Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[lasker98]

I'm wondering about the processing order for effects in DSP Studio. I understand the effects are processed in the order they're listed, and the effects can be dragged to re-order. I see Output Format and Volume Levelling can't be re-ordered, they're always performed before any other effects.

When using convolution, it would seem it would be better performed before any Volume Levelling? I have no idea of the mechanics of how each of these effects work, but thinking that convolution would/should be performed on as full range signal as possible? And volume levelling best performed on final signal? I know convolution was probably the most recent of those effects to be added to DSP Studio, so was wondering if the ordering of processing was more of a default because of that, or if there is a well thought out, valid reason for the convolution being processed after volume levelling.

Thanks,

Bill

[Matt]

With JRiver's 64-bit audio chain, the ordering of effects is mathematically irrelevant.

Here's an example showing that millions of volume changes are bit-perfect at 32-bit (the highest hardware bitdepth):
http://wiki.jriver.com/index.php/Audio_Bitdepth#Bit-Perfect

[lasker98]

Thanks Matt,

Interesting information in that link (the C++proof part flew right over my head though). Based on your response and reading that link, it seems there would never be a technical reason for re-ordering the effects? Is that correct?

After looking at the page you linked, I have new insight on output bitdepth. My dac accepts 24 bit data, but I had output format>output bitdepth in DSP studio set for source bitdepth, which more often then not would be 16 bit. This was based on the idea that source bitdepth would be "bit perfect". After reading that link, it seems obvious it's best for me to use 24 bit for output bitdepth. Is that correct?

Thanks,

Bill

[Matt]

Interesting information in that link (the C++proof part flew right over my head though). Based on your response and reading that link, it seems there would never be a technical reason for re-ordering the effects? Is that correct?

There are sometimes good reasons to reorder, but it's not about bitdepth precision.

For example, you want to apply timing effects (Room Correction / convolution) at or near the end.  This is because these change the timing/phasing information.  And some other effects might not work well with audio if the timing is goofed up.

Another example is using DSP to bi-amp.  At the end of my DSP chain at home, I copy the subwoofer line to different channels.  But I want to do that last, since all the way through the chain I want the subwoofer in the right place so something like bass management in Room Correction will redirect bass to the right channel.

After looking at the page you linked, I have new insight on output bitdepth. My dac accepts 24 bit data, but I had output format>output bitdepth in DSP studio set for source bitdepth, which more often then not would be 16 bit. This was based on the idea that source bitdepth would be "bit perfect". After reading that link, it seems obvious it's best for me to use 24 bit for output bitdepth. Is that correct?

That's exactly right.  Always pick the highest bitdepth your audio device supports.  ASIO does this automatically, but WASAPI requires a user setting.