Heads-up on a DAC not mentioned much here

[barrows]

Interesting comparison. I would have loved to hear about the comparison of both DACs being fed from the same source: SPDIF from the Orpheus. Orpheus/Audio GD seems like a pretty expensive set up.

 

[aljordan]

Orpheus/Audio GD seems like a pretty expensive set up.

 

I do not normally feed the Audio-GD with the Orpheus, but we did in this case because it was the easiest we had available. I have been running it via the transformer isolated S/PDIF output of a cheap M-Audio Delta 410 sound card with very good results, but I didn't want to yank it out of a computer to bring it over to my friend's house. I would think that anyone who would run the Reference 7 would probably do so with something like a decent USB to S/PDIF converter.

 

Alan

 

[bixby]

What are some very good sounding usb to spdif converters? No need for hi res for me, more interested in very good implementation at a reasonable price point

[gojita]

Alan,

 

Where did you learn about the DAC chip in the Cantata?

 

The Cantata user manual makes it sound like the non-usb inputs will do 24/192. I asked Audio-GD about their non-usb inputs and they claim only 24/96 due to the limitation of the PCM1704UK used in the Ref 7. Wouldn't this indicate that the Cantata and Ref 7 use different DACs?

 

I'm curious as I've been considering trying out one of these.

 

Thanks,

 

Jason

 

 

[barrows]

pretty sure the BB 1704 can accept data rates much higher than 24/96. Unlike contemporary DAC chips, the 1704 does not have on chip oversampling/filtering, and is usually paired with an external oversampling/filter. Of course, one might use an external filter that is limited in data rate.

Regarding the comparison between the Audio GD and Cantata, it does not seem like an apples to apples comparison; both should have used the same transport and input-in this case Orpheus-SPDIF. The one review I have read of the Cantata found better sonics through the SPDIF input, while this might be in error, it still seems like the comparison would be more valid through the same pathway using the same transport.

 

[XP9433]

Wouldn't the Audio-gd NFB7EN DAC using the Sabre32 ES9018 DAC chip and supporting up to 32Bit/192KHz be of more interest?

 

 

 

Frank

 

 

[Audio_ELF]

The PCM1704 chip is advertised as 24/96 because they expect external oversampling. In actual fact it will accept up to (IIRC) 768k. It all depends how the designer / manufacturer utilises the chip.

 

Eloise

 

[aljordan]

Hi Barrows,

 

A few weeks ago I compared the Cantata to the Orpheus. I was using the Orpheus to feed the Cantata S/PDIF, and I was (correctly in my opinion) scolded for not using the asynchronous USB input of the Cantata by a Cantata owner, presuming USB would be the best manner to drive the Cantata from a PC.

 

I think that if one is comparing DACs, it is reasonable to use the best manner of feeding each specific DAC in the comparison, and not limit the performance of each DAC by using the same source.

 

I can compare the two DACs using the Orpheus and feeding each DAC's S/PDIF input on Saturday if that is of interest. I am wondering though if people would feed the Cantata from a PC via S/PDIF if it includes a technically superior asynchronous USB input in the unit that slaves the PC to the local DAC clock.

 

Alan

 

[aljordan]

Where did you learn about the DAC chip in the Cantata?

 

From the following review:

 

http://www.techradar.com/reviews/audio-visual/hi-fi-and-audio/audio-systems/resolution-audio-cantata-704024/review

 

Alan

 

[aljordan]

Audio_ELF wrote: The PCM1704 chip is advertised as 24/96 because they expect external oversampling. In actual fact it will accept up to (IIRC) 768k. It all depends how the designer / manufacturer utilizes the chip.

 

This is the default manner in which the Reference 7 feeds the PCM1704, although the oversampling rate is adjustable or can be bypassed altogether. I did try running it as an NOS DAC and did not like the results.

 

Alan

 

 

 

[ted_b]

to be listened to via its own best sweetspot (i.e best engineered input, etc). If a ball club is trying to find a pitcher to complete its rotation, it will not ask Nolan Ryan (fast ball pitcher) and Greg Maddux (king of location, breaking ball pitcher) to both throw the same pitches. One or both will not show his strengths, in that the object is not to throw the fastest per se, it's to get the batter out.

 

[barrows]

Thanks for the details.

 

"I think that if one is comparing DACs, it is reasonable to use the best manner of feeding each specific DAC in the comparison, and not limit the performance of each DAC by using the same source."

 

Normally I agree that the approach you state above should be tested for comparisons. But first I think the DACs in question should be compared through the exact same source to get a good baseline of differences.

As we know, introducing two different interfaces brings in a huge amount of variables, on the one hand we have:

1. Computer setup, Orpheus setup, SPDIF cable from Orpheus to Audio GD, SPDIF receiver quality in Audio GD

2. Computer setup, USB cable, USB receiver quality in Cantata

It is difficult/impossible to assure that all these options are optimised. As an example, different player software might tilt the balance in favor of one DAC or the other.

If one tests: computer setup, SPDIF cable, Audio GD/Cantata-the only variables are the DACs-this lets us realize the true difference between them, and then we test the individual interfaces.

BTW, the review linked above states that they got better sound through SPDIF than USB with the Cantata.

If you do have the opportunity to compare these DACs via the same source, I would love to hear the results.

 

[Audio_ELF]

@Barrows...

 

I think it depends on what your aim is with a review how you should demo the kit (i.e. internal USB vs external converter to SPDIF).

 

If you're working out what is the best (for you) within your budget, then you should be trying to get the best quality possible (IMO). However a "commercial" review the comparisons should be made with as much of the same kit as possible so both using an external converter if required for one.

