I got an iPod last week and have spent the last week just engrossed in trying to nail down what encoding method to use.

When this process began, I started with the iTunes default (128 kbps AAC - ). I moved back to the recommended mp3 standard (-V2 Lame, via EAC). I've tried several other settings and nothing seems to make me happy.

My unhappiness has ranged from my being able to hear artifacts to the fact that the encoding process was taking forever. I have 300 + cd's to rip...

So I tried to just cool it for a day and give it some thought. I did some reading about iPod's, battery life, perceived transparency, etc. I went back to the drawing board with the following goals:

- Stick with mp4 (AAC) because it is well supported by Apple
- Find the file size where transparency is (at least, perceptually) achieved
- Keep my files under 9 MB (everything I've read hints that files at 9 MB or over cause the iPod HD to spin-up); thus shortening the battery life

After having said all that, I settled on 224 kbps. I encoded a wide range of music, from different decades, mono and stereo, different styles. 224 kbps seemed to do the best job at the lowest bit rate (and file size), to my ears.

What I want to know is, has anyone else settled on this as their choice as well?

I will be using my iPod for listening at work (I have some decent Sony earbud's) or in my Camry.

Did you verify (with an ABX test) that you can hear artifacts at lower bitrates? Listening tests show that AAC(Itunes) ahieves transparency at 128kbps most of the time.

Absolutely. In my car, I can hear the artifacts. It's probably more of a 'placebo affect' in headphones but in my car, anything under 224 sounds (I'll hold off on the hyperbole) not so good.

I'd imagine the bitrates needed for transparency during work or travel would be a good deal lower than when listening at home.

It's probably a placebo effect in your car too.

As jmartis suggested, you're best served by doing multiple ABX tests with different AAC bitrates to determine what is/isn't transparent to you.

There's nothing wrong with 224kbps if if gives you piece of mind, but if say, 160kbps is just as transparent to you, going with the lower bitrate will take up less space and probably give you better battery life.

This thread is relevant and an interesting read:
http://www.hydrogenaudio.org/forums/index....showtopic=45644

Quite so. At one point I had delusions of greatness and I thought that I need at least 256 kbps for my music to sound good. Some time ago I simply bit the bullet and performed an exhaustive series of ABX tests of Nero and iTunes AAC codecs.

The result? Both Nero and iTunes AAC are transparent to me with normal (i.e. non-problem sample) music at 128 kbps. I may have lost my dignity but I've saved hundreds of megabytes of space encoding my music at 128 kbps AAC

Additionally, noisy environments like cars really aren't the best spots to conduct an ABX test. If you wanted to conduct an ABX test in your car, you would have to hook your computer up to it, drive your car out in the middle of no where on a calm day, and then conduct the test. Still, cars are not that acoustically tuned so music will be bouncing over many objects in your cars thus reducing quality even more.

Use your PC in a quiet environment, like your room, along with your best set of headphones. That is the best environment. Using headphones and conducting a ABX test will eliminate the placebo affect. Just skipping around tracks in your car (or when using your headphones) is an example of when the placebo affect will come into play.

I doubt it.

In all my years I have never heard a car stereo I would call "accurate". Fun, perhaps - but the small space renders most listening into an odd hybrid of headphones/room acoustics with very little real fidelity. Now, perhaps if you had a car system that had an insane amount of high end boost PLUS your hearing could still detect information at around 18kHz PLUS you played recordings with a lot of energy in that band - you might hear some tiny difference between 128kbps and 224kbps until all that high end boost numbs your ears. Otherwise I give it a probability of zero.

The other posters are right, real ABX is the only way to tell for sure.

I used to design high end car stereo components (preamps, crossovers) for Phoenix Gold, and while loudness and bass boom were certainly valued actual fidelity was completely ignored. The market dictates this for car stereo, and as a result none of the engineers bothered to install the systems in their own cars - our most senior design engineer stuck with his factory cassette/CD decks, as did I.

That's what I get for having no interest in hip-hop or dance music... c'est la vie.

It's easy to imagine that the highly colored sonic environment of a car could unmask frequencies that would be masked in silent headphone listening. I've noticed artifacts in my car with LAME -v 4 that I don't notice at the office. Or maybe my car stereo just sucks.

I only have one suggestion, if you are using the iTunes AAC encoder, try encoding with the VBR switch on, it is a quasi-VBR mode that may help improve quality at a lower average bitrate, i.e. 160 Kbps VBR averages around 167 Kbps, which is less than say 192 Kbps CBR, but will probably be very similar quality.

Seriously, do some ABXing. Find that magical transparent bitrate, then encode at 15% higher (just to be safe, because you can never unhear an artifact in your favorite song).

If you have an incredibly small music collection then by all means, encode at a ridiculous bitrate. But if you're like the rest of us, you have way too many songs to be too stingy about quality, at least if you're going to listen to it from the speakers of a Camry (I own one myself, and it's quite possibly the worst system I've ever heard)

EDIT: Fixed some grammar that made no sense at all.

Ditto.

And I think I know why it is - the bass/midrange speakers are in the doors, the treble speakers (tweeters) are in the dash board. Both have horribly uneven frequency response. So you have spatial separation of sounds that are supposed to be together, and random severe EQ changes. It's no wonder that some artefacts become more obvious.


However, I have serious doubts about this being an with high-ish bitrate AAC, unless the tracks are exercising that iTunes AAC bug.

Cheers,
David.