Conventional DAC

Mr. Kiwi.I'm pretty sure we have the same Theta stuff (gen5 & Data 3)-right? My remark mostly refers to said "whatever"..sans...and with 24/96. I remember a while back you were mentioning the distance to Theta from NZ,and how long you'd be without the dac Then I told you I live 15 mi. away; and had the 24/96 done to mine. Well this is a personal opinion;but it is the most gain for the bucks(300) I ever got. Or: a Gmc671 blower may not help your family sedan. 24/96 just helps kick it up another notch if everything else is there. Never said a $400 dac w/ is better than a $2000 w/o.

Now I get you Avguy, and concur that the 24/96 DACs have something very different and musically almost magical compared with earlier DACs - when all else is equal. My DAC did get to do the round trip by the way - thanks to your recommendation. I even upgraded my Meridian 566-20 to 566-24 on the strength of the experience - also very worthwhile.

Agree with Rewiiki ("there is more to a good DAC than a good DAC chip")--most manufacturers pay about $5 for a Burr-Brown chip after their quantity discounts (that is, most digital audio manufacturers realize superb economies of scale with chip manufacturers, maybe the reason for the CD revolution int he first place). I have found the most important parts of a DAC to reside in the analog output stages and the power supplies. I prefer DACs with tube output stages (esp. Manley, Audio Note) and stay away from the op amp output solution. The numbers game (16bit/20bit/24bit) has not been as crucial and I--like Redwiki--find 20 bits adequate. Have experimented with D-to-D Bit conversion and it has resulted in lowering the noise-floor, through dither and noise-shaping, but not enough to the point that it makes digital much less fatiguing to listen to.

At the risk of being boring on the technical side, there are real problems with PCM when trying to up the word length. The DAC chip has to incorporate resistor values that start with a value and then halve the value for the next resistor, then halve that again, etc till the 24 bits are provided. Providing these resistor values accurately on a chip and those values staying stable over time and with changes in temperature is very expensive and not really achieved accurately in standard chips. Therefore there is distortion and noise created (called non-linearity). This is one of the downside problems of pushing PCM up to the 24/192 standard. Sony's SACD uses DSD, which avoids this problem but creates its own problems. Noteably dCs (Elgar, Purcell, Delius) uses a mixture of DSD and PCM (putting it crudely) to avoid the problems both have when you push either technology to its limit. Sure the technology you employ is important, but the way it is employed is still likely to be the big differentiator for a long time.