Skip to main content

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: [READ A WIKI ARTICLE INSTEAD OF MAKING UP SHIT] From: WAV vs FLAC and  (Read 3144 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

[READ A WIKI ARTICLE INSTEAD OF MAKING UP SHIT] From: WAV vs FLAC and

An audio cd is 1411kbps because every bit whether silent or not is represented. 44100*2*16=1411200 bits per second or 1411kbps. FLAC compression eliminates 0s and other reduncies to achieve a smaller file size (hence the lower bitrate), but the result is lossless. In theory, full white noise played at maximum volume is the only thing that CANT be compressed with FLAC because every single bit is a 1. Foobar (and any other software player) indeed decodes flac on the fly back to its uncompressed wave form. If you heard a difference, welcome to the wonderful world of expectation bias! Now run those two files through a double blind test using foobars abx comparator, and you will find there indeed is no audible difference between your wave and flac files. Also please keep in mind the Terms of Service of these forums, specifically number 8, before making subjective claims about audio quality between codecs.

As far as MP3, and other lossy encoders go, the simple explanation is they remove inaudible parts of the signal being fed to them, and replace it with low level noise. The more complex explanation will require some reading on your part. The level of compression used determines how much is removed, and how much noise is injected back in.

[READ A WIKI ARTICLE INSTEAD OF MAKING UP SHIT] From: WAV vs FLAC and

Reply #1
An audio cd is 1411kbps because every bit whether silent or not is represented. 44100*2*16=1411200 bits per second or 1411kbps.
It should be added that although the resulting FLAC has a lower (and newly variable) bit rate, it is, by virtue of its being a lossless codec, decompressed to exactly the same audio at exactly the same bitrate as the original before being output to speakers or whatever.

Quote
FLAC compression eliminates 0s and other reduncies to achieve a smaller file size (hence the lower bitrate), but the result is lossless.
The second sentence of this explanation might be useful conceptually but is very oversimplified. What 0s? Are you considering a multi-bit waveform as being compressed in terms of its individual bits in isolation, rather than more relevant methods of compressing actual audio? I think descriptions like this arise because it is initially easy to compare lossless audio compression to zipping a text file full of spaces and repeated words; however, even that comparison is far from accurate, aside from its upshot of the decompressed results being identical in both cases.

Quote
In theory, full white noise played at maximum volume is the only thing that CANT be compressed with FLAC because every single bit is a 1.
Never mind oversimplified: this is totally false. White noise is, by definition, a completely random waveform. Thus, although it really makes no sense to talk in terms of single bits in a multi-bit audio signal, we can do so for a random signal because there should, over a theoretical infinite period, be completely equal volumes of 0s and 1s. That would be impossible to compress with anything other than chance reduction in file size. By no means will white noise be all 1s. If that were the case, your audio would be a positively biased DC for its entire duration, not white noise. And, more to the point, a constant stream of 1s would be a free lunch for a lossless compressor, completing the failure of your description here.

[READ A WIKI ARTICLE INSTEAD OF MAKING UP SHIT] From: WAV vs FLAC and

Reply #2
Quote
White noise is, by definition, a completely random waveform. Thus, although it really makes no sense to talk in terms of single bits in a multi-bit audio signal, we can do so for a random signal because there should, over a theoretical infinite period, be completely equal volumes of 0s and 1s. That would be impossible to compress with anything other than chance reduction in file size. By no means will white noise be all 1s. If that were the case, your audio would be a positively biased DC for its entire duration, not white noise. And, more to the point, a constant stream of 1s would be a free lunch for a lossless compressor, completing the failure of your description here.


By "white noise" I did mean a full spectrum of 1's, without thinking that wouldn't produce any sound at all, and yes would be 100% redundant, and easy to compress. What I meant was a 100% full spectrum, full volume blast of noise would be impossible to compress, since there's no silence. Oversimplified? Yes. But making the zip file comparison, while not 100% scientific, is often times the answer that helps somebody new to encoding understand the difference between lossless and lossy compression. If they choose to get more in depth, and learn the scientific intricacies of it there is plenty of available reading on the web.

[READ A WIKI ARTICLE INSTEAD OF MAKING UP SHIT] From: WAV vs FLAC and

Reply #3
What I meant was a 100% full spectrum, full volume blast of noise would be impossible to compress, since there's no silence.


That's also wrong. That would imply nothing could be compressed because there's no silence in most music I've heard.

[READ A WIKI ARTICLE INSTEAD OF MAKING UP SHIT] From: WAV vs FLAC and

Reply #4
What I meant was a 100% full spectrum, full volume blast of noise would be impossible to compress, since there's no silence.


That's also wrong. That would imply nothing could be compressed because there's no silence in most music I've heard.


Agreed.  Lossless compression works by capitalizing on redundancy and unused bits. There is very little redundancy in music, so mostly the unused bit pool is the primary area it works from.

Lossy compression handles random noise in an interesting way. The compression algorithm notices that it has something that cannot be compressed by exploiting masking and that there is program material across all the spectrum. It then goes into overload posture and reduces the SNR across the band, and in the process replaces source noise with the encoder's noise.  So, what comes out of the decoder sounds like the right noise, but its numerically something else again.

[READ A WIKI ARTICLE INSTEAD OF MAKING UP SHIT] From: WAV vs FLAC and

Reply #5
What I meant was a 100% full spectrum, full volume blast of noise would be impossible to compress, since there's no silence.


That's also wrong. That would imply nothing could be compressed because there's no silence in most music I've heard.


Does your music run at full 16 bit depth at all times? You must listen to some severely brickwalled stuff! 

[READ A WIKI ARTICLE INSTEAD OF MAKING UP SHIT] From: WAV vs FLAC and

Reply #6
I guess you probably mean "frequency bands with no energy" when you say "100% full spectrum ... no silence".  Thats correct in that true white noise can't be compressed, but the reverse isn't true, just because something has a lot of energy doesn't mean it can't be compressed.  You don't need silence, just predictability in time or stereo.

 

[READ A WIKI ARTICLE INSTEAD OF MAKING UP SHIT] From: WAV vs FLAC and

Reply #7
What I meant was a 100% full spectrum, full volume blast of noise would be impossible to compress, since there's no silence.


That would imply nothing could be compressed because there's no silence in most music I've heard.


Does your music run at full 16 bit depth at all times? You must listen to some severely brickwalled stuff! 


Read the highlighted comment you made, then read mine. You are saying that it's impossible to compress anything except silence.

 
SimplePortal 1.0.0 RC1 © 2008-2021