Audio formats

Just wondering what format would be best to use for my iPod. It’s a 80GB sixth-gen model. I use the volume normalizer, usually…but sometimes, I turn it off for more volume. I mainly listen to heavy metal and thrash. Y’know…Megadeth, Nevermore, Slayer, Children of Bodom, etc. . Here’s the available formats:

– MP3-320kbps (lossy)
– AAC (lossless)
– M4A (lossless)
– AIFF (lossless, I think)
– WAV (lossless, I think)
– Apple Lossless (currently used)

This probably isn’t enough info for you guys to help me, so if you need more info, feel free to ask. Thanks. :wink:

<<>>

Oh, I know what they are. You’re just begging for us to poke fun at you, so here goes: Anybody who listens to that probably has no hearing left anyway. Did you know certain CostCos will check your hearing for free? It’s the loss leader for their hearing aid department. You need to make an appointment, tho.

OK, that’s enough of that.


– MP3-320kbps (lossy)
Yes, but at the bitrate, it’s almost not lossy. Very few people can tell that MP3 is back there compressing stuff when the bitrate is set that high. Of course, at that bitrate, the songs are pretty large, too.

– AAC (lossless)
No. Advanced Audio Coding is lossy, but unless you take a lot of pains with special MP3 coders, AAC will do better at the same bitrate than MP3. I use the highest AAC on my iPod.

– M4A (lossless)
MPEG4 Audio I believe is still lossy. This is getting into the weird and wonderful world of audio standards that not everybody supports.

– AIFF (lossless, I think)
– WAV (lossless, I think)

These two are cousins of each other and they’re both cousins of Music CD format CD-A. Yes, these two (three) are lossless but you still need to be careful. You can still make WAV files sound terrible by picking wrong bitrates and sample frequencies. The difference with the non-lossy formats is that the music stays the same level of bad from that point on. The lossy formats get worse each time you do anything with them. The sticker shock with non-lossy formats is the filesize. There is no space saving at all. Converting an average MP3 into WAV will easily result in a five times larger file, and many times far worse. Also remember that the song has the damage of the worst format in its path. You can’t download a destroyed MP3 file and magically make it all better by converting to WAV. What you will get is a very, very pure and accurate destroyed song that takes a lot of disk space.

Once you do that, however, the song won’t get any worse through editing and production. It will start getting worse again when you convert it back into a lossy format.

These are the formats you want to do production in, then convert to one of the lossy formats for your iPod. For example, when I have to capture voice recordings, I do them in 16-bit, 48000 and save them as WAV files. That’s the format I use for cutting and production later, and then, when I’m happy, I export as WAV and use the Apple iTunes Audio Import AAC converter to produce my iPod cuts.

Never double convert. Feeding an MP3 file into iTunes and then telling iTunes to produce an AAC song is asking for trouble. Compression damage multiplies. If you produce the MP3, tell iTunes to leave it alone.

Every engineer I know did the quality tests. We all picked a song that we know and messed with all the formats at different quality levels to hear what they did, and then picked one for our own use. Our choices turned out to be remarkably similar in spite of not talking to each other ahead of time. However, your mileage may vary.


– Apple Lossless (currently used)
I don’t know enough about that one, but the Apple Lossless formats in Video Land are compressed and may have cleverly hidden damage. I don’t know.

Koz

I agree with Koz’ (somewhat hidden) recommendation - in that I doubt that you will hear any difference with AAC at high bitrate v. Apple lossless - and the AAC wil take up less space. Both, though, are of course Apple proprietary formats and only play on iPods/iTunes.

And remember if you do run out of space - there is always the option of the 160gb iPod Classic (will need to be my next major purchase as I have run out of space a while ago on my old iPod).

WC

Erm…okay. Koz, I hear fine. I actually have to hear for my dad, usually…anyway. The Apple lossless seems to have the ability to achieve a much higher volume without any clipping than the MP3 did. Also, it doesn’t seem to be as effected by the EQ as much as the MP3s did. Currently, I think I’ll stick with Apple lossless. Thanks, everybody! Seriously, thank you. The way I see it, the GBs are there for storage, and I still have 53.4GB left, so the large file size doesn’t hurt much.

<<>>

I wonder when people say this, because that’s not supposed to happen. I don’t doubt that it does, tho. Each audio format is something of a slave to the interpreter that goes between the audio amplifier and the sound file. 0dBFS is the digital audio overload point and it doesn’t change.

It’s scary that people are choosing file formats because of how loud they are.

I’ll have to try that.

Thanks for the info. We use the responses to each question to help others.

Koz

First off…what? Okay, you lost me. What is 0dBFS, and a “slave”? I know what a “slave” hard drive is, but nothing else. That went completely over my head. Second…why? You said it was scary that people are choosing file formats based on volume. How so? I’m just curious, that’s all. For me, it’s a mix. I like my music loud, but if the sound quality is crap, then I won’t use it. Generally though, quality over quanity (volume) for me. Thanks for feedback everyone.

This picture is what happens when you separate the meterbar from the work window, change the timeline to dB from Percent, and pull the window very tall.

http://kozco.com/tech/audacity/AudacityPanelFull.jpg

Note that the sharp points on the blue waveform approach zero (guage on the left) and the far right-hand side of the green bouncing light sound meters is also labeled zero. That’s zero dB Full Scale. 0dBFS. That’s the magic place where the digital audio system can no longer get louder. Because of that, zero is another magic place, it’s the place where the audio distortion goes up to the moon. You want crunchy bass notes during a live band performance? Just let the bass notes push either the blue waveform or the green meters into zero.

