SRC (Sampling rate conversion) after Loundness Normalization

[kavithaig]

Hello,
The resampling module integrated in the libiamf code is called before loudness normalization. Could we reposition resampling as a module which changes the sampling rate after all the complete IAMF processing is done? Basically Mixed output->(if)Loudness_normalization ->(if)peak_limiter -> resampling . Would that be ok?

Thanks!

[yilun-zhangs]

Hi @kavithaig
Do you have any special reason to adjust the resampling position to the end?
The reason why we leave the limiter to the end is any processing may lead to signal clip.
So we ensure all processing modules are before limiter.

[kavithaig]

Hello @yilun-zhangs ,
We have a generic SRC and we wondering if using that in the end would be an option. We will run tests and make sure we pass PSNR for sure. If we don't pass PSNR definitely we will fall back to your architecture and use the same algorithm you have used. This is to understand our options in case code footprint needs to be optimized.

Thanks for the quick reply..