Hello, this is my first post

Sorry to ask a typical newbie question :

I use EAC to encode my CDs, with LAME encoder. I usually use --alt preset extreme or --alt preset insane (mostly). Encoding time doesn't matter, quality is very important.

I couldn't find out from reading the topics about LAME, which version I shoud use...

How will I get the best quality :
- with 3.90.3 ? (official ?)
- with 3.96.1 ? (better ?)
- with 3.97 alpha 10 ? (best so far, but alpha = issues ?)

I don't want to wait for "3.97 final", because I have a lot of CDs to encode in the next weeks...
So what is the best choice ?

Thank you for your help

There is a recommended settings and recommended compile thread for a reason...

Yes, but the latest update is august 2004 !.. so it won't help me to decide if I should use 3.97 alpha 10 or better 3.96.1 or better 3.90.3...

What would be your advice ?

I already gave you my advice!

Did you even read it? The posts I referred to do have comments about 3.96.1 and when it should or should not be used.

If you want "the best quality" go with lossless like FLAC. If you insist on using MP3 rip some tracks with 3.90.3 then with 3.96.1 on API, if you can't hear the difference use which ever one is faster.

I cannot give an advice, but here's what I am doing:

lossless: wavpack (..if you've got the space..; in the past I used Ape and Flac (chronical ordered))
mp3: lame 3.96.1 --alt-preset standard

I only use mp3 for not changing the CD all the time and it fits fine to my needs (hifi here, not PC-speakers)
Don't give too much on 'quality', if there's are real mentionable difference there's the next question if you (or the average-guest at your party) are able to hear it, this also depends on your playing-setup.
If you are paranoiac maybe use '--alt-preset extreme/insane', but I have never done this.
It's about music and this is mostly also fine if it is played a nuance more worst than possible - if you are sitting in the kitchen with friends you won't give sth. on quality, otherwise it will be an idea worth to recalibrate yourself

edit: typos

Of course, but I didn't find if 3.97 alpha10 is better for quality than 3.90.3 or 3.96.1 (sorry if the answer is obvious - I'm newbie in MP3 world) ?

Sorry but maybe I didn't search the right place or did not understand all : could you please give the link ?

http://www.hydrogenaudio.org/forums/index....showtopic=28123

read the small print

Thank you.

So just one question left : using --alt-preset insane and sometimes --alt-preset extreme, is 3.97 alpha10 better than 3.90.3, or is there no noticeable difference ?

You should try by yourself. There are few people on this board able to compare two different encoders at this bitrate, and nobody did it in the past as far as I can remember.

Oh my...

Use --preset standard with 3.90.3 or 3.96.1. The results should be equal, but 3.90.3 is slower. I doubt you can hear a difference with extreme/insane.
3.97 is alpha so if you want to use it for archiving, you better wait for the final.

OK, thank you all

My, some snippy replies in this thread. The choice is actually a little difficult at this point....

LAME 3.90.3 --alt-preset standard (recommended, but conservatively, by HA)
LAME 3.96.1 --preset standard (faster, quality is sometimes better and sometimes worse than 3.90.3, smaller files, not as well tested though.)
LAME 3.97 (alpha 10) --preset fast standard (faster still, probably better quality, but still alpha so bugs are possible.)

I am personally using 3.97 at this time, since it seems to me that, alpha or not, a lot of listening tests went into it and it seems to be the highest quality at the fastest speed.

If you are really conservative, and want to be absolutely positive nothing really bad happens in some weird cases (instead of mostly sure), use 3.90.3.

Jebus: It has nothing to do with conservatism. It has everything to do with sensibility and thorough testing. Currently, 3.90.3 is the best tested and has delivered proven results. This is why it is the current recommended encoder. Unless you can prove why you need a different version, use 3.90.3. Using anything else will give suboptimal results in terms of quality. Every other choice is a decision to trade off quality for speed.

If you are looking to get the best quality from MPEG Layer 3 audio (which was ostensibly never designed for true perceptual transparency anyhow), use LAME v3.90.3.

Thank you for this answer, it's clearer for me now.

things are going slow in the mp3 world nowadays
the poineer work for LAME 3.x was done 1999 to 2001, now it's mostly unattractive (but still important) tunings

Between 3.97 and 4.0, the LAME team (however populated it is these days) is busy working on making the world's best MP3 encoder even better. Worry not.

