0
0

When I select the option to optimize the sample rate of my wave banks in Designer, what exactly is happening during this process? Does it resample everything to 22.1kHz, or is it more complicated than that? I’m entertaining using sample rate optimization and ADPCM compression to create small, CPU-light streaming music banks, and while I’m very impressed with the results so far, I’m interested in knowing what exactly Designer is doing during the conversion process so that I can better prepare my source WAVs to endure the compression.

  • You must to post comments
0
0

Hi –

Pleased you’re finding it effective.

It isn’t just resampling everything to 22.05kHz.

It’s trying to seek the lowest rate at which the difference between the resampled version and the original is below a threshold value.

In terms of being able to prepare the sounds for compression, if you really want to, you can reduce the rate at which it resamples by applying a fairly strong LPF (say >6th order) at a frequency above which you consider the signal unnecessary. However there are two caveats:

1) The lowest resample frequency will be governed by the format in which the soundbank is to be stored. Fortunately for ADPCM the minimum rate is 4kHz so there’s lots of room at the bottom. However for MP3 it’s 32kHz (so 16kHz is the lowest you’d want to filter).

2) The optimiser has been designed to do this task for you. There may be cases where you might want to preempt it but it’s your choice.

  • You must to post comments
0
0

[quote:2s8a4u7u]It’s trying to seek the lowest rate at which the difference between the resampled version and the original is below a threshold value.[/quote:2s8a4u7u]

Can the difference threshold be tuned?

  • You must to post comments
0
0

Also, would it be possible to set a different difference threshold for different sound types? One for dialog, another for music and a third for SFX or something of the sort? By the way, thanks a bunch for this feature! This made my morning.

  • You must to post comments
0
0

No, the key is to pick the optimal sample rate, before it sounds lossy. At the moment it is designed so it is non lossy, so that it doesnt pick a sample rate below the rate that it should be.
This is similar to doing an fft and finding the lowest sample rate possible before you start losing information. There isnt really a ‘fuzzy’ answer, unless you want to purposely resampler lower than the optimal rate, and actually make the sound worse than it should.
Letting people adjust our acceptable threshhold just leads to abuse and confusion/misunderstanding of what it does, so we’d rather keep it out. Just pretend it doesnt exist.

  • You must to post comments
Showing 4 results
Your Answer

Please first to submit.