The Mike Moffat (#2 at Schiit) Blog

For music recordings, if the digital signal has been properly dithered then it's the same as analog, and perhaps even better depending on a number of other factors. Proper dither makes the noise completely uncorrelated with the signal: it's random compared to the signal. Analog sorta gets this for free, and digital has to work a little harder to get it. When you do, you can listen for signals as low as you like as long as you're willing to listen for long enough. Technically that means using a narrower passband (narrower in frequency means longer in time) so that you can hear signals buried very, very deeply in the noise.

Lots of modern communications systems work because of this principle, like JPL hearing Voyager's very weak signal sent from a zillion miles away. You can see this in action in FFT measurements posted here and elsewhere when the number of FFT buckets is increased, and the noise floor goes down. The tradeoff is that the measurement takes longer.

 

At least it wasn't an 8-track tape!

 

@baldr . The multi-bit module is a thing of beauty; nicely done sir.

Hopefully you have as much fun exercising your digital wizardry, as we do listening to the results.

 

Radiation would have destroyed that as soon as it left the van allen belts

 

Where is the DIP-8 op amp ROM?

 

It's not an OP AMP probably ROM for the SHARC.

 

There's a joke- 40,000 years from now we get a message from beyond the solar system. It says, "send more Chuck Berry ASAP."

Question- I know about dither, but how much digital audio is dithered? All, some, none? And where in the process does it have to happen to preserve the data? I'd think early on.

 

Like anything else, it's possible to always screw it up. Dither is necessary when you digitize in the first place, and also when you reduce word width (eg. going from 24 to 16 bits). The important point is that the noise floor is not correlated with the signal (there is a more precise math definition), so however you get there, it's good.

 

Arrived 3/22/18

SCH-06
B60053xx

There’s also a small sticker on the left near the RCA Output jacks labeled:
GUNGNIR MULTIBIT
MB01660

 

Congrats, enjoy. Btw, the numbers are structured as mine are, just ~200 higher.

 

May your auditory experience be as enhanced as the assets of your avatar (at least they seem to be).

 

Funny you mentioned the Avatar girl. I had to take that down on the S Hoffman Forum. Someone complained it was too risqué.

 

The raised eyebrow is very provocative.

 

Yeah, that’s it - the eyebrow!

@BIGGER Dave, congratulations on the Gungnir Multibit. I’ve had mine about 3-4 weeks, it takes the place of a Bifrost Multibit in an SE system. It is also responsible for a Freya ending up in my system ahead of a couple of headphone amps. The Gungnir Multibit was a bigger improvement than I expected. The Freya is an extremely over priced and overkill passive attenuator at the moment. Eventually a Vidar or similar will show up, and the Gungnir Multibit and Freya and main headphone amp will move down stairs. Ahhhh, I digress. Congratulations again.

 

I find your cup threatening.

 

I laugh every time I see that avatar. I use to have this poster on the door to my office.

 

I don't even know what "too risque" means. Is that a thing?

 

Sorry if this was already answered but did we figure out what Mike was referring to in the latest release when he said that he did something he said he'd never do?

 

Pretty sure it was the multibit card.

 

This could be its own thread but a few resources:




production oriented with examples:

https://theproaudiofiles.com/dither/

http://www.harmonycentral.com/articles/dithering-demystified

http://productionadvice.co.uk/when-in-doubt-dither/


more strict explainers:

https://www.soundonsound.com/techniques/digital-myth

https://www.soundonsound.com/techniques/digital-problems-practical-solutions#para8