in praise of Bias Peak's Samplerate Change

[Syn-Fi]

Bias Peak 5.2 appears to have a very intensive Samplerate change algorithm that can take 1h30mins at 'level 10 quality' to convert 20 minutes of 24bit 48khz audio to 44khz on a G4 MDD 1.25ghz

But it is worth it and i have not been satisfied with Soundforge or other tools for this purpose.

attached whitepaper.

[IIO]

i was always wondering what it is doing. barbabatch was usually the tool of choice for pro audio.

[cyberish]

interesting... - and this 5.2 version would be better than the algos in version 6?

[Syn-Fi]

hi,

ver 6 retains this function and i have ver 6 on Mac Mini i7, so i should test how quick it is on that setup with the algorithm set to '10' in preferences. 

Samplerate change is particularly important given how many people record at higher sample rates to edge up Nyquist but then need to downsample for CD or for vintage samplers.

PowR3 dither is also on Bias and i prefer this dither when i'm dropping from 24bit to 16bit.

[IIO]

i havent read the paper, did he set up something custom?

i am under the impression that 4 times upsampling and using bicubic interpolation should give perfect results already.

my personal "trick" was always to avoid rates which are not multiples of 44.1.