I am glad that xp_98's questions have now been answered. It helps not only xp_98 but other newbies who pop in here from time to time to try to learn something about encoding. xp_98, thank you for asking your questions and to the many who gave helpful answers, thank you too

Well even assuming you are right on the first point (and i'm not convinced... there was a LOT of testing on both 3.96 final and 3.97 alphas... and a year of heavy use with 3.96.1 by the majority of readers here), I'm not sure why that means that 3.90.3 is in fact for SURE better than the later ones, as you imply. If anything, 3.97 tested clearly better according to Guruboolez's recent tests (primarily), and frankly I trust his ears more than your opinion. But I did warn that there is a slightly higher chance of unknown problems with the more recent versions.

I certainly did NOT mean conservativism in the negative (neo-con?) sense. I meant conservative as in safer.

As for conservatism, in the past I have always held out for stable versions, but in the case of 3.97a10 I make an exception.

My usage history has been:

3.90.3 --alt-preset standard
replaced by
3.96.1 -V2 --scale x.xxxx
replaced by
3.97a10 -V2 --vbr-new --scale x.xxxx

Regards,
Madrigal

So if I'm not misreading you here, you acknowledge that it is safer, yet do not use it? That seems to be an exceptionally odd juxtaposition of beliefs.

If everybody only used what is considered "safer" because it is the most tested there would never be any progress. Innovation would come to a halt.
3.90.3 is great, but LAME continues to improve. In my opinion, Jebus has hit the nail on the head in his expanation of the current state of affairs.

That's just wrong. Lame 3.90.3 was tested in 2001-2002, and the fresh-builded --alt-preset standard appeared to work better than 3.89 --r3mix for similar bitrate. That's all. Claims based on very old tests have no validity anymore. All we know is that 3.90 is better than 3.89. Point. It's like saying that DivX 3.11 is the best MPEG-4 ASP encoder over DivX6 or Xvid 1.1, or that PsyTel 2.15 is the most pertinent AAC solution. Yes, it was true three years ago. But it doesn't mean that such statements are still true, just because a band of lazy people prefered to parrot blindly some recommendations rather than spending some of their free time to submit any feedback to active developers.

Did you ever tested 3.90.3 over 3.97, as it was requested by lame developers? I can't find any results. It's simply very rare to find feedback on HA.org these days... Then how could you prove that encoder x is superior to encoder y when you don't perform any blind test? I have tested several alphas of 3.97 and also 3.96.1 on many aspect and a big variety of sample. And based on my experience, following HA rules, I can say that 3.90.3 is not only slower but also have quality issues partially corrected by 3.97 alphas which makes latest alpha preferable (but maybe unsecure) for people looking for the best possible quality to any older lame versions.

Don't spread wrong information on this board please, or submit your test results if you have interesting things to reveal about 3.97 possible flaws. Gabriel & Robert are still working on lame, and that's the moment to submit all known issues.

what is --scale x.xxxx?
if one has the disk space why not use --alt preset extreme? by being conservative do you mean that extreme might not work as smoothly as standard?

also how does one prevent EAC from crashing if the ID3 tags are not fully populated (freedb.org).

i am using....
--alt-preset extreme --id3v2-only --pad-id3v2 --ta "%a" --tt "%t" --tl "%g" --ty %y --tc "EAC LAME 3.90.3 Extreme" --tn %n --tg "%m" %s %d

BTW what is the advantage of using wapet? with say the following command line.
%d -t "Artist=%a" -t "Title=%t" -t "Album=%g" -t "Year=%y" -t "Track=%n" -t "Genre=%m" lame.exe --alt-preset extreme %s %d

sorry for being a diffcult but i want to archive 600 CDs and want a setting that is future proof (when bigger 4GB Ipod shuffle come out for example) or when playback systems can diffrentiate between APS and APE.

Then, by your standards, the administration itself is spreading wrong information.





I'm merely attempting to help a new user figure things out, and reiterating the standard Hydrogenaudio position on these things. I suggest you take up your quarrel with the administration, not me. I've yet to hear from anyone whose opinions and ears I respected (other than yourself) that stated 3.97 is to be preferred over 3.90.3. The artifacting at those bitrates is beyond my ears; I acknowledge this fact, and have ABX results that show that I very likely can't hear a difference.

Read this (especially Post #13).

Then download WaveGain here, calculate x.xxxx by running your wav files thru WaveGain using 'Calculate Only' mode, and add --scale x.xxxx to your LAME command line of choice.

