What does it actually do?

P

Powertech

Super Member
I have a Cambridge Audio Azur 550C CD player. It has a Wolfson WM8740 DAC inside it. All the brurb says that this DAC is '192KHz capable'. But does it actually upsample from the CD 44.1KHZ to 192KHZ and then carryout the D to A conversion from that bitstream? Or does it actually do something else? I can't find anything to give a clue as to how these DACs are implemented.

I am thinking of putting another DAC behind it, but I want to make sure that what I am doing is actually likely to be an improvement on the original setup.

I have heard good things about the Behringer Ultramatch Pro data converter and I was thinking of buying one to put behind the CD player and then either using the internal DAC or feeding that into something like a Topping D30 or D50.

Any thoughts?
 
I would dig out the data sheet for the Wolfson DAC, and compare the filter performance with the DAC filter settings.

After all, it is likely that they are using the reconstruction filter in the DAC, and nothing else.

The 24/192k 'capable' is just puff.
 
So am I right in assuming that the CD player's internal DAC will see the 44.1kHz bit stream and carryout analogue conversion directly on that?
My understanding is that the Ultramatch will reclock the input bitsream and upsample it to 192kHz because you have told it to do that. Then it's internal DAC (or a further external DAC) will convert on the 192kHz bit stream.
Have I got this right?
 
It's possible that the Azur is upsampling by 4x to 176.4kSa/s. You'd have to dig further into reviews and tech descriptions. The DAC is very likely to upsampling as part of its operation, and provide a number of different filter options, so you'd have to read the data sheet, and somehow find out what options have been selected for the Azur design.

Upsampling isn't a trivial operation, and the filters used can influence the sound. Non-integer upsampling is even more interesting (e.g.44.1 to 196). The 196kSa/s spec of the DAC in the Azur is intended for 48kSa/s sources. It's unlikely to be used in a pure CD player.
 
So this might explain why the Ultramatch is getting such rave reviews. It sounds like there is another set of steps in the process with this type of unit where it takes the input data, reclocks it and then upsamples it to a higher frequency and word length. I understand that the upsampling system is a kind of interpolation process, a bit like filling in the sampling gaps that the original 16 bit 44.1kHz sample process missed from the original signals. Then, instead of doing D to A on the original 16 bit data stream you are now doing D to A on 24 bit data stream at a much higher sampling rate. Surely this has got to give a much smoother analogue signal with higher definition as an output. I have been reading up on this and I realise that I am kind of reading between the lines and doing a bit of guesswork here.
I am just trying to get my head around what is going on as I start on an upgrade path.
 
O.K. I've gone for a Behringer SRC2496 Ultramatch Pro. For the price it's worth having a go with this. Who knows, it might be a good call.
 
Upsampling is merely a technique to ease the design of the reconstruction (anti-aliasing) filter used after the DAC

In an ideal world, a DAC would be followed by a perfectly flat and phase linear brick-wall anti-alias filter. Such things do not exist, so real world filters struggle to provide adequate alias attenuation, whilst remaining flat and linear.

Upsampling pushes the alias away from the passband, easing the requirement placed on the filter.

The simplest upsampling is zero-stuffing: insert zero-magnitude samples between each sample

2x s0s0 etc
4x s000s000 etc

But, as you observe, it is then possible to apply an interpolation filter to fill in the gaps, providing a smoother (but entirely synthetic) sequence of samples to a higher precision DAC. This interpolation filter cannot reconstruct lost information from the original recording, that was lost when it was sampled to CD. It is 'making it up', based on the extant samples, and the filter algorithm.

There are loads of different interpolation filter designs, and each will have an influence on the sound.
 
The Wolfson processes the digital signal at the values sent to it with no upconversion to higher bit or sample rates being done.

The technical document can be found here:

https://statics.cirrus.com/pubs/proDatasheet/WM8740_v4.4.pdf

If you want to add an external DAC, use a digital coax or optical toslink cable from the CD player to the DAC to allow the external DAC to process the digital signal.
 
cpt_paranoia said:
Upsampling is merely a technique to ease the design of the reconstruction (anti-aliasing) filter used after the DAC
Click to expand...

"Anti-aliasing" shouldn't be there. It is indeed a reconstruction filter. Aliasing happens during the a/d proces, not during d/a, so an anti-aliasing filter is only present in recorders, not players.
 
