Good day.
I tested the lossy audio formats using RightMark Audio Analyser 6.2.3.

By RightMark Audio Analyser 6.2.3 was generated a test WAV file with the following characteristics: 16 bit, 48 kHz. Then the source file has been packed in various audio formats with different bitrate, then these files are converted back into WAV. Since the format - lossy, ie, lossy compression, it will lose some audio information has not been restored in the reverse conversion to WAV. This gives us the opportunity to see what exactly happened with the sound by each of the codecs.

The article originally is in Russian, but you can use a Google Translator, for example.
Here is a full Google Translated version. I think it is rather understandable.
Here is a full original Russian version.

I would like to read your feedback on this method.
Thank you.

The raw numbers may be interesting from a technical POV, However: 'what sounds better' ? is a different matter. When taking into account perceptual quality - raw numbers mean little.

shadowking
I partly agree with you, but this test shows the ability of a specific format to reproduce the original sound as closely as possible with a sufficient great compression.
I think the pluses of this method are that the result does not depend on the subjective perception (different people hear differently), the quality of the speaker and the quality of analog audio output.

You should look into a simplified encoder such as lossywav, wavpack lossy, optimfrog dualstream . Those will attempt to reproduce the original signal with more accuracy - but achieve less compression.

aac/mp3/ogg have little regard for reproducing original waveforms.

http://www.hydrogenaudio.org/forums/index....mp;#entry149481
http://www.hydrogenaudio.org/forums/index....showtopic=16295
http://www.hydrogenaudio.org/forums/index....showtopic=11442

Discussion closed.