So, the firewire is working, hey? I know this has been a real 'crusade' for you, so well done!

Now, what are the most obvious sonic benefits, as compared to the 18cm I2S lines that you've been using to date?

Are you using a similar firewire implementation to Weiss?

Mani.

Thanks Telstar.

This is really interesting stuff. One of the most valuable lessons for me is that 'anecdotal' is valid (and this from a credible scientific researcher).

Keep the faith Peter...

Mani.

Peter, thanks for taking the time to expound your thoughts - very helpful.

I'm not sure if I'm knowledgeable enough to agree or disagree, but one thing is for sure; I totally agree that a NOS 24/192 would be nice for playing 24/176.4 and 24/192 material - I don't see any need to manipulate the data in any way, shape or form at these resolutions.

As for 16/44.1, I've always felt that the 'old' multibit DACs/CD players with 8x oversampling were absolutely fine (with no upsampling required). But I can see that 4x upsampling with a NOS 24/192 would work nicely...

Mani.

When I have some time, I'm going to dust off my old 'Sampling Theory' texts from university. I'm absolutely convinced that 4x oversampling (with no upsampling) should theoretically achieve the same result as 4x upsampling (with no oversampling). (Let's just stick to integers, because anything else just complicates things further.]

The fact that we've had oversampling for over 20 years and upsampling for only 10 or so is because, in IMHO, the marketing people hadn't realised that they could scam consumers with upsampling, knowing full well that their DACs/CD players were oversampling anyway and just didn't need upsampling. Actually, to my ears, upsampling just messes things up with an OS DAC.

Now, if you have a NOS DAC, I can totally understand why you would want to upsample. But you'll just get the same result as an oversampled DAC, surely?

I'd love to be shown wrong 

Mani.

OK Telstar, that's cool.

But I'm still just a little confused. Peter, are you simply proving that if you upsample, you don't need to oversample... and if you oversample, you don't need to upsample???

This fits with my (limited) understanding, because oversampling and upsampling are achieving the same result, but at different points, no? FWIW, I've never liked upsampling, whether done in hardware or software... BUT, I've only ever used an OS DAC, where IMHO upsampling simply isn't necessary.

But your thinking on getting the output stage right is facinating...

Mani.

This sounds a bit like the Q1=-4 setting at the moment!



Peter, I would have loved to have seen the following included in your analysis:

1)  a delt-sigma "24"/192 DAC
2) a true 24/192 Multibit DAC... with oversampling switched in but without upsampling

My (limited) understanding is that there should be no difference between 2) and an upsampled NOS DAC.(If you'd like to borrow my D70, I'm sure we could arrange that. Is this a true 24 bit DAC?)

Mani.

Thanks Peter. This is facinating stuff!


So, my understanding is that XXHE and the hardware will work together seamlessly. But will the DAC be able to work with other software players and/or digital inputs?

Mani.

Peter,

I only use laptops, and find them absolutely fine...

"In particular, we've found that, with laptops, using battery power results in very good total correlated jitter similar to TosLink or S/PDIF (<200ps or lower at 44.1kHz, and with about the same noise floor), but using the typical switch-mode PSU with the same laptop can result in 3000–3500ps of jitter." - Matthew Bramble, Technical Director, Cambridge Audio.

Certainly, using XXHE on an old Sony laptop (with a 'real' TI 1394 chip) feeding a Weiss AFI1/PM Model Two I'm getting outstanding results... I really never knew that digital could be this good... ever.

Mani.

Yep, from a theoretical point of view, 300dB/octave filters should sound great. Try it with your DEQX and see what happens. I remember asking DEQX why they recommend sticking below 90dB/octave. They said it sounds better, but could give no theoretical reason why.

In my 'real' experience, nothing other than 6dB/octave sounds right... Of course, this creates all sorts of other issues, but then that's hifi for you - just full of compromises 

Mani.

If there is, let me know - I'm looking to go 4-way... all analogue!

Mani.

Sounds very cool OMF. I especially like the fact that you're using the DAC as the master clock and syncing everything to it - I really don't like asynchronous anything... or upsampling for that matter... or 1-bit DSD (but that's another subject!).

When I was doing something similar (but far less sophisticated) a few years ago, the biggest issue for me was controlling the volume. How are you doing this?

Mani.

(Peter, are you happy for me to place this link?)

Here's some info about the XVR1 x-over: http://passlabs.com/pdf/product%20lit/xvr1_lit.pdf

Cons:
- not cheap: were around USD 5-6K each when I bought mine about 5 years ago
- only 2-way: you'll need to buy a pair for 3-way operation and 'daisy chain' the output of one channel to the input of another
- not suitable for any sort of correction
- not sure if they're still in production
- I've never seen one for sale on used market (and I know why!)

Pros:
- very transparent
- much better than passive x-over from speakers (Wilson Benesch Chimeras)
- much, much better than DEQX 2.6P

Hope this helps.


Looks very interesting. Didn't you used to have a Meitner??? In any event, I realise that you need a 6 channel DAC, right?

As I said earlier, if you've managed to get digital x-overs to work well, then that's great. I just never did - and the XVR1s were a breath of fresh air...

Mani

Why? I think for the same reasons that XXHE sounds better than other SW players... why filter-less DACs sound better than filtered (with hi-res, at least). IOW, there's a lot more going on in the digital domain than we know... but that the likes of Peter are exploring.

In my real experience (i.e. not theory) my XVR1s have literally trounced digital x-overs like the DEQX. Actually, the difference in SQ was embarrasing...

I don't intend to sell the D70 - I love it. Also, the Pacific Microsonics stuff is really studio gear, and not designed for consumer use. It may not fare very well and may need to be shipped back to California for servicing from time to time... in which case I'll need something else to listen to. If I don't have a NOS1 by then, the D70 will have to do.

Actually, I'd love to do the following DAC comparison: NOS1 vs. Model 2 vs. D70

Mani.

If you're intending to use a firewire interface with W7 RC, you'll need to use the legacy driver...

Mani.

Hi OMF,

Good to hear the Finales are still in action - which is more than can be said for Overkill itself, right?

I can't really help you with your enquiry - a while ago I decided that I would do all my filtering in the analogue domain, hence why I use a couple of Pass Labs XVR1 x-overs. There is a 'rule of thumb' that I read on another site that I like very much; "analogue can't do delay, digital can't do filtering". If you've managed to overcome this, well done!

But in any event, XXHE is definitely the way to go...

... Now, if only I could find a decent filter-less DAC... Peter, how's the NOS1 coming along???


Mani