Originally Posted by Rob Sinden
I've posted the document about the technical benefits of the system working in the digital domain. If you choose to ignore it that's fine.
You don't seem to understand. You aren't "working in the digital domain". What you are doing is adding a DDC into the mix, and then changing the wire length.
Source -> Decoder -> DAC -> Amp -> Driver.
Source -> Decoder -> DDC -> DAC -> Amp -> Driver.
The DAC is physically located in an assembly adjacent to the speaker cabinet; and the speaker is active. That's not typical, but hardly significant as all it does is shorten the speaker cable (compared to a conventional setup), or the analog interconnect length (compared to an active speaker).
That's not nothing in a very long run (like in a full-sized theater). It's analogous to why pro gear (also often long runs) uses XLR cables (high voltage) and active speakers as often as they do... but in a home / home-theater environment; there's not enough down-side to speaker cables (which attach to far longer, far thinner cables at the voice-coil) to experience a performance gain.
I've no doubt these speakers sound good. Likely even great. A high-SPL open-baffle system in a well setup room should blow off socks. But it is not related to their choice to run active speakers or use proprietary interlinks.
What part of the science in the document I enclosed do you disagree with?
Well, let's start with their comparison of bit-depth data.
They accurately relate bit depth to dynamic range (max SPL) with their (rather arbitrary) "quality at volume" chart; but in order to relate that to the actual system we have to engage in a few lies.
We have to lie and say that a 24-bit source in a standard system would be down-sampled to a 16-bit source (it would not: a 24-bit rig would convert to analog, which has resolution bounded only by signal strength).
We would also have to lie and say that DDC on a 16-bit source to a 24-bit interconnect would improve useful dynamic range to that allowed under 24-bit. It would not do so. (your extra bits would contain no data)
Then there's complete snake oil like "distinguish between embedded noise". There is no noise introduced at any resolution in a digital signal. Noise would come from pre-ADC, and even 64-bit should reproduce that accurately, not lessen it. There's a failure to understand digital waveform encoding there.
I just saw that actual bottom of the PDF.
To make their case: they assume
1) That a digital equalizer has been inserted between the DAC and the amp (the only spot in any of my gear that does this is my DCX2496).
2) That the connection between the DAC and EQ is analog (I suppose necessary since post DAC)
3) That the connection is 3v (my DCX uses XLR connectors which are higher voltage)
4) That the ADAC steps are not 24bit + (they are 24-bit in the DCX)
5) that there's a 20x increase in the noise-floor caused by the EQ (there isn't).
I too can turn my volumes up to ear-shattering levels and then (with no sound) stick my ear to the speaker and not hear a hiss... which was , you may recall, the test proposed by the OP.Edited by JerryLove - 11/16/12 at 12:46pm