Regards,
Madrigal

See Madrigal's post above.

Also, these threads may be of use:
http://www.hydrogenaudio.org/forums/index.php?showtopic=3440
http://www.hydrogenaudio.org/forums/index....showtopic=34011
http://www.hydrogenaudio.org/forums/index....showtopic=34130


Why not indeed? However, if disk space is no issue, why not use lossless?


http://www.hydrogenaudio.org/forums/index....showtopic=34170


The above uses WAPET to add APEv2 tags to the MP3, instead of the usual ID3.


I personally archive to lossless (APE), and then transcode to MP3 from there. I would worry less about playback systems and more about your hearing quality. Encode a few well known tracks to APS and APE and see if you can spot the difference.

I would definately recommend archiving to lossless though, and transcoding for portables, etc.

ok now what is lossless? is that APE? I thought APE was Alt Preset Extreme? I thought all MP3 compression was lossy including APE or even API (Insane).

However for my purposes APE will suffice. I am looking for a 5:1 compression (from 600CD, 9000 songs, about 420GB to about 80GB). I assume APE as about the same compression as 256kCBR (maybe a bit better) but with better efficiency. I could be wrong.

from what i know 128k is about 10:1 comprssion (a 700MB Cd will fit in 70MB).

I intend to store my MP3s on a 120GB harddisk. Keeping the harddisk 1/3 empty (using 80 of 120 GB) will allow for defragmenting the drive when required.

MP3 is a lossy form of compression, as you know. However, MP3 is not the only format in the world and there are other lossy codecs and also lossless ones. Here's the Hydrogenaudio Knowledgebase article for Lossless for you to read. There are quite a few lossless codecs, so here is a comparison page for you to browse.

You mentioned how you wanted your rips to be future-proof. Lossless is the perfect way to go if you have the space, since the quality of your encodings will be identical to those of your original source (compact discs), which obviously isn't the case with lossy codecs such as MP3.

APE is the file extension (.ape) used by a lossless codec called Monkey's Audio, so people often (wrongly) refer to it by saying just APE. You are correct in noticing that this is also the acronym used for one of LAME's presets, extreme.

APE is the file extension used by Monkey's Audio files. My fault for using APE as both a reference to Monkey's Audio and to --alt-preset extreme in the same post. Sorry.

Lossless files are basically compressed WAVE files, with no loss of quality. We term codecs such as MP3, OGG, AAC, etc. as lossy. Codecs such as Monkey's Audio, FLAC, WavPack, etc. are lossless.

You won't get 5:1 with lossless though - more like 3:2.

Edit:


Hmm... I'm not sure it's strictly wrong! There are many abbreviations that clash unfortunately. Hopefully "APE" as "--alt-preset extreme" will be defunct soon, and we can just use "PE" to abbreviate "--preset extreme"... if we must. I think referring to a file with the extension of ".ape" as "APE" is perfectly acceptable - just as we refer to PCM WAVE files as "WAV" or MPEG-I Layer III files as "MP3".

is lossless compression of 5:1 possible?

In a word... "no."

Check the lossless audio comparison link above your post.

If it's entirely full of MP3s and its content isn't being regularly changed, you need no defragmenting. Defragmenting is extremely overrated, and this is something I've verified myself.

Defragging the OS/programs drive or partition can certainly be helpful. The benefits of defragging a drive full of mp3s is probably negligable, although in a data recovery situation, it would be imperative.

The only time I've ever had it actually give any kind of noticable performance gain is after lots of additions/deletions with very little free drive space for an extended period. If the pagefile gets fragmented, it can load more slowly. Few OS features are changed after initial install, and most programs are installed before the hard drive becomes filled.

Very few people actually meet the criteria that are required for defragmentation to be useful, and the people who would need it generally know why and when it's needed to be done. Recommending it as a panacea (as is done by far too many people) to people is nearly useless with Windows XP.

Actually in theory defragging a drive formatted with a filesystem such as NTFS or FAT is extremely necessary (as compared to say, ext2/ext3 for Linux) due to the way it allocates space. That being the case, and those filesystems being the most pervasive by far, most hard drive manufacturers now include a fireware sort of abstraction layer, basically ignoring what the OS thinks the drive is doing in favor of something much more hardware-specific. So you're only really defragging what Windows thinks the drive looks like, not the actual layout of files as the drive heads see them. THAT is why there is not much performance gain to be had these days.

