

Erik: I'm thinking about going over to --alt-preset high/extreme or somewhere thereabouts... and as for the 320kbits too high, I actually said CONSTANT 320 would be overkill. It's why I've always used VBR.

Anyway I think I'll close this now as I'm going to stick with the current version of EAC. I might change back to LAME 3.90.3 "permanently" as long as there are no issues with encoding truely gapless mp3s. That's the only other issue I have... and I vaguely remember that being something still under development when 3.90.3 was originally released...

well that STILL contrast :-)
don't care about filesize" <-> "320 CONSTANT kbps too high

3.90.3 deals with gapless just as poorly as any other MP3 implementation.

Well, not exactly. If the app can read the LAME Tag info with the padding values, it can give you a perfectly gapless result. Applications to include this are for example: foobar2000, Winamp with the in_mpg123 plugin or Burrrn.

I can't see that you got a clear answer on this question so I'll do my best:

Yes it is still an issue. There are problem samples where 320 kbps still isn't enough to recreate the audio without audible degradation. And as many other already have pointed out, if you keep high frequency content (which you can't hear) bits are spent on these frequencies which could have been spent on improving frequency range that you can hear.

I am sorry if you feel like we are pushing you, but we are actually only wanting to help. The presets have been tested so thouroughly that you can't imagine. You can trust them to bring you the best possible quality out of the MP3 format.