So zero is non-movable.

It should be possible to start with a performance that is undistorted (like the one in the picture)…

http://www.kozco.com/tech/soundtests.html

…which is the Piano Trill from that web site and export several different versions of the show (MP3, AIFF, WAV, OGG, etc. etc.) and the performance on each and every one of these sound files will be exactly the same volume.

In the case of the MP3 file, there are compression settings and you can tune file size and how much distortion you’re willing to accept, but the volume of the show shouldn’t change. Just the quality.

That’s how it was designed to work. If you exported a show multiple times and picked an audio file type based on how loud it got, something is broken.

Koz

So koz has already answered this, but this is soch a juicy question :slight_smile:

The important word in the above statement is “seems”. These effects are entirely down to the playback system.
In my house, audio CD’s are much louder than tapes !
Why is this?
It’s because my cassette player is attached to a domestic 30 watt per channel hi-fi system, whereas I have a CD player that is connected to a 2500 watt per channel PA system. The tapes I can play loud, and with CD’s I can make your ears bleed.

So this may not be exactly what was meant, but the fact is that recordings do not need to be “maxed out” in order to sound loud, most playback systems include a volume control that can take care of that. If your play back system will not go loud enough with a well recorder source, then you need a more powerful playback system.

So why is it that recording “A” sounds louder than recording “B”, even though both recordings go full scale in Audacity tracks? (Technospeak: they are normalized to 0dB. Non-techy- speech: the peaks reach all the way to the top and bottom of the track).
This is where we nip over to Wikipedia and look up "Digital Audio: Digital audio - Wikipedia
As we see from their illustration, our digital recording aims to accurately trace the recorded waveform, but is only able to approximate to it:

Furthermore, representing a waveform in number this way means that there is an absolute maximum value that can be used for a sample, and an absolute minimum value.

Since CD recording use “16 bit” binary digits, that effectively gives us a maximum range for any sample of between 0 and 1111111111111111, which in normal numbers (decimal) is a range of 0 to 65535, but since our waves go up and down (zoom in close on an Audacity track to see the wiggly blue line), this gives us a range from “minus 32,757” to “plus 32,767”.
A waveform that moves between these two extremes is at the maximum possible amplitude, and for that specific frequency, nothing can go louder. This is called th zero decibel level (0 dB). On a well designed system, this would equate to the “red line” on our level meters.

With our Digital recording, the quietest possible signal is a flat line (usually at “0”, or close to it) - this is silence.
The quietest sound we can record would be a signal that fluctuated between a high and low value that cover a range of “1” - for example, fluctuating between 000000000000000 to 000000000000001 (which would be very quiet indeed).

So on CD’s we have a maximum possible range that goes from, silence to 0dB in 65,535 steps.

We said earlier that digital audio tries to copy the shape of the sound wave as precisely as possible. Any variation from this is some kind of distortion. Is distortion a bad thing? Well Jimmy Hendrix would say no. Distortion can sometimes produce a more desirable sound than a completely clean and undistorted sound. On the other hand, we have all heard recordings that are badly distorted and sound terrible. It all depends on the kind of distortion, and whether it is appropriate (desirable) for our recording.

So how do we make a recording sound louder? We can use a “trick” called “dynamics compression”. We have a maximum range of loud to quiet, but we can apply a special kind of distortion (called “dynamics compression”) where we make more of the sound reach the highest levels. Effectively we “squash” (compress) all the sound into the louder part of our dynamic range. This kind of “compression” has nothing to do with “compressed formats” (such as mp3’s) - This is compressing the dynamics (loud / quiet ratio) whereas “file compression” (mp3’s) are about making the file size smaller.

The Good thing about dynamic compression - it can make music sound louder when played at a specific volume level. It can even out volume fluctuations that can sometimes be desirable. It can help make different tracks on an album sound at a similar volume level.

The Bad thing about dynamic compression - it can make music sound “flat” and monotonous. It reduces the dynamic contrasts in the music.

The Ugly thing about dynamic compression - it is a form of distortion and it is throwing away some of the sound quality that we can have. It assumes that people are too stupid or too lazy to adjust the volume control on their CD player. In extreme cases it will turn all of our lovely musical accents into flat farty sounds.


OK I’ve had my 10 cents. There’s a certain amount of personal opinion and personal taste involved when it comes to compression, but hopefully this provides a little insight into what it’s all about.

Back to the original question - you missed out FLAC (lossless) - approx 50% lossless compression, but probably not compatible with iPods. Best quality on iPods is probably using Apple Lossless - not because it’s any better than other lossless compresors, but being Apple, it will be well supported on iTunes and iPods.

There’s a load more information about audio file compression codecs here: http://en.wikipedia.org/wiki/Comparison_of_audio_codecs

But the other thing to consider when choosing formats is: are you going to stick with Apple/iPod/iTunes forever - if so then using Apple Lossless or high bitrate AAC will be the choice fotr you.

If you are concerned about future portability then you might choose high bitrate MP3 or WAV.

WC

P.S. and I hate dynamic compression - especially when radio staions like BBC R1 use it to sound LOUD - and when recording engineers “remaster” older recordings to make them sound LOUD in the current relentless loudness wars that record companies seem intent on enagaging in.