The question is, am I doing something that sacrifices a lot of audible sound quality?

I recently got an Ipod and am trying to make a good judgment about the Itunes MP3 settings to use.

I know the common practice here is to use LAME and I am convinced by the objective data that it is the best MP3 compression method for high quality / file size ratios. However, I am trying to use the Itunes MP3 encoder to import CDs for reasons of convenience and simplicity. I am willing to use a little extra hard drive space to get the same sound quality with Itunes as I would with lame at a higher rate of compression. I just want to make sure I am not doing anything terribly dumb (aside from using the Itunes enocdoer in the first place!).

So here is what I am doing: For low sound quality CDs (generally 1960s or earlier), I am encoding VBR, highest quality, joint stereo at the minimum of 160 kbps setting. The rate for a song is generally around 170-175 kbps. For high quality CDs I am enconding VBR, highest quality, normal stereo at the minimum of 192 kbps setting. The average song encodging rate seems to be around 200-205 kbps.

I am concerned that the Itunes MP3 VBR codec might do something funky I don't know about. So far my ears tell me things are going pretty well, but I have neither the ears nor knowledge of many people here.

I am interested in sound quality, I have a nice headphone setup, but I use a common sense approach. If it would take very strict A/B listening for me to notice small artificacts, I am not concerned about that. My rule of thumb is that I want to enjoy the encoded sound as much as I would enjoy the sound on the CD -- crisp highs, rock solid bass, full details, etc. I am not concerned about a tiny audible artificact in an unusually complex passage.

So do my settings sound reasonable for what I am trying to accomplish? I am willing to up the bitrate or change encoding methods if it will improve sound quality to a noticeable degree.

Edit: Also, I use the "smart encoding" and cutoff below 10 hz options settings in Itunes. Are these harmless to sound quality?

Thanks to anyone who took the time to read this and offer suggestions or helpful advice.

Regards, Steve

I'd just use 192kbps CBR with stereo, alas the standard "higher quality" setting. In any case, stay away from joint stereo on higher bit-rates than 160kbps and only use VBR if you encode at ~224kbps or more.. But again, the quality should be very good at just 192kbps.

EDIT: Since you have an iPod, you should use the AAC encoder if you want high quality. 192kbps should be transparent with this codec, perhaps even 160kbps..

Thanks!

I am using MP3 to maintain compatability with other devices, so I'm sticking with MP3 I think.

Of course, I have questions:

What does using joint stereo do to the sound quality so that I shouldn't use it above 160 kbps? Is it an audible degredation?

Why should I stay away from VBR at less than 224 kbps? Does it sound worse than CBR at the lower settings with the itunes encoder?

Thanks again for your help.

AFAIK, the joint stereo mode isn't lossless, so the stero imaging can very well suffer if you use it on high bit-rates (I think the smart encoding setting choose normal stereo on 192kbps and higher even if you select joint stereo).

The VBR mode does a pretty lousy job on low bit-rates, see Roberto's listening test and see how badly it did there So if you use it, remember to at least select the highest VBR setting (they put in no less than seven quality levels)..

Since it uses mid-side and not intensity stereo, it should not affect stereo imaging. Probably nobody knows how well JS is tuned in this encoder, but as it defaults to plain stereo already at 192, it seems its creators didn't have much confidence in its abilities. I guess its performance can probably be compared to older Lame versions before "safe joint stereo" was introduced, which also defaulted to stereo at 192.

That's likely true, yes.

The problem with an encoder like iTunes is that no one really knows how it works, so to be safe you should probably just stick to the standard default settings.

Hey thanks again Buseman.

Is this the listening test to which you refer?

http://www.rjamorim.com/test/mp3-128/results.html

That was from January 2004, with VBR set at 112 kbps using Itunes MP3. Roberto seems to have admitted to making a mistake in using that instead of 128 CBR. So you're right, this seems to show problems with Itunes VBR at these low bitrate levels.

Is there any objective data or listening-based anecdotal evidence that more recent Itunes (I'm using v 4.7) MP3 VBR at the highest level at 160 kbps and 192 kbps minimum settings is worse than the itunes CBR at those higher levels?

Also, thanks to sTisTi for the clarificaiton on the joint stereo issue.

Thanks once again for your help, I'm learning a lot.

Regards, Steve

January 2004, matter of factly. I'm afraid iTunes didn't exist in 1993

Umm, it cannot be from 1993, because itunes didn't exist at that time - i guess you mean 2003 instead :-)

The problem is this: with a good tuned encoder, VBR and JS should almost always be better - but with a badly tuned encoder, the opposite applies. iTunes isn't a well tuned encoder like lame - they probably tried to make it sound not too bad, and then stopped improving it - why should they care? They want to sell AAC, not mp3.

