Best format for music

Organichu

Hey guys. I had to reformat and so now I'm looking at downloading a lot of my iTunes songs again (and more importantly) ripping my many CDs to my HDD. I never really thought it out before. Now that I'm starting fresh I want to do it right.

1.) I do not have a premium sound system, and so I imagine the most perfect quality (would that be pure .wav files, directly from the original CD?) would be wasted. I'm looking to listen to these tracks through a laptop's speakers, mid level ($50-100) headphones, and the same headphones through an iPhone.

2.) Space is an issue, but not a huge one. I expect I have maybe 3,000 tracks or so? I don't expect to fit them all on a 16gb iPhone, as it'll have video on it too... but I don't want 20 MB tracks.

3.) What is 'lossless' audio? It seems impossible to downsize on the amount of data in the container without degrading quality, so I doubt it means what it sounds like.

4.) I'm looking for something free.

Thanks a lot in advance, guys.

Typhoid Manny

Free Lossless Audio Codec (.flac) is probably the best audio format around right now. Honestly I don't know what the "lossless" means in techincal terms, but my semi-trained ear can definitely pick out the difference between flac files and 192kbps mp3s. The downside is that they're about double the size of those mp3s, and I don't think iTunes supports the format.

Willeth

Organichu wrote: »

3.) What is 'lossless' audio? It seems impossible to downsize on the amount of data in the container without degrading quality, so I doubt it means what it sounds like.

That's exactly what it means. I only have a basic understanding of it, but think of it as a vector image compared to a bitmap image. One contains instructions to build the image, the other contains details of each individual pixel.

For your needs I think you're best ripping to 192Kbps MP3. Pretty standard and won't sound terrible if you upgrade at some point. Other people will be able to provide software solutions - I don't think the software I use is even available any more.

Zoolander

Just stick with MP3. If you don't, chances are you'll regret it at some point. Besides the lossless codecs (which really are lossless, like zip files for your music), all the other codecs are in the same ballpark. Fancy codecs like AAC or OGG are surely better than MP3, but not so much that it's worth the hassle of potentially not being able to play it on some mp3 player or other.

256kbps MP3 is more than good enough for what you're describing.

TechBoy

I also agree with using MP3 simply for compatibility, but use a VBR (variable bit rate) MP3 encoding. It'll be of higher quality and smaller size than simply using a constant bit rate (CBR). Really old stuff might not support it, but everything you're running will.

DrFrylock

Audio can be thought of as a waveform. When you record or manipulate audio digitally, you're taking a fixed number of samples of that analog waveform per second (for a CD, 44,100 samples per second) at a particular resolution (for a CD, 16 bits per sample). These are used by an audio player to reconstruct the analog waveform. The more samples and the higher resolution, the more accurately the analog waveform can be reconstructed.

Lossless storage or compression preserves every single one of these samples at its highest resolution. WAV is more-or-less uncompressed, which means that any redundancy or extraneous information in the waveform is preserved in its native form. FLAC uses some fairly sophisticated algorithms to reduce file size, but in such a way that the original digital waveform can be reconstructed exactly. It's like ZIP, but made specially to take advantage of the characteristics of audio. So, a FLAC played back will sound exactly the same as the original WAV, but it will take slightly more processing power for the decompression.

Formats like MP3, non-lossless AAC, and WMA are all "lossy." This means that they throw out information to reduce file size even further. As information is discarded, the file size gets smaller, but the original waveform can't be reconstructed quite as exactly. The more you throw out, the smaller it gets, but the more the resulting reconstructed waveform differs from the original. Eventually, the difference becomes large enough that it gets noticeable.

The MP3 encoding scheme took advantage of two particular developments: an increase in available computing power for decompression and new algorithms for "psychoacoustic" compression. Uncompressed audio requires basically no computation to decode: you just feed the data to a fast enough digital-to-analog converter (DAC) and send the output to your speakers. Computationally, MP3s are fairly expensive to decode. I don't think most 486 computers could decode them; early Pentiums could. Psychoacoustic compression means that the MP3 encoder tries to throw out information selectively, based on the known characteristics of the human ear. "OK," it says, "most humans can't hear this particular thing anyway, so let's throw that out, and keep this other thing that is very noticeable to humans." This allowed a 128kbps MP3 to be, to the ear, very similar to the audio quality found on a CD, which requires just about 10X as much data. So, a 680MB CD could be effectively encoded in about 68MB.

Constant bit-rate (CBR) encoding sets a target data rate (say 128kbps) and then throws away enough information to meet that constraint for the duration of the audio. Variable bit-rate (VBR) is even more selective. It will look at various segments of audio (perhaps 1 second or 0.5 second intervals) and compressing some more and some less. So, it might decide to encode the first second of audio at 64kbps and the second second at 256kbps, based on content and compressibility. On average, VBR files should be about the same size as CBR files, but slightly higher quality since the more dynamic parts of the audio have less information thrown away (i.e., are encoded at a higher bitrate).

I personally have resisted VBR because I've had problems with various players and the ability to seek into a track accurately; others will tell you this is a solved problem but it really depends on the player, whether your encoder puts seek information into the file, whether your player uses it, the size of the file, whether you're playing from the filesystem or seeking over an HTTP stream, and so on. Your mileage may vary.

Organichu

Thanks a lot guys. I'll give this all a look and sample some FLAC stuff and some lower bitrate mp3s to determine how much of a difference I personally note.

Organichu

I'm finished ripping my Lil' Wayne and Ben Folds/Five discographies... many gigabytes!

I think I may be forced to get suuuuuper selective with my song choices or definitely go with the mp3 option.

Unknown

This content has been removed.

Organichu

Ok thanks a lot Mcdermott.

That sounds fair. So at 192 I'm probably looking at about 1.5 MB/minute?

At 3,000 tracks, an average of 4:30 (I'm guessing), that's about 20-21 GB? That is incredibly doable, as I'd be able to keep well over half my tracks on my iPhone at all times.

I'm going to rip a few tracks into 192 mp3 right now, see what they sound like.

Will I notice any major difference ripping down (from say 320kbps) to 192kbps versus ripping to 192kbps straight from the CD?

Thanks again everyone. I'm actually going to be diligent naming stuff, keeping all caps consistent, etc... I'm really excited to not have loose ends that mess up my coverflow view. I really appreciate this opportunity to get everything organized evenly across the board.

Feral

Organichu wrote: »

Will I notice any major difference ripping down (from say 320kbps) to 192kbps versus ripping to 192kbps straight from the CD?

Compressing a file twice (once to 320 kbps and again to 192 kbps) can introduce encoding artifacts that will diminish sound quality compared to compressing straight from CD to 192 kbps.

Personally, I've seen a lot of blind listening tests that suggest that only a tiny minority of people can tell the difference between 192 kbps MP3 and anything better on consumer-level hardware. If you have really nice hardware and really well-trained ears, you might be able to tell the difference between 192 and 320, but most people can't.

If you're going to be doing listening tests, do them blind. Close your eyes and get a friend to play you each track without telling you what it is. I'd be very surprised if, in a blind test, you can consistently tell the difference between 192 and 320 or 192 and FLAC.

Organichu

Yeah I mean, I'm a classically trained pianist but I'm pretty sure I don't have 'the ear' that is so sought after. It took me four or five years of guitar to even tune by ear, and I sure can't transpose busy songs... so I doubt I have this perfect aural fidelity or anything. I'm sure I'll be fine with a modest bitrate, but I would like to test it out on a few different genres first.

Unknown

This content has been removed.