I record all CD originated tracks to Apple Lossless, just to be ready when I upgrade my Mac to one with a better soundcard.
I keep most of 200 GB of music in Apples Lossless with doubles in mp3 for ipod use. Having a pristine lossless version is for later making any other types of AAC or MP3 files for mobile use.
Allthough it might be unwise to pick Apple's proprietary codec for archival storage, it still seem like the best choice for the ease of a large music library organized with the ease of iTunes, saving some space over .aiff files and still having bit perfect sound files sonic reproduction.
In a noisy car and to maximize storage space, I think compressing music files is acceptable and have chosen to use 192 VBR MP3 for iPod use, which may be overkill.
A Seagate 400 GB ATA (or SATA) HD kit for internal installation was just on sale for $129, so the only reason to record at lower bit rates these days is to enable more iPod capacity or to reduce clutter from keeping multiple copies of a song. I make mp3 as added to the ipod and intend to look further into keeping a seperate library of mp3 only songs, for ease in creating song lists for burning or downloading. Apparently own way to create a seperte library for MP3s is to make another user name and stock his library with only MP3 songs, but I'm still working on keeping a backup of all my itunes without having problems when backing up to a clone drive.
It's unlikely that nearly all the old analog music from the 60's- or 70's which was a either recorded on low resolution analog equipment or the masters were stored improperly and have started to breakdown (or both) gets any benefits at all, from Apple Lossless encoding. At least it won't be losing more quality, but it' still a shame to realize, even the best stored copies are more like a shell left from original live sessions. What a shame
What would be really useful... is a disc of selected timed tracks of obvious examples of flaws in Mp3 encoding at several bit rates and then a lossless version.
Sometimes I her (or think I hear) an Mp3 song as a whole just sounds closed in and dull with lossy coding, without one specific example to point out. That would be easy to miss on something like lower resolution speakers that don't sound "open" and natural to begin with.
I'd imagine using as high quality as is affordable, closed back (high sound isolation) headphones, would be the easiest and cheapest transducers to be trained to listen critically for to pick up for the average listener.
Identifying specific instruments and the exact locations containing coding errors, would be a first great step in training others to hear the differences between audio coders.
I'm convinced that most casual listeners are unable to hear differences between lower-mid rate coders and full rate music because no-one ever showed them which songs and especially which instraments are most difficult for lossy coders to accurately reproduce, a description of what to listen for and then played A/B listening tests until they could identify relltale sings over agin it's a combination of where to look for possible coding failures and what to listen for. Sometimes each much easier to hear a pattern (or find Waldo) once another person points it out.
Who's wondering if it might be better not to know how to bring the flaws in our coders to attention and then .
I imagine a simple tutorial/training tool for the budding audiophile, or for an itunes library owner who is willing test the belief that mp3's and lossless files are indistinguishable, and recording at higher bit rates over, (128kbs or so) makes little or no difference or cannot be heard in blind testing.
I'd just be happy to find a list of tracks that hopefully would not require too much time to amass from CD purchases, along with a time code with a short description on what to listen for would work. Does anyone else have a list of test tracks? Time coordinates would be helpful and exactly what to listen for.
A fairly short CD containing a tutorial about how and where to identify coding errors, with examples bit perfect encoding and a few encoding errors using A/B music clips, containing a number of instruments that are more difficult to compress accuratly, at lower bit rate settings.
Any links to the method of subtracking one track from another?
Doubt if I can hear a lot of difference in mp3's over a certain minimum around 160 or so. I do think there are some instraments that are difficult to encode accurately at even higher bit rates. Cymbals seem to suffer from encoding errors when given too little space problems with too little encoding space,. and will have an unnatural "warble".
-
Dave
Miles Davis is not the best example. Those old analog tape decks were not nearly as good as today's studio gear and were not "CD quality". Iv'e got as much of Diavis' recording as I can find and they are great to listen to but don't make good examples technical audio quallty. On the other hand almost anything made in the last decade by Deutsche Grammophon or Telarc are technically very good.
How many bits per second you need to match CD quality depends on what is being encoded. and it's surprising Things like trumpets encode easy The really hard to encode stuff is some "eletronicia" I've had a heard time with a few Bjork tracks and even the intentionally "lo-hi" Portishead. The electronic sounds are un-natural. Thinks like triangle waves can't be encoded using the methods used in MP3. Some drum sounds on an Eric Clapton CD just would not work until I went to 320Kbps too.
But for the most part if you use itunes with setting at 196kbps and variable bit rate with "highest quality" it works fine except for the few odd cases above.
One way to prove if you can hear the difference is to encode a CD to MP3 and then re-encode the MP3 file back to .wav and then subtract to that wav file from the CD track. Idealy you'd get a long stream of zeroes that not even your dog could hear but if the MP3 was not perfect you get some rubish that may or may not be audible
