advertisement


What do DACs do with non-parity frames?

"The AES3 standard allows for the flagging of concealment errors by the validity bit."

So?

That's concealment then. Helps you a lot in counting interpolated errors. For that you have to get back to the transport.

But anyway, and in line with the OP, I was referring to the single parity bit in AES/SPDIF, which doesn't detect multiple bit errors.
 
Werner said:
"The AES3 standard allows for the flagging of concealment errors by the validity bit."

So?

That's concealment then. Helps you a lot in counting interpolated errors. For that you have to get back to the transport.

But anyway, and in line with the OP, I was referring to the single parity bit in AES/SPDIF, which doesn't detect multiple bit errors.
Click to expand...

Just to clarify, statistically, the parity bit should allow detection of 50% of frames which contain one or more bit errors, presuming a degree of randomness in their generation. More accurately, it allows detection of any odd number of bit errors per packet (sorry item, I meant frame! ;-). Considering their apparent rarity, it seems a bit academic anyway, since as has been said above, significant levels of bit errors will lead to frequent drops-out rather than audible value-riding.
 
item said:
Interesting, but I don't see how this would be relevant. What is your proposed mechanism for discrete (and rare) parity errors making a characteristic difference to a DAC's performance?
Click to expand...

I'm just keeping an open mind about your experience. It might be that there are often logical reasons for your hearing differences between transports. I think it makes sense to rule out logical discrepancies in the digital side before postulating analogue gremlins. As I've said before, until two sources are shown to checksum, searching for the source of audible differences in the analogue part of the signal might be a wild goose chase. I'm not convinced you've eliminated the bits as suspects yet. Or poor DAC implementations.
 
BTW, from a software point of view it seems to me that a buffering DAC would best handle erroneous frames by averaging the preceding and following samples rather than repeating the last good sample.
 
I applaud your investigative spirit, but suspect you're barking up the wrong tree!

Having said that, I've been wrong-footed too many times by ignorantly discounting the relevance of the unlikely to be sure that what you're proposing is impossible. I tend to assume that the bits are not the problem - more likely the relationship between the transport and DAC.
 
sonddek said:
BTW, from a software point of view it seems to me that a buffering DAC would best handle erroneous frames by averaging the preceding and following samples rather than repeating the last good sample.
Click to expand...

Do you think the difference would be audible?
 
Also, let me make one other thing clear. I do expect the amplification stages of DACs to sound different. I even think my DAC sounds very slightly different depending on which power supply I use. But that's not the contention. The issue is whether two bit-perfect transports will sound different playing into the very same well-designed DAC. To say they don't suggests a pretty rigorous approach to ensuring that the bits are correct first.
 
The only rigour required is a bit-perfect test, trivial with any pc front end.

Do you seriously think there are cd players out there that are not bit perfect?
 
Si - only one I can think of is the Philips CD723, a budget model which IIRC did something slightly odd in the decoder for other reasons which had the side effect that the SPDIF data content was not what came off the disc.

A very rare issue though.
 
I would assume so Martin, dvd players et al might not be bit accurate, but I guess those aren't part of Sondeks remit.
 
Werner said:
"The AES3 standard allows for the flagging of concealment errors by the validity bit."

So?

That's concealment then. Helps you a lot in counting interpolated errors. For that you have to get back to the transport.

But anyway, and in line with the OP, I was referring to the single parity bit in AES/SPDIF, which doesn't detect multiple bit errors.
Click to expand...

Except for the small fact that an interpolated error is a concealment error.

And that the set parity bit does allow detection of multiple bit errors.

And that you were not talking about the parity bit at all, you were talking about whether read errors could be flagged or counted on the AES digital interface.

It's chilling at times how you spout nonsense with such self-appointed authority.
 
sq225917 said:
Werner the original cd 1 was a 3 box unit, with a little display light that flashed for a unrecoverable read error. Similarly the PM dac unit had an error display on it, though I'm not sure what the trigger was for that, obviously not read errors. I get that this isn't what Sondek is asking for but these are the only two commercial units with data error indicator lights on them, IIRC
Click to expand...

You can add the Meridian 207 & 207Pro.
 
You must log in or register to reply here.


advertisement


Back
Top