A few months back, I decided to back up my CD collection to WMA9 160kbps CBR files. I'd looked at quite a few options including MPC and Nero AAC, etc., but ripping with WMP9 straight from CD to WMA9 files was too super easy to pass up, and the quality (at least for my ears) was excellent.

I'm a musician and I've always thought that I have a pretty good ear for sound.

Tonight, just for fun, I thought I'd compress Pulp's "Different Class" to MPC using the -xtreme -xlevel parameters with v. 1.14. I expected to hear a fairly discernable difference better these 200+kbps VBR MPC files versus my less fancy 160kbps CBR WMA9 files. I was really surprised, after carefully listening to a good cross-section of the files and comparing the two formats, that I could not hear a discernable difference (using: SB Live! card, Sansui AU-222 amp, Sony MDR-V500 monitor headphones).

So, either I don't have a good ear for sound, or there isn't much of a difference there at all, unless you're listening on a high-end audiophile system of course. I'm willing to concede that maybe these particular songs are simply not complex enough to let MPC shine, but what songs would be?

Am I just a feeble old bum, or do I have a point here?

Spaceman

If you can't hear a difference between 160 kpbs CBR WMA9 and MPC --xtreme maybe you should see if you can get by with --standard and save yourself some bits. Of you could also keep using WMA if it works for you. I really do recommend VBR over CBR so even if you use WMA maybe you should try a VBR setting. What's most important is how it sounds to you. If you don't need as many bits as someone else to reach transparency then I wouldn't worry about it. You'll just save more space.

That being said, MPC is (still) considered the best alternative (except from lossless) if you want to back up your CD's. At least if you want to go for transparecy at decent bit rates.

If wma9 160 sounds excellent to your ears, you can't really "expected to hear a fairly discernable difference" with mpc --extreme.

Use whatever sounds good for you, and enjoy the music

There are, what, 15,000 people here, I expect 1% of them could tell a difference between one well implemented codec and another at >=160kbps . And 15,000 people is only ~0.00025% of the population of the planet, so nope, you're not feeble.

Spaceman, unless you really need wma I'd stick with mpc --standard (--quality 5). Its transparent in most cases and will save you 17% diskspace compared to --extreme. Bitrates will average around 170k .

i would definetivelly go with MPC instead of WMA9, but ... i would surelly choose Vorbis between the 3 ... at least Vorbis has (limied) hardware support, and is as good as MPC ... my 2cents

Thanks to Roberto´s latest listening test we know that Vorbis AoTuV b2 is tied to MPC at 128 kbps.

I assume there are be plenty of people who think that for medium and high bitrates MPC still is a bit "safer".

...these are the sane people


later

Hey all,

Thanks a bunch for your responses. Thanks, especially, to Lev for the golden eared percentage. At least now I know I'm not a fruit because I can't really hear a difference.

I know you all prefer MPC or Vorbis to WMA9 - but how does WMA9 suffer compared to these two exactly? Is it in the high frequencies like with high-hats and snares, or does it tend to just sound duller and not have the range that codecs like MPC or Vorbis can achieve. I know it differs with the sample listened to, but it seems that certain codecs have weaknesses such as "warbling" in the high freq's, etc.

I think I'll stick to WMA9 just because I've already encoded over 350 hours of music with it. Re-encoding all that would send me loopy. I'll keep MPC and Vorbis in mind for any further encoding I do.

Thanks again,
spaceman

And by the way,

I just read the all important rule about comparing audio quality in Hydrogenaudio forums (ABX tests requisite, etc), so forgive me if my subjective questioning has (unintentionally) flouted the rules. I guess I should conduct an ABX test of my own to satisfy my curiosity.

Thanks again,
spaceman

Exactly

There's nothing in the rules against discussing the nature of a particular artifact. Rule #8 is to stop people from making a claim inconsistent with past tests with no evidence to back it up.

Anyhow, the particular artifact usually complained about with WMA is high-frequency "ringing" or a metallic sound. Some people are very annoyed by it, others can't detect it. I can only detect it with concentration. I thought I might have poor high-frequency hearing or something, but I tested that and it was pretty good, and I can detect lowpass fairly well, so there must be some other factor. It is also important to distinguish between WMA standard and WMA pro (two incompatable and completely different codecs). WMA pro doesn't appear to have this defect.

At 160kbps it's quite likely you won't notice artifacts of any kind unless they're extreme or you've trained yourself to listen for them. I remember back with the 128k extension test - I was showing it to a friend and he didn't even hear the artifacts produced by the blade anchor until I explained what they sounded like. Now his skin crawls whenever he hears an old, badly-encoded 128k or 112k mp3.