Sample rate conversion is done on the following cases:
By default, qaac works as if --rate=keep specified. --rate keep tries to keep the same sampling rate as source.
On the other hand, --rate auto lets encoder choose the optimal sampling rate. This changes the behavior when the target bitrate is low, and --rate auto also allows you to specify lower bitrate than for --rate keep.
qaac chooses SoX rate converter of libsoxr by default. If it is not present, or if you specify --native-resampler, qaac uses CoreAudio sample rate converter.
--native-resampler can take optional arguments to control sample rate converter complexity and quality. Without optional arguments, codec default sample rate converter is used. If arguments specified, qaac will use dedicated, configurable AudioConverter for sample rate conversion. Unlike other options, --native-resampler and it's arguments are delimed by a '=', and you cannot use a whitespace here. Complexity is one of line,norm,bats. Quality is an integer between 0 and 127. Complexity and quality must be delimited by a ',' (comma).
|line||linear (Worst, and included only for the completeness. Don't use this)|
|bats||mastering (Best, but quite slow)|
|--native-resampler||Use codec default converter|
|--native-resampler=bats,127||Complexity:bats(mastering) and Quality:127(highest)|
|--native-resampler=bats||Complexity:bats. Default quality is chosen.|
|--native-resampler=96||Quality:96. Default complexity is chosen.|
Last edited by nu774,