I think we should go over this again
|Originally posted by Cor
If you look at a DS PCM chain from analogue to analogue it looks like this:
Analogue -> 1-bit ADC -> DS Stream -> Decimator -> PCM samples -> Interpolator -> DS Stream -> 1-bit DAC -> Analogue
I was talking about DAC's above, so an Interpolator was the correct beast. You shouldn't find a Decimator anywhere in a DSPCMDAC.
You are confusing terms here, a decimator
and an interpolator
are different from a differentiator
and an integrator
, I did not mention a decimator
, I said a differentiator
, A differentiator is not a decimator, A decimator is the inverse of an interpolator and neither is used to convert from PCM to SDM stream or vice –versa, their use is in oversampling and downsampling. Do not forget what I mentioned earlier SDM words are relative values(differential values) and PCM words are absolute values, so you cannot apply an interpolator in the manner that describe, the is how the block diagram should look
Analogue -> DSD ADC -> DSD Stream -> Integrator-> PCM stream -> Interpolator -> PCM Stream -> PCM DAC -> Analogue
Or to use a 1-bit for Digital to Analogue conversion you have
Analogue -> PCM ADC -> PCM Stream -> differentiator -> DSD stream -> Interpolator/oversampler -> DSD Stream -> DSD DAC -> Analogue
|In theory, 24-bit 96 KHz. PCM does actually contain more data than 1-bit 2.8224 MHz. DSD. PCM also losslessly compresses more. (Both DVD-A and SACD employ lossless compression algorithms) Hence, given the same raw capacities, you *should* be able to fit more music of a given quality onto a DVD-A. Of course, we're splitting hairs here. Most discs aren't close enough to being full for this to matter.
If you think in terms of the previous paragraph that is DSD relative values and PCM absolute values, your logic here breaks down completely. And lastly even the most ardent proponents of PCM acknowledge that 1-bit SACD is superior to 24-bit 96 KHz PCM, the point of dispute is actually 24-bit 192 KHz. The actual issue that you should be thinking here is how much playing time can you achieve for a given amount of data.
DVD-A's can be authored in 96/24, 192/24, and several other bit-depth/sample-rates. You could author a DVD-A in 16-bit 44.1 Khz if you really wanted to. With SACD's you have one option: 2.8224 MHz DSD. My point is that while DVD-A at it's best competes well with SACD, it is possible to author DVD-A's that are grossly inferior just by choosing a low bit-depth/sample-rate. Yes, it's a damned silly thing to do, but never underestimate peoples' capacity for stupidity.
Your logic is here is faulty, think of the joys of oversampling that mean you can author any of the PCM bit depths that you have mentioned to DSD. In fact this is where some of the misconceptions about SACD lie since a lot of folk are oversampling low bit depths PCM encoded material to DSD and publishing SACDs and some of these SACDs give DSD a bad name. Secondly I hope you are not unaware of the fact that professional DSD is now 8-bits as opposed to 1-bit. Lastly there is nothing stopping anyone from actually recording in 5.6 MHz SDM if they so desire, as I said earlier as long as your final SACD is 1-bit 2.8 MHz all is well. Personally I fail to see how being able to author to low bit depth and sample rate confers any benefit to DVDA, it is more like an albatross. The flexibility that you expouse here makes it easier to get it very wrong.
|Originally posted by Joe Bloggs
5.1 DVD-A is restricted to 24/96 because that's the limit specified in the DVD-A spec, and that's the only reason. Even if DVD-A were actually more inefficient at coding the format could have specified a higher reading speed for the DVD drive to pick up the data fast enough for 5.1 at 24/192.
Thanks Joe, I know that the 24/96 MCH restriction is specified by the DVD-A spec but the reason for the restriction is because the current DVD specification cannot support 24/192 data throughput for MCH. Your second point is mute though, if you had a higher reading speed, SACDs will not be stuck at 2.8MHz, they will be running at least 5.6MHz and the issues of ultrasonic noise and equivalent bit-depth pushed by the DVDA camp will be muted.
PS:let me state, I made a factual error in my last post, instead of differentiator I ought to have said integrator.
EDITED: Additional information.