Light Harmonic Da Vinci DAC

[wizard]

View attachment 4967

 

[Bruce B]

I will be using one at RMAF this year in the Usher/JPS Labs room.

 

[opus111]

I see they use no digital filtering - its effectively a NOS DAC, which means with no analog post-processing it will have HF droop to -3.2dB @ 20kHz on redbook material. Anyone come across its frequency response measurements to confirm?

 

[LL21]

I see they use no digital filtering - its effectively a NOS DAC, which means with no analog post-processing it will have HF droop to -3.2dB @ 20kHz on redbook material. Anyone come across its frequency response measurements to confirm?

Wow...that was quick! You are all over the digital...that is great for the rest of us who enjoy learning from your analysis of these designs! Thanks, Opus.

 

[opus111]

Yes, I'm drawn like a moth to a candle flame when expensive new DAC designs come to my attention, to see if they contain any new innovations Here I went to the website and picked up some details but they don't talk about how they did the R2R DAC. Also they focus quite a lot on jitter which for multibit DACs isn't really an issue - its a problem that plagues S-D type designs though as I've just realized from posting on the Vivaldi thread!

 

[MylesBAstor]

I've asked Gavin Fish, the National Sales Manager, to drop by and say a few words. Hopefully, he'll add more to the discussion.

 

[MylesBAstor]

Is 3db down at 20Khz state of the art? That should result in lots of loss sound for second harmonics of instruments playing even at 10Khz, thats not too good , and for $20K, oh well. Not that I am against tone controls, but given people over 45 or so already are way down in their hearing, this aint gonna help much with "air"...

And what about that happens at 22 kHz?

 

[Bruce B]

And what about that happens at 22 kHz?

Guess we sound ban Redbook!

 

[MylesBAstor]

Guess we sound ban Redbook!

Perfectly fine with me

 

[Bruce B]

Perfectly fine with me

Ditto..... you're preaching to the choir!

 

[opus111]

Is 3db down at 20Khz state of the art?

No - there are other NOS designs that fix this up - like AMR's for example. The problem is not so much the 20kHz droop (where few of us oldies can hear) but that the NOS roll-off begins around 5kHz. A fair proportion of HF energy ls lost over the two octaves from 5kHz.

That should result in lots of loss sound for second harmonics of instruments playing even at 10Khz, thats not too good , and for $20K, oh well. Not that I am against tone controls, but given people over 45 or so already are way down in their hearing, this aint gonna help much with "air"...

Getting NOS flat does indeed help with clarity and 'air'. Uncorrected NOS is coloured.

 

[opus111]

I dont know who is going to hear a harmonic at above 22khz on normal music, what about it?

There is another issue with plain vanilla NOS - that it puts out non-harmonic frequencies above 22kHz owing to lack of an anti-imaging filter. While we can't hear these frequencies I do have a suspicion that they're not doing anything to help improve tweeter linearity.

Redbook is slowly dying out and in 30 years will be just like LP is today, as Tim mentioned, the "in thing", atleast for a small sector of the audiphile population.

I hope if I'm still around in 30 years I'm still able to listen to my considerable collection of redbook material I'm very unlikely to 'trade up' on 44k1 material I've already bought but some of the newer lossless multichannel stuff coming out on BluRay looks attractive.

Hearts Heart album Gratest hits on CD sounds damn good to me. CD can and does sound good, NOS or over sampling or whatever. LP sounds good and so does tape. But man, I will not buy a source component that can not get the basics down in digital land.

Agreed - but then I would say that as I'm developing a quasi-NOS (in sound) design which does fix up the frequency response

 

[dallasjustice]

There are more than a few folks that like the B&K House Curve which also is a HF roll-off. Some folks think totally flat in the HF sounds annoying. If you don't like the HF roll-off, you can oversample to 176k on the server side and get flat response but probably more ringing. There are trade-offs. You could build a filter to compensate for the droop, but said filter would cause some ringing even if it were an NOS filter. To me, it's an overblown criticism of NOS DACs.

 

[opus111]

Some folks think totally flat in the HF sounds annoying.

I suggest those folks most likely have other problems in their set-ups - like sibilance which means they want to tame the HF because it sounds harsh.

If you don't like the HF roll-off, you can oversample to 176k on the server side and get flat response but probably more ringing.

The ringing is dependent on the kind of filter used - many (probably most) filters use a linear phase response which gives both pre- and post-ringing. However its quite possible to upsample (oversample) using a minimum phase filter. The result then is just post-ringing which is most likely masked, if its at an audible frequency (which seems unlikely).

There are trade-offs. You could build a filter to compensate for the droop, but said filter would cause some ringing even if it were an NOS filter.

Yes all engineering is about how to handle the trade-offs. Some ringing is inevitable with a droop correction filter because it typically requires high Q to follow the inverse slope accurately enough. Introducing such a filter though increases the non-harmonic components (above 22kHz) which are already an undesirable artifact from NOS.

To me, it's an overblown criticism of NOS DACs.

Meaning you prefer the sound of droop to the correction methods you've so far heard?

 

[dallasjustice]

Meaning you prefer the sound of droop to the correction methods you've so far heard?

Yes. I can only relate to my experiences. My Dac has FIR HF compensation filter that produces flat response with redbook. It only adds a small amount of ringing. I prefer the filter off. In my case, even a small amount of pre-ringing is less desirable than some HF roll-off.

 

[opus111]

It seem to me your listening experience is good feedback for Vincent. Has he confirmed that his NOS correction filter is linear phase? In which case we could suggest to him that he provides a minimum phase solution. A min-phase correction filter will be shorter (in terms of taps).

There is though another possibility for the sound degradation - all digital filters introduce round-off noise so that may be why you prefer the sound with it switched off. As soon as a filter's in the path the data is no longer 'bit perfect' at the DAC.

 

[dallasjustice]

You would have to ask Vincent. I don't know. However, I strongly believe that NOS offers a more long term solution. IOW, we could let the computer do what it does best, crunch numbers and let the DAC do what it does best. I can't keep up with filter panoply. Who knows what filters can be created in the future? Playback software is just starting to incorporate some of these filters. I think it's great to give folks a choice including the ability use no filter.

 

[opus111]

I'm with you on offering a choice, though some choices are too bewildering for end users, so fewer choices is normally better. Ayre for example offers the choice of 'measure' vs 'listen' on some of its products. I don't though take the view that users who prefer not to use a computer as their source should be penalized so in my designs I prefer not to rely on external processing to get a flat response. To me that's like having built-in EQ which can't be overridden. I take it that a lot of (not all) customers would like to follow the 'official' (i.e. the designer's) recommendation for their filter setting Putting the onus for that choice on the end user is asking rather a lot.

 

[wizard]

From Larry Ho

Da Vinci will support DSD64 and DSD128. (but no DSD256)

We actually put TWO DAC's core into one chassis. Use best R2R structure for PCM and delta-sigma for DSD conversion.
This new "version" of Da Vinci is coming out and will demo in RMAF.

 

[Larry Ho]

I will be using one at RMAF this year in the Usher/JPS Labs room.

Yes, Bruce.

Let's have your good music there...