I don't know if they improved it since then, but the bottom line is: If its not too much inconvenience, then use lame with APS/V2 for the 190kbit range, instead of the iTunes MP3-Encoder. As you already correctly sumarized in your initial post, your situation is a quality vs. convenience decision.

- Lyx

edit: rjamorim was faster - the test was from january 2004

LOL.... my brain didn't keep up with my fingers. Yeah, January 2004, not September 1993. I'll go back up and change my post.

So what I am reading is that nobody knows how good or bad the VBR implementation is on recent itunes MP3 codecs at these bitrates (160 and 192 kbps). Further, as Itunes AAC is better, and LAME MP3 is also available, no one has reason to examine the issue too closely or objectively test it. So if I want objectively assured sound quality I can go AAC and get something quite close to transparency at 192 kbps and still have the convenience but loose the compatability with other devices, or I can go LAME and lose some convenience but keep the compatability with other devices.

What do you guys think, will AAC become more or less universally supported in the near future?

Exactly how convenient can Lame be? Can you use it with a program where you can you just put in one CD after another conveyer-belt style and have it download the MP3 tags and rip automatically and eject the CD and have the tray open and waiting for the next CD, as you can do with Itunes? I want the ripping to be as painless and mindless as possible. The convenience is a big deal to me, I have a lot of CDs.

Thanks again for your help, everyone.

While EAC can be very difficult to do the _one-time_ initial setup, it is VERY easy to use afterwards. So lets asume that you don't mind spending some hours for a one-time setup(including research and gathering info) as long as the everyday-usage is simple. Then everyday usage looks like this:

-> You insert a CD (all tracks are automatically selected) and press a button. Done. <-

What you will get with the above simple usage:
- the best mp3 quality possible at medium-high bitrate ranges
- absolute safety that the tracks were ripped without errors (if anything went wrong, EAC will tell you - other rippers don't do that)
- automatically get tags from freedb and add them to the encoded MP3s
- (optionally) automatically remove leading and trailing silence of tracks
- real gapless playback with players which support it (currently, thats fb2k, winamp(with a plugin), xmplay(??? not sure) and a small amount of portables)

I dont remember if EAC has an option to automatically eject the cd after ripping. Someone else may be able to give more info on that.

- Lyx

Steve999, are you on a Mac or a PC?

I'm on a PC. Thanks for the replies everyone.

I'm getting 8x to 12x rip speed with itunes.

Is this realistic with LAME / EAC or other lame implementations?

Yes, if you select the same rip-method as the one iTunes uses (burst) - however, then you will not know if the rip was error-free. So, it can be as fast and insecure as iTunes, or slower but safer.

- Lyx

The iTunes encoder has seen no improvement since 2001 i believe, and I would take a wild guess they won't do anything with it in the future either. When that said, I don't think it is as bad as this board seem to portray it as. Yes, the Joint stereo mode isn't safe to use, and VBR is a bit shaky, but if you encode at 192kbps or above, it should give good results. Sound & Vision tested it a while back and they gave it good scores above 192kbps (not so good on lower bit-rates). Since you will encode at high bit-rates, I reckon you should be reasonably safe.



I think it will be better in the future since it got powerful players like Apple & the MPEG group behind it. It's certainly a convenient format to use, since it encodes as fast as its mp3 encoder yet still gives noticeably higher quality. Unless you have some other devices right now that don't support it and you are going to use iTunes anyway, AAC is probably the best way to go

If your CDs are in perfect condition and without scratches, you can use "synchronized" mode in EAC and get the same (or better) speed and safety like in iTunes. With scratched CDs, it is advisable to use EAC in secure mode to ensure error-free rips. However, this mode slows down the read-out. How much depends on whether your drive uses C2 error information and caches audio (you can test this in EAC). If it does not support C2 information and caches audio, ripping will be pretty slow. What drive are you planning to use for ripping?
If I were you, I'd really use Lame and a secure ripper like EAC. This way you lose a bit of convenience but can be sure you have perfect rips (and best possible MP3 quality) which you can enjoy for a long time without regrets.

I'd also go with EAC and LAME 3.96.1 ... but I think there's a plugin that allows you to use LAME + iTunes...

AFAIK it only works on Mac-iTunes

Jojo in "i'd use lame" shocker



Only woks on the Mac version, and included 3.93 as default the last time I checked.