Mind giving me a source for this info? I've never heard that before.

stupid question: why does one need APEv@ tags instead of the Usual ID3. remember my Ipod should be able to read the tags so I would think that ID3 tags would make more sense in my case. Right?

anyway after all this wonderfull feedback I now got a command line that looks like this:
--alt-preset standard --id3v2-only --pad-id3v2 --ignore-tag-errors --tt "%t" --ta "%a" --tl "%g" --ty %y --tc "EAC LAME 3.90.3 extreme" --tn %n --tg "%m" %s %d

should i use 3.90.3 or 3.96.1 with the above command line. Is there any part of the command line 3.90.3 wont be able to resolve (--ignore-tag-errors perhaps).

Yes. APEv2 tags are used in place of ID3v2 tags, as some people believe that ID3v2 is evil. ID3v1 tags are occassionally inadequate due to the 30 character limit. With my MP3s I use LAME's default tagging system, which uses ID3v1 where possible, but adds ID3v2 if ID3v1 is insufficient. I think that rocks.

I don't fully understand the ID3v2 legs bad viewpoint. I assume it's only bad if poorly implemented - but I trust LAME to implement it properly.


You will need to use 3.96.1. If you read the relevant thread I pointed you to previously it details that --ignore-tag-errors won't work with 3.90.3.

I, and 70% of the HA voting public, am using 3.96.1.

thanks so much for taking time to help. i am new to MP3 and age is not in my favour either. The reason I asked this is because i hear that 3.96.1 is only better (than 3.90.3) at lower bit rates and I am thining of using --present extreme.

HA is a huge archive. By the time I finish reading some threads I forget what was said on other threads.

Could you give a link? I'd like to see where the HA administration claimed that "Using anything else will give suboptimal results in terms of quality. Every other choice is a decision to trade off quality for speed."?
Dev0 recently said that 3.97 will be the recommended version, because current tests are more than promising. Again, read carefully recommendation, and don't spread wrong information about current quality and current HA.org position about LAME. Thanks. If you have something interesting to claim, do it in lame alpha testing thread. I recall that the only contribution you made is this one. You're not very qualified to make claims about quality of various lame version.

Logical inference.

The current recommended compile is 3.90.3. Recommended compiles are chosen solely on the grounds of proven audio quality (to the best of my knowledge). Therefore, the recommended compile provides greatest audio quality, according to the administration.

The subsequent versions of LAME were optimized for speed. Therefore, subsequent LAME versions are faster than 3.90.3. These speed optimizations caused degradation of audio quality, resulting in builds that produced poorer audio quality, and requiring retuning to provide quality equal to 3.90.3.

So, the rationale for using versions of LAME produced after 3.90.3 is to have speedier MP3 creation, and the statement about the tradeoff between quality and speed was created.

That is how I reached that conclusion. Mind picking out the fallacy and/or error? When the recommended compile page changes, I will change my opinion accordingly. I've acknowledged that I am unable to ABX at the bitrates given and am going solely on the opinions of others.

However, currently, members should not be recommending versions to new users that conflict with the Current Recommended Compiles. If you have issues with the current recommendations, please discuss this with the administration so appropriate changes can be made to that post.

I am told that 70% of HA is using 3.96.1. besides 3.90.3 cannot handle the --ignore-tag-errors switch and hence can hiccup if the tag data is not part of it's database.

I was given to undersand that some subsequent versions of LAME (after 3.90.3) were not ONLY optimised for speed but also kept or enhanced audio quality. 3.96.1 is one such version. From this I can only infer that 3.90.3 is probably the version that is least likely to hiccup (except for the above mentioned switch). Maybe I am wrong.

If one could suggest how to get 3.90.3 to not get tripped up by tag errors I would be grateful.

I can't believe it. It has been explained several time. Reason is also explained on the recommended compile thread you've mentionned. Are you blind?

>3.96.1 seemed as though it might be about the same quality as 3.90.3
> in some of the tests done after it was released.

The "administration" doesn't claim here that 3.90.3 is better than modern lame release. But Dibrom (or another administrator) followed:

> However, in the minds of the administration of this forum it has not yet
> been tested extensively enough, thus 3.90.3 is still the recommended version.

Isn't it clear enough? Or can't you read? 3.90.3 is recommended because tests were done by some people in the past. And not because quality is proven to be superior to any other lame version.
It should be clear enough.




