CBR 320, MP3 or AAC
Reply #9 – 2007-01-31 12:04:01
... From your experience, is aac better than mp3 at 192? ... ... considerations about mp3- and aac-rippers .... At ~192 kbps you usually won't hear a diference, but as you can hear differences with mp3 on regular music when going from 128 kbps to 160 and 192 kbps your hearing abilities are pretty good, and as such you may worry about the rare but existing cases where mp3 has weaknesses. You can try for yourself with for instance these samples which you will find here on HA when searching for them:eig harp40_1 herding_calls trumpet The first two samples provide a problem for the encoder regarding temporal resolution. Helix is the only mp3 encoder to give me a satisfying result with eig at very high bitrate (and I'm pretty deaf towards temporal resolution problems). With harp40_1 it's not that bad, but it requires a bitrate above 200 kbps to get satisfying results. herding_calls and trumpet sound like simple samples but are difficult samples for various encoders. These samples show that there are tracks where at least a higher bitrate is neccessary than the one that is usually required. Using VBR in theory can solve these problems but in practice it depends heavenly upon encoder implementation whether or not that's true. As to my current best kowledge Helix VBR is working very well, and also current Lame 3.98 is going to get a good VBR mode. Fraunhofer VBR and VBR of Lame versions prior to 3.98 however have a questionable VBR mode when it is about guaranteing quality. Taking it altogether I'd say for people with good hearing abilities using mp3 at 192 kbps may be a bit too low with respect to critical parts of music. AAC in principle has the same problem in that a higher bitrate or quality setting is needed for these samples than is ususally needed. However a good AAC encoder provides excellent quality at a quality setting of ~ 128 kbps, and going ~180 kbps provides for an excellent quality even for these difficult samples. You can find out for yourself, and as your hearing probably is better than mine you may wish to vary parameters a bit. You should think of the ripping process and the encoding process as two different processes (that can be combined). You needn't necessarily rip and encode using Nero. Most people here use EAC (Exact Audio Copy) for ripping. With EAC you can combine the ripping process with the encoding process by integrating any commandline encoder like Nero's free AAC CLI encoder. I personally keep it as independent processes, rip to wavs using EAC, and encode the wavs using foobar.