I listen to quite a wide range of music, but hardly ever on a setup or in a context where I can fully appreciate or even perceive the fullness of the fidelity of reproduction. Not that I don’t want good fidelity, but I hardly ever sit still and just listen to music without some other ambient noise or activity going on.
Probably for this reason, my senses are not sufficiently attuned that I can discern any differences between a CD and a decently-encoded MP3; I’m aware of the whole ‘perceptual’ aspect of MP3, but I just can’t personally tell the difference.
In exactly what kinds of ways does an MP3 (and for the purposes of this thread, we’ll assume that a decent job was made or encoding it) sound different from the original?
Well mp3 compression works by attempting to eliminate the frequencies the human ear wouldn’t hear anyway. IIRC 20hz to 20khz. Minidiscs work in the same way but to a better job, probably because they’re a high bitrate. As for the sound, it’s slighly “muddy” sounding. The treble is reduced a bit, the bass is a tad more muffled. In general, the clarity has been reduced ever so slightly. Most of the time the reduced quality only becomes apparent when comparing to the original CD.
Download the programs to make MP3’s and you can hear it for yourself. Search for LAME MP3 encoder, install that first, and then find the RazorLAME GUI front-end that makes LAME easier to use (LAME itself is a command-line program, RazorLAME is a “window” that controls LAME). Both of these are GPL/free programs, and aren’t big files. You’ll need a CD ripping program to get the CD songs onto your computer, but you might have one already. Anyway, you can’t just copy&paste the files as you see them on the CD.
LAME allows encoding from the maximum of 320 Kb/sec to all the way down to 8Kb/sec, but it may have a hard time with really low compression rates, depending on the sound being converted. The “standard” MP3 rate is 128 Kb/sec; 320 Kb/sec is basically identical to the original CD but with little file compression, I prefer 192 Kb/sec. At 128 Kb/sec I find that cymbals tend to sound all warbly. Using 160 or 192 keeps them clear. It depends on teh sounds you are converting, but I found that LAME (and a couple other encoders I had) started to sound majorly-distorted down around 56Kb/sec.
I do not claim to have golden ears by any means–and especially do not claim to have a golden PC speaker setup (I got the $50 Logitech 3-way setup) but I can easily hear the poor cymbals of a 128 Kb/sec MP3, and I can hear that the 192 Kb/sec MP3’s sound way better.
~
You answered your own question when you said “decently-encoded”. Some people use a lower bitrate to save hard drive space, and the quality suffers. On the other hand, a higher bitrate preserves fidelity. I find that, for my purposes, a bitrate of 128 is usually sufficient, but for stuff I intend to broadcast, I use a higher bitrate.
Question’s already been answered… but I’ll just toss in that the question of MP3 encoding seems to be highly subjective. I can tell a difference between 128 bitrate and 160 bitrate fairly easily… (I think - never actually set up a blind test,) but as far as 160 bitrate and 256 or 320… much less certain if my ear can pick up any distinction.
Consequently, I have a huge collection of 160 bitrate MP3’s of pretty much every CD I own.
…On the other hand, I can tell a difference between a 320 bitrate MP3 and ‘full CD’ audio quality – which is why I insist on ripping to WAV sound (which can preserve full CD audio) when I’m burning a mix CD… as long as it’s my own stuff of course. If I throw in anything that was shared off the net (not that I file-swap any more, but I do have a small treasure trove of stuff from napster through kazaa,) then I have to take whatever quality I got.
I have done a lot of ripping/encoding/etc myself (legally, for stuff we used in the playout system for a community radio station), I just keep coming across people who wrinkle their noses when MP3s are mentioned, claiming that they are rather poor and that the difference is easily discerned, even with decent bit rates etc. These people are typically oddly short on the specifics of what kind of differences it is they can hear though, which made me ask the question.
Of course this was for FM radio broadcast, so the absolute quality of the sound source is not quite as crucial as it might be elsewhere, but as I said, I can’t discern the difference (listening off-air), but I didn’t really know what I’m looking for.
I think it’s funny that people get their noses all disjointed over it. If everything started sonically from the same place they might have a point, but it doesn’t, and they don’t. Take a clean copy of something recorded and produced badly versus an mp3 of something fantastically produced.
You know you could encode those .wav files into a losslessly compressed format like FLAC to save some hard drive space.