 

Eloise

 

[barrows]

Yeah, I mostly agree. But if one is married to a certain interface, then one is going to tweak that interface to get the best out of it. One might try different USB/SPDIF cables, or different playback software. Only through many trials will one get the best out of a given interface-so this makes comparisons with different interfaces difficult, as one cannot assume the different interfaces have both been optimised for best performance with the respective DACs.

In this case a $5K Prism Orpheus is used as a SPDIF interface-I do not expect many people will purchase an Orpheus to use as an SPDIF source for the Audio GD: the results using something like a Hiface might be very different.

 

[fahadm]

Desperately seeking your DAC shoot-out results ;-)

 

Thank you

 

Fahad

 

[Audio_ELF]

I mostly agree with you too...

 

I guess my thinking is if you were (for example) to compare DACs you should compare them in real world situations. The Audio GD + HiFace Evo vs the Cantata would be valid where as using the Orpheus as the interface for the Audio GD vs the Cantata invalidates (to a large extent) the comparison.

 

Eloise

 

[aljordan]

Barrows wrote: BTW, the review linked above states that they got better sound through SPDIF than USB with the Cantata.

If you do have the opportunity to compare these DACs via the same source, I would love to hear the results.

 

This past weekend I did compare the Cantata driven via SPDIF from the Orpheus to the Audio-GD Reference 7 driven the same way. I did not compare the Cantata's coax SPDIF input to its asynchronous USB input. The rest of the system was Pass XA 100.5 amps at some point and BAT VK-600 SE at other points, BAT Rex preamp, and Wilson Sasha speakers.

 

Instead of mentioning which one I thought sounded best, I'll just mention how they differ. The Cantata had a more forward presentation with less low frequency response and an overall brighter sound. The Cantata did not resolve low level details as well as the Reference 7.

 

Comparing the Orpheus to the Reference 7: The Orpheus is not overall brighter than the Ref 7 as the Cantata is, but the Ref 7 has slightly more midrange body than the Orpheus. The Ref 7 also slightly out-resolves the Orpheus, but I can hear the resolution difference more in the high frequencies rather than the low frequencies.

 

I realize that nobody would normally drive the Ref 7 with an Orpheus, but that's what we had on hand. I did order a Halide Bridge to see how that fairs but I imagine I won't receive it for a week or so.

 

Having said all this, I will also mention that in my opinion all three units sound very good, particularly in my own system where I use a soft digital room correction filter to help with room anomalies. The owner of the system with the Sasha's thought the Ref 7 sounded best, and was actually somewhat unsatisfied when he had to return to the Cantata; but I think that has much to do with overall system balance issues where one DAC just mated better than the other.

 

Alan

 

[barrows]

Alan. That is a lot of good information.

 

[all300b]

Thanks for that great comparison. I'm starting to think that a Wavelength Wavelink - good SPDIF cable- Audio G-D ref 7 may be a true high end solution at under 3K. The internals of the Audio G-D look really impressive.

 

[aljordan]

all300b wrote: I'm starting to think that a Wavelength Wavelink - good SPDIF cable- Audio G-D ref 7 may be a true high end solution at under 3K.

 

I would like to try the Wavelink as well. I have in the past owned a Cosecant V3 which was a nice sounding DAC. I didn't order the Wavelink because the Reference 7 only does up to 96 kHz, but some the design of the Wavelink is intriguing. I've read good reports of the Halide unit, but if a Lynx card outperforms it (like I've read somewhere) I'll be disappointed, as I've owned a Lynx card for a long time and do not think it is the best performing solution.

 

Alan

 

[manisandher]

"... I've owned a Lynx card for a long time and do not think it is the best performing solution."

 

Alan, you might want to try an RME AES-32. I have the PCI version of this card and have done a number of listening tests with it. Feeding my ADC/DAC with an analog source, I can easily switch between an internal 'ADC/DAC' chain vs an external 'ADC/RME/DAC' chain. In double-blind listening tests, I and several friends could NOT identify when the RME was in the chain. Conclusion: the RME AES-32 and the cheap AES cables that I use are transparent (or at least transparent enough for my setup).

 

BUT...

 

There are a couple of things that I had to do to get here:

1) set the AES voltage to 'consumer' (~2.1V)

2) set the RME buffer size to 1024

 

What I've learned from people like JR_Audio and Keith Johnson (via Dave Peck) these last 6 months or so have been a revelation. There's more to AES than meets the eye...

 

Mani.

 

[aljordan]

Hi Mani,

 

I'm interested in knowing what you have learned about AES. Are there links you can refer me to?

 

Do you think AES/EBU is a better manner to feed a DAC than BNC if the distances are short?

 

Edit: I see that you are slaving the AES 32 to your DAC. I have done this with the Orpheus and the Lynx with the Orpheus set as master and was able to get very good performance out of the Lynx card in this manner. However, since the Orpheus as an async firewire solution with the ability to slave the computer to the DAC, it is unnecessary.

 

The Audio-GD Reference 7 does not have a clock output, so I can't slave a transport to it. The DSP-1 unit in the Ref 7 has the task of reclocking with a high quality local clock. Hence I imagine the AES card would not be a good solution if it could not be slaved to the DAC. How much difference do you hear when the AES 32 is acting as the master compared to acting as a slave? There was quite a difference when I did the same with the Lynx card and the Orpheus.

 

Thanks,

Alan

 

 

[manisandher]

Alan,

 

I'm certainly no expert - I consider Pat (ar-t) and Juergen (JR_Audio) to be the real experts here wrt this subject.

 

"I'm interested in knowing what you have learned about