O.K. I bought a SRC2496 from Amazon which was delivered overnight. After a quick play with it the results sound promising. I can definitely hear positive differences when upsampling to 24 bit resolution against the direct 44.1KHz from the CD player data stream. There also appears to be a difference between using the input clock and reclocking with the Behringer. The output is using the Behringer internal DAC and analogue outputs. How it will sound with a good external DAC, only time will tell.
So far, so good.
 
That's good! Is there a defeat/bypass switch to allow for real-time A/B comparison?
 
Not as such. The nearest you can do with the Ultramatch is to go 44.1kHz digital in - sync lock to input signal - 44.1kHz out - then DAC.
I plan to also set up the CD analogue output to a spare input on the preamp to give a definite comparison.
Thus far the overall sound seems to be a bit more natural sounding with the upsampling, especially on the treble end.
I am trying to be as impartial as possible, trying to avoid audiophool terminology, but as I said before, it is looking promising.
I am just using the Behringer internal DAC at the moment, so this is probably just a first step in a chain of potential upgrades.

On a side note: I was looking on various sales/auction sites for CA Azur 550C CD players and I am not finding any at all - I really mean none - surprising. I don't know why.
 
Last edited:
Powertech said:
I can definitely hear positive differences when upsampling to 24 bit resolution against the direct 44.1KHz from the CD player data stream.
Click to expand...
Actually, upsampling refers to - well the sample rate - in this case 96 kHz. Padding the bit depth with empty spaces isn't going to improve anything.
 
I don't know precisely what the Ultramatch does to be honest, but it does do something.
Trying to be objective about this and bearing in mind that this machine is cheap, I set the system up as follows:-
I plugged the analogue outputs from the CD player into a spare preamp line input to provide a comparative baseline.
I fed the CD player optical and coax digital outputs into the Ultramatch and fed it's analogue outputs into the preamp CD input.
O.K. - so far so good. What am I making of this? Well, I already know that the Cambridge CD player is a very good machine in it's own right so I am already setting the bar fairly high. I tried the system out first with a Scissor Sisters CD which is well recorded and has a lot of low level bits and pieces going on in the background, and because it happened to be in the CD player at the time. I then played various tracks from albums that I know well.
Initially set to 44.1kKz in and out using the source clocking there was no noticeable difference between the optical and coax inputs and very little difference between the Ultramatch and the CD player itself, but there was something. When I switched to reclocking and upscaling the sound appeared to be slightly more resolving in treble detail and the bass sounded a little bit cleaner. I played with the output sample frequencies and went backwards and forwards many times and I can say that there is definitely a slight difference.
I know that this is no blind test, but I am trying to be as objective as possible (edging towards pessimistic in fact), but I think there is a very slight general improvement in presentation, maybe a little bit smoother especially in the treble region. The difference is not like night and day, it is very slight. Now, whether it's down to the upscaling, reclocking or the DAC I can't really say, but it definitely does do something.

At some stage I plan to get a separate DAC, so when that happens I can do some further comparisons.

On the subject of just adding zeros into the data stream, I am not sure if that is all it does. There may be some interpolation involved as well. Of course when you have a higher sample rate, it makes the job of the lowpass filtering a bit easier, so there may be something in that as well.
 
The (inherent) zero order hold on the DAC output introduces a sinc function. A sinc function produces aliasing.

Upsampling increases the sampling frequency, this moving the sinc aliasing up in frequency, allowing the analogue filter on the DAC output to be gentler, but still reduce the sinc alias products to a manageable level.

The anti-alias filter on the DAC output serves a different purpose to the anti-alias filter on the input of an ADC, but it is still an anti-alias filter.
 
You must log in or register to reply here.

Similar threads

J
Toshiba DVD players
2
Replies
23
Views
5K
joenielsen
J
Emotiva!
can someone help me understand resampling options?
Replies
3
Views
685
KOWHeigel
KOWHeigel
Bill Rogers
La Scala's, mc275 VI, tube pre-amp, Appo Sonica DAC. What CD player?
Replies
4
Views
364
Bill Rogers
Bill Rogers
virtualshawn
CD as transport only, how much does quality matter?
Replies
11
Views
1K
virtualshawn
virtualshawn
S
Receiver/CD player DAC confusion
Replies
18
Views
4K
Yamaki
Yamaki
Back
Top Bottom