Friday, May 23, 2014

NOS Dacs

I don't know how much oversampling is used with the Burr Brown PCM 1704 in my Onkyo RDV-1 I use for decoding digital for my Acoustat speakers.  The manual doesn't say.  I'm inclined to think that one of the advantages of R2R Dac's is that they don't require the huge amount of oversampling that Sigma Delta Dac's require.  (Even if they are not used in an oversampling circuit, Sigma Delta Dacs have at least 64x oversampling and usually much more now built in, that's how they work.)  But it may also be that R2R dacs simply have more resolution at each point they generate, and they sound better because of that, and the oversampling as such may be immaterial.  Sigma Delta DAC's have 1-5 bits of resolution, and all the rest is made up (faked?) with oversampling.  One other question I have: do Sigma Delta DAC's use feedback to get correct results, or is the oversampling all that's needed?

Update: I think it's quite likely Onkyo RDV-1 uses 8x oversampling, because Burr Brown made a digital filter, the DF 1704, specifically designed for the PCM 1704.  And the DF1704 does 8x oversampling.  It also accepts inputs up to 96kHz.  So it's a perfect match with what I've got.

Anyway, now that I've started to go off the Objectophile ranch (they say that all DACs sound the same, and also that there is no reason why R2R DACs should sound better, and they could be worse because they have poorer measured linearity, and they don't like that idea that you would go with something that measures worse because of your superstition, though they still maintain they all sound the same) with a R2R DAC, I've been thinking about the non oversampling approach too.  Many appear to confuse the two concepts, oversampling vs non-oversampling, and ladder R2R vs sigma delta.  But most ladder R2R Dacs (as equipment) do use some oversampling, they only raise the possibility of getting away with none.

Here's a great thread which lists a bunch of non-oversampling Dacs (equipment, not chips) you can buy.  Some of the ones mentioned at semi-affordible prices are:

47 Laboratory Shigaraki Series Model 4715 DAC
Audio Note DAC Kit 2.1, DAC Kit 3.1.
dB Audio Labs Traquility and Tranquility SE
Metrum Acoustics Octave and Hex
MHDT Laboratory Havana
Mojo Audio AD1865 NOS DAC
Promitheus Audio Solid State DAC and Tube DAC
Red Wine Audio Isabellina
TeraDak TDA1543 NOS DAC V3.1D
Wavelength Audio Brick v2
Vertex AQ Alethia DAC

Another one, sometimes considered a benchmark, is the Light Harmonic Da Vinci 384k ($20,000), not to be confused with another maker's Da Vinci product.

TeraDak also OEM's other products, like the Valab 1543, and those two units are among the least expensive NOS DAC's.  I don't like the fact that they use transformers for I2V and output coupling, though it is one way to get around the ultrasonics (a problem with a lot of NOS DAC's).

DAC's like NuForce and Peachtree claim to be NOS, but use Sigma Delta DAC's which use oversampling internally.

Here's a DIY thread about NOS DAC using 1541A.  This thread at DIYAudio has become quite popular, growing to over 500 pages!  NOS DAC's may be mostly "underground", but they are a very hot item among a certain class of DIY and other underground audiophiles!

I remember seeing one NOS DAC tested in the pages of Stereophile.  The objective measurements didn't look good.

Update: I believe I have now read some where that Sigma Delta DAC chips do indeed use feedback to get linearity.  So at the end of the day, this is glorified mud!  Throw enough mud on the wall until it weighs enough.  Not too much different from PWM systems.  I'm convinced now.  Give me real PCM!!!  Oversampled or not, it's still better than mud!

No comments:

Post a Comment