False. Gabriel and Robert worked to improve quality as well as encoded speed. All recent tunings made on recent alphas had no impact on speed but on quality. Read changelog if needed. And if you can't understand it, just count the lines coloured in red...



Good Lord? What's this nonesense? LAME is not GOGO... To improve speed and lower quality in the same time, it should be easier to make -q 9 as default. I wonder why developers were irrational enough to spend so many hours to work with various switch in order to achieve the same results.



For someone who can't ABX difference between 2 release of lame, you have many things to tell...



Did you read HA TOS? Conclusions and claims must be based on listening tests, and not on syllogism or phantasmas.
All modern encoders were able to gain speed and quality in the same time: Vorbis, MPC, AAC [Nero, faac and QT]. And nobody on this board conclude that higher speed necessary lead to lower the quality. Except maybe some people with fertile imagination but no free time to contribute to help on encoder's development.



Then read all tests that were performed on lame during the last year. You'll find several evidence that modern release of LAME are all except "suboptimal" compared to 3.90.3. I've spent several hours (probably more than 100) to test lame, and I've published most results on this board. Some other people really interested about quality and progress also spent some of their free time to give positive or negative feedback. You should better do the same, rather than polluting this board with unfounded equations about quality and biased interpretation of HA.org recommendation. And if you really don't want or simply can't work to improve LAME, read at least the informative tests that are posted on this board: you'll probably learn interesting things and therefore stop spreading wrong informations to other people. HA.org as well as LAME developers need testers and feedback, not parrots and imagination.

mmmm... well I think members of HA can recommend what they like as long as they give adequate reasons/evidence for doing so; this may then prompt others to test and then agree or disagree (such is life eh?).

In this instance, 3.96.1 appears to have proved it's worth, but has not been recommended over 3.90.3 for many reasons that have been endlessly discussed. guruboolez's, and other's, tests on the 3.97 alphas are evidence enough for me to use a 3.97 alpha (alpha 10). I know the risks in using an alpha as well as it not being recommended by HA. But I store my audio as FLACs and then transcode to whatever I like, so perhaps my situation is not as important as someone who is looking to rip x number of CDs to mp3 for the foreseeable future. But whenever I've given advice on this subject I've always stressed that just cos I'm using an alpha doesn't mean they should.

Anyway, my 2 cents.

Peace, love and unity

I think anyone who wants to make a collection of mp3 from 600 CDs would probably be a music maniac (no hard feelings! actually, i am one of them owning a collection of 10000 mp3s!)

For any music collector (mp3 music of cource)

->Quality Issue
(hence using either extreme or insane preset)

-> Use of 5:1 compression
(insane is cancelled, bcoz that would be 4:1, too large;
"ABSOLUTELY HIGH Q MP3" = "U R INSANE"
it means why to use mp3 then? use Monkey's Audio!)

-> The time that will be spent ripping 600 CDs
What I have experienced is that 3.96.1 Extreme is equivalent to that of 3.90.3
then why not use 3.96.1?
(Of course! this is my opinion! May not be the recommended one!)

So i would suggest combination of EAC + Lame 3.96.1
and using --preset extreme

use of extreme will yield best possible quality (VBR).
JointStereo is safer
speed of encoder is also good (without sacrificing quality)
so, i think this is a practical combination for a music collector

i am using this combination myself and i am satisfied with it.

I haven't ever compared 3.96.1 with 3.97 Alpha

Such opinions are not welcome here unless backed up by fact. Read the HA terms of service.

While these posts are incredibly passionate, I thought I would add my $.02.

I am in the process of ripping a collection of a similar size to yours. My hard drive was 160GB. Ultimately, what I realized is this is an extremely time consuming process. As such, I changed my plan from ripping EAC/Lame to EAC/Flacattack.

I purchased a new 400GB hard drive for $220 on eBay and encoded both .flac and .lame at the same time.

The reality is a newer, better format, will always be developed. 3.90.3 vs. 3.97 debates (thank god) will someday end. I would ask yourself are the next few months of your time not worth a couple of hundred bucks?

My current compression results are as follows (per 100 discs):

7.5GB per 100 discs using 3.90.3 --alt preset standard
30GB per 100 discs using flac 1.1.2 --best (8)

Good luck with your project and remember, this is supposed to be fun. You are not the Anti-Christ if you choose 3.96 or 3.90. I assure you, no one will die.

Oh, BTW, take a break during all these posts and actually enjoy some music.

Brent

Great advice !

Regards,
Madrigal