[streetdragon]

Quote:

 
http://www.jiscdigitalmedia.ac.uk/audio/advice/aac-audio-and-the-mp4-media-format/#aac-basics
 
Lossless formats use non-destructive compression, lossy formats use destructive 'compression' ..
Except it isn't compression, it's data-reduction ..
 
Lets say you want to compress a can of air .. Do you remove some of the air and call it compressed ?

so do lossy files only lose data? or compress and lose data as well?

 

[Terminator02]

Quote:

Originally Posted by DougofTheAbaci /img/forum/go_quote.gif
 
It's not worth mentioning Ogg because it's got the most limited support of any of the major audio formats, doesn't have the best quality/size ratio and wherever it is supported better formats are also found. In other words: Don't bother with it.
 
As for how, it's the same way you can compress a folder into a ZIP and RAR archive and get two different file sizes that contain the same data: Compression algorithms. MP3 is an older file than M4A (AAC).

While I'll agree with the limited support, I don't know what justifies your quality/size ratio argument. If there is some general consensus that AAC is better than Ogg Vorbis, or if you personally tested it yourself, I don't know. I'm not trying to belittle your argument, I'm legitimately curious as to how the two compare (I haven't done any serious testing between the two myself).
 
As for 320 kbps CBR, you're not getting the point. Both MP3 and AAC files that are at 320 CBR kbps will have the exact same size on the same track, only the sound quality will differ. I realize that the compression algorithms are different, that is why AAC should do a better job at preserving the audio quality in the allotted amount of bytes per second of audio (320).

 

[mikeaj]

To me, these would be examples of losing data (let's suppose we start with 16-bit, 44.1 kHz sampling rate stereo data, so 2 channels):

• Instead of 16-bit samples, chop off the least-significant 4 bits and have 12-bit samples.  You've lost one-fourth of the data.  This degrades the quality somewhat, yet the file size is still 75%.
• Instead of 44100 samples per second, throw away every other sample, so you have 22050 samples per second.  Now you've lost half the data.  This degrades the quality significantly.  In fact, you've lost audio data above 11025 Hz.

 
Lossy audio compression essentially does some processing and filtering of the data, switches from representing the data in one alphabet (so to say) to another, more efficient alphabet, and figures out which letters can be used to best represent the original data, based on the psychoacoustic models for what is important and what is not.  The new alphabet exploits the features of most audio files, the way people hear, and the energy compaction properties of the (modified) discrete cosine transform to reduce the amount of data in an efficient manner.
 
So I would call that compression, sure.

 

[bigshot]

In practice, the loss in sound quality using AAC is nil. I took an AAC 256 file and reripped it to AAC over and over again ten times- ten generations of lossy encoding. It sounded fine. Try it yourself. You'll be amazed.

 

[AKG240mkII]

Quote:

so do lossy files only lose data? or compress and lose data as well ?

That's a very good question !
I have never seen the entire source-code for AAC-encoding so I can't say for sure .
 
What I DO know is : There is NOTHING left to compress in lossy-encoded video-files .
It doesn't matter how you try to compress a mpeg-video, the only way to get it smaller is to transcode it to a
even worse format .
Likewise, it doesn't matter what compression-level you choose in 7zip, win-zip or win-rar when you try to compress a AAC,mp3 or any other
lossy media-file . In fact, you will sometimes end up with LARGER files !
 
In matters like these, WIKI isn't all that bad :
 
Quote:

In computer science and information theory, data compression, source coding,^[1] or bit-rate reduction involves encoding information using fewer bits than the original representation.
Compression can be either lossy or lossless. Lossless compression reduces bits by identifying and eliminating statistical redundancy. No information is lost in lossless compression.
Lossy compression reduces bits by identifying marginally important information and removing it.
The process of reducing the size of a data file is popularly referred to as data compression, although its formal name is source coding (coding done at the source of the data, before it is stored or transmitted).
https://en.wikipedia.org/wiki/Data_compression

Without having seen the source-code, my guess is that lossy formats do both,or attempt to, since the main purpose is to reduce file-size .
 
PS :
It's important to note that the word 'compression' doesn't (always?) mean the same in computer-science as it does in audio !

 

[DougofTheAbaci]

Quote:

so do lossy files only lose data? or compress and lose data as well?

 
I was under the impression that they did data compression as well. It would seem a very logical way for reducing file size.
 
Quote:

While I'll agree with the limited support, I don't know what justifies your quality/size ratio argument. If there is some general consensus that AAC is better than Ogg Vorbis, or if you personally tested it yourself, I don't know. I'm not trying to belittle your argument, I'm legitimately curious as to how the two compare (I haven't done any serious testing between the two myself).

 
I did a bit of testing early on when it looked like Ogg was going to be the de facto HTML5 audio codec. Once it became clear that wasn't going to happen I quickly stopped caring given it's lack of support. Beyond that it was more an impression I'd gotten while researching the format.

 

[grapefruit]

I've noticed with my AAC rips that the music sounds harsher and less nuanced, whereas MP3 is less so. Just my 2 cents.  

 

[bigshot]

Harsher sounds like clipping. Your music is probably hot mastered.