Print

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

[d_pert]

Hi there,

Was wondering if there's any kind of diagram illustrating the signal path through MC (incl. various DSP stages, upsampling, downsampling, mix-up, mixdown etc.) right to the sound card device.

I'm an avid MC user and love it. Just want to be sure of what I'm actually getting in the end when I mess around with 24/96 VST/DX plug-ins etc. Particularly, I'm interested in the Output Format: it's influence, at what point in the chain, etc.

Thanks!

DP

[Matt]

Output format conversion happens at the front side of the chain.  

DSP's happen just-in-time as data is delivered to the output.

The entire chain is always 32-bit, and only down-converted to a lower bitdepth (or recently encoded as Dolby Digital) as a final step if the output requires.

[keither]

I wonder if, for fidelity, output format would be best as the last thing in the chain.  Hmm.

[d_pert]

Okay so let's say:

I play a 16bit WAV. I have a 32 bit VST plugin running in the DSP studio. The VST recieves an "upsampled" version of the WAV (to 32), processes in 32 (naturally) and then passes it to the output stage of MC. If I had a firewire DAC operating at 32 bit, and the MC output mode was set to 32 bit, the DAC would get 32 bit (it being understood that the VST "wet" signal portion is the only meaningful 32 bit resolution information in the stream)?

Let's say the above ends with an output setting of 24 bit instead. What kind of conversion 32 to 24 is happening? are you using some kind of dithering to make some use of the LSB (bottom ? Are they being dropped?

Just interested...



Thanks.

[Matt]

Let's say the above ends with an output setting of 24 bit instead. What kind of conversion 32 to 24 is happening? are you using some kind of dithering to make some use of the LSB (bottom ? Are they being dropped?

No dithering is done for 32-to-24 bit conversion.  You can find huge arguments about this with smart people arguing both sides.

I come down on the side of no dithering, just rounding.

But it's important to remember that 24-bit or 22-bit or 20-bit sounds provides more bits in the final mix than a human ear can hear.  The final bit being rounded, cropped, or dithered is an academic issue.

My ears used to be able to hear to about 15-bits with fancy headphones in a silent room with complete concentration.  Even this might be an exaggeration since I was using a Creek headphone amp that boasts >70 dB S/N ratio -- about 12 bits.

I don't mean to downplay Media Center's 32-bit playback chain, because I think it's a great thing, especially when you're using DSPs.  JRiver may have made the first 32-bit playback chain in the consumer market, back in the Media Jukebox (v6?) days.

[Matt]

I wonder if, for fidelity, output format would be best as the last thing in the chain.  Hmm.

I don't think so.  Remember that all data is 32-bit through the entire chain.