.spec wrote:Surface_Tension wrote:
How it works in my understanding, is that as the quality of the mp3 decreases, the cost is to the high end, which generally is in the inaudible spectrum for human hearing. However, you don't need to see the wind to know it's blowing. It's invisible, but it can make the leaf on a tree blow and you can see that. So you can't hear this missing high end, but it does skew the way the low end travels, thus making an audible difference, if only slightly. So at the end of the day, the mix will be muddy and unclear.
This isn't really how VBR encoding works, but whatever. I would love to do a blind listen to a song encoded with FLAC -V0 and 320 with you. I feel pretty fucking confident that you nor most the people on this forum (myself included) could tell a difference.
This goes double for a tune played in a club/at a party. By and large most huge sound systems are only good for playing shit loud, not pinpoint sound representation. I'm sure we could all hear a 192 vs 320 on a big system but the difference between a -V0 and 320 is about 64k/sec i.e. negligible.
it's nothing to do with vbr or cbr. 320 is cbr (constant bit rate), which only means that you have a specific amount of data -- 320 kiloBITs per second of audio -- in which to represent the audio. conversely, vbr means that the result has a dynamic bitrate: parts of the tune with huge frequency range etc can be represented with a lot of data (usually up to 256kb/sec), and near-silent or more simple parts can be represented with anything down to 64kbit/sec. how you manage to stuff it in is up to the encoder: it invariably means throwing shit away (hence lossy). what it throws away is up to the encoder.
going from anything to flac or wav is harmless because flac throws nothing away: it's lossless compression, like a zip file. going from anything lossy (e.g. any mp3) to anything else lossy -- transcoding -- will always result in degraded quality, because your encoders are going to have different behaviours, and different results on different inputs. so encoder #1 will throw away one set of information to cram it in, and usually compress your range in the process (frequency, not amplification), and then encoder #2 will throw away _another_ set of information when it encodes. it's kind of like playing a crap tape down a lousy phoneline. (yes, this includes taking a 320, converting it to wav, and then re-encoding it to mp3. don't.)
320s are convenient for aim and stuff (australia has pretty ghetto third-world internet), but i have to say i really love producers who keep everything in wav and always have the wavs ready to send if you like the 320 -- you know who you are, big up!
anyway, enough with the facts in a discussion like this.
