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: lossy to flac - why big size? (Read 12260 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

lossy to flac - why big size?

Reply #25 –
MetalheadGautham, If you follow my previous link, you will in turn find this link...

http://groups.google.com/group/alt.radio.d...TF-8&rnum=7

...if you want to understand things, you'll have read explanations when offered!


Mike Giacomelli, the mp3 decoder does not add noise, apart from rounding noise or dither (rarely used in this context, apart from MAD and a few others). It does unpack the data in the mp3 file, and overlap it in time and frequency domains as required, which is an irreversible process. However, the real loss is in the mp3 encoder which represents the audio in this domain with lower resolution in the first place. If it wasn't for the quantisation in the encoder, the whole lot could be perfectly reversible (depending on encoder design - many have hard coded features that make perfect reconstruction impossible).

Cheers,
David.
 

lossy to flac - why big size?

Reply #26 –
Quote from: niktheblak on
Quote from: MetalheadGautham on

well, howcome that noise does not occupy a huge space in mp3?

It's because how the concept of 'noise' is defined in digital signal processing.

When you strip away information from a signal, you are adding noise to it. Reduction of information equals to addition of noise in signal processing. For example when you truncate a 24 bit number to a 16 bit one, you are adding 8 bits worth of noise to the 24 bit number. The noise it not stored in the compressed MP3 bitstream. It only appears while the MP3 stream is decoded and expanded into 16/24-bit PCM data for playback.



  and in simple words you are just pointing out what I said about the importance of reading of the original data stream itself insted of using the decoded stream... I have drawn the following conclusions in the process:

1. It must be possible to recoganise what portion of the audio is the artificial "noise" by recoganising the encoder used in the process.

2. it must be possible to use the MP3 bitstream to create the flac file, provided flac supports 16 bit (and thats what I guess, they call lossy flac)

3. this process is not nessarily decoding and recoding, but more like unwrapping and rewrapping

and a question: can psycoaccoustics be used successfully to detect and remove the noice? I mean, there are several frequencies which appear deaf to the human ear, so this must mean that it can be exploited, mustn't it?


Quote from: 2Bdecided on
MetalheadGautham, If you follow my previous link, you will in turn find this link...

http://groups.google.com/group/alt.radio.d...TF-8&rnum=7

...if you want to understand things, you'll have read explanations when offered!


Mike Giacomelli, the mp3 decoder does not add noise, apart from rounding noise or dither (rarely used in this context, apart from MAD and a few others). It does unpack the data in the mp3 file, and overlap it in time and frequency domains as required, which is an irreversible process. However, the real loss is in the mp3 encoder which represents the audio in this domain with lower resolution in the first place. If it wasn't for the quantisation in the encoder, the whole lot could be perfectly reversible (depending on encoder design - many have hard coded features that make perfect reconstruction impossible).

Cheers,
David.



lossy woes all over again.