That's right. It uses apple scripts so it should only work on a Mac. You can find it here:

iTunes-LAME

You also have to have LAME installed on your system separately to make it work. I used it before I discovered cdparanoia on Linux (or EAC on PC).

Check again : http://blacktree.com/apps/iTunes-LAME/
You can choose between 3.90.3 or 3.95.1
and here's a guide for upgrading to 3.96.1

Edit: too bad it's for mac os only.

BTW, Here's an Apple link to the Blackwell software (but for Mac OS X):

http://www.apple.com/downloads/macosx/ipod...ameencoder.html

The latest version of iTunes-LAME includes LAME 3.96.1 but one can install another version if they want.

Umm... the plugin is only for Mac OS X, it's just that he was refuting the claim that it was using an old version. Since you replied that you use a PC, the plugin will be of no use to you.
Also, for Windows users those who do not use Mac OS X, the lame build system is the same as that on other unices, so Mac users are less reliant on needing binaries or DLLs. It's very easy to upgrade the version of lame used by the iTunes plugin.

Oh.

Back to the same old quandary.

Most people want to encode their CD's into mp3. If iTunes got a reputation of having a lousy mp3 encoder, people would be hesitant to use it and look at the alternatives. If it really was bad, you can be sure they would still work on it.

Same thing with Lossless. They're not selling those files anywhere, but it is still important..

Even people who generally do not care so much for audio quality complain about the lousy iTunes-MP3-encoder. It's hard not to hear the artefacts at 128 kbs ... Apple does not care!

Well, maybe i expressed what i meant a bit unclear. When i wrote "sell", then i didn't mean it literarilly - but instead more symbolically(like, they want to market(sell) the AAC-Codec). People can also rip and encode to AAC with iTunes.

Still, you have a point with what you said. However, even if the itunes mp3-encoder is less bad than its reputation, then i still asume that apple have no interest to create a good mp3-encoder - a bit like "if it sounds okay to the average user, then thats enough".

IMO, the point is not that Apple doesn't care about their MP3 encoder, but they have limited resources and it makes much more sense to tweak their AAC implementation compared to their MP3 one.

Still, does anyone here hear artifacts at 192 kbps with the itunes mp3 encoder?

Musicmatch seems to take pride in keeping its mp3 ripper up-to-date... might it be a little more reliable, higher quality, of a known level of quality?

Thanks again everyone, it's been very enlightening! I think what I am learning is that at higher bitrates people haven't done much objective research with medium-level mp3 encoders, so there's some guesswork involved.

This is because the majority of people generally fail at such high bitrates and therefore cannot help much in "tuning" such bitrates. It may also have to do with the fact, that ABXing at such high bitrates is very stressing and exhausting - if the testresults dont even contribute in improving the encoder in question and the chances are high that another encoder is better anyways, then motivation is very low - humans are lazy

- Lyx

I wouldn't worry about it too much if you are using those bitrates. With iTunes, the VBR settings use a specified bitrate as the minimum bitrate (I think you already know this from what I have read). But that means that you are only going to go up from there. It's not as efficient as LAME's VBR, but by definition your 192 VBR file in iTunes will be better than your 192 CBR because it will definitely be of a higher bitrate than 192. So it's safe, given that iTunes is what you want to use.

But you probably won't be able to hear the difference. I'm probably violating some TOS by saying that. That's why I try not to post much.

i think your clean and clear w/ that statement. its just your opinion.

edit: typo

It's impossible to prove that one cannot tell the difference, only possible to prove that one CAN tell the difference. Thus, no one can request an ABX from you or something like that. Of course, that doesn't exclude the possibility that if one posts BS, then he can get quite a lesson from people who know better :) And, from my experience at ha.org, its impossible to avoid that this happens to you from time to time. It also happens to me. Humans error. The difference is how often it happens and if one learns from it :)

- Lyx

Thanks, I found that very reassuring.

This morning, dying of curiosity, I did a sighted volume-matched A/B swithing between the original CD and the ipod with a well-recorded acoustic jazz CD (Kenny Garret's "Standard of Language") through a nice pair of Sennheiser headphones and a solid amplifier, which was enough to convince me, at least for that CD, that upon listening closely I could not tell the difference between the original CD and the ipod with 192 kbps itunes vbr "highest quality" setting MP3. They sounded like dead ringers to one another is what I thought to myself. Of course, I don't know what the dreaded artifacts sound like, so I could have missed something....

Hope I didn't violate any TOS... I am just saying that subjectively on a sighted, volume matched A/B test with one particular high quality CD it didn't seem to me that I could tell a difference. Even setting up the sighted test was a chore.