Also note that how good an .mp3 sounds also depends on the encoder as well. LAME is best encoder out there: to my ears a 128kbit LAME encoded .mp3 sounds as good as many 160kbit .mp3s encoded by other programs.
Saving hard drive space is kinduv a non-issue for this purpose, though, since the hard drive is only a ‘temporary stop’ as it were.
Come up with a playlist. Rip from however many store-bought CDs to WAV files on the hard drive. Burn WAV files to a blank CD. Listen to burned CD once through to make sure everything came out okay. Erase WAV files from hard drive.
As long as I have 700 megs free on a drive, space itself isn’t an issue. I know I could experiment with loss-less formats, but as far as I know there isn’t one that is standardized and supported by all audio CD burning programs like WAV is.
Pretty sure you’ll find that CDex actually uses the LAME encoder - it used to be the case that you had to install it separately, but many freeware apps now come with it bundled into the installer.
I worked for awhile at a company that built MPEG-2 encoders and decoders. IIRC, MP3 derives from the audio part of the MPEG-2 spec. While the video compression followed a set pattern of fixed frames (basically JPEG pictures) and transitional frames, the audio compression was part art and part science. It was based on what they called “psychoacoustics”. For example, they figured if there was a tone 20 cycles away and 2 deciBels softer than the primary tone, it wouldn’t be heard anyway, so it could be eliminated. They had a bunch of tricks like that. Finally, they would take the compressed music and present it to experts: music critics, opera singers, producers. The experts would then fill out surveys that basically asked, “How annoying is this? a) Not annoying, b) Somewhat annoying, etc.”
Interesting. A widely used piece of gear used in recording studios is an “Aural exciter” made by Aphex. It works with psychoacoustics as well, altering frequencies that in turn affect our perception of the sound. I could only hazard a guess behind the actual mechanics of the device, so instead I’ll just say it certainly makes something sound nice.
My minidisk deck (which uses ATRAC3, I think) very infrequently adds these weird artifacts, usually rapid, high-pitched, electronic-sounding squawks. There’s one song by Bjork in which the same annoying “bleep!” sound pops up at the exact same place every time I’ve tried to record it. Normally the ATRAC sound quality is excellent, but the compression process does seem to have some odd bugs that not only degrade, but essentially ruin the sound. The same tune ripped into an MP3 causes no such problem, but I’ve not used MP3s enough to say they’re generally superior at high bitrates.
Are there similar artifacts produced by MP3 compression?
I tend to notice a warble/shimmer type distortion with high frequency sounds - like the transients from cymbals. I also get vocal distortion occasionally. But, up the bit-rate and they tend to go away.
MP3-encoded music is a lossy compression and so, in the theoretical level, there is a loss somewhere. Whether or not that loss is heard depends on how compressed the audio is and at the “quality” of the ears of the listener.
Don’t mind audiophiles turning up their noses at you. How does the saying go - “audiophiles don’t even like other audiophiles” or some such.
One thing that is noticeable is lack of dynamic range.
If you encode a CD that has been digitally mastered from either the studio or from the master tapes, and compare this to MP3 files, you may well find that the MP3 file does not have the impact or punch the original master has.
This is noticeable with music that has sudden pulses from relatively quiet passages, so that cymbals crashing don’t hit as hard, or a massive bass drum beat does not have the physical effect on the listener.
I think the age of the intended audience needs to be taken into consideration. Myself, I’m very sure my ears ain’t what they used to be although I have taken care to protect them under adverse conditions. I simply can’t tell the difference, on my hifi system, between a 128K bitrate-encoded MP3 and the original CD, at least in an informal, non-double-blind home test environment.
It’s well known that hearing of higher frequencies decreases as people age.
That said, I’m enough of a purist to save everything I want to exist for posterity as WAV files (stuff that is rare or impossible to recreate). Right now I am transferring some 1950’s and 1960’s original, one-of-a-kind open-real audio tapes to digital storage. Even though the originals were sometimes recorded on substandard or amateur equipment, I try to preserve every possible bit of quality during the transfer. If someone wants to reduce that quality later, that’s their business.
Thank goodness the cost of WAV storage on a CD is so little as to not be a factor. I might feel differently if I had to store a 30MB file on 25 floppies or if CD blanks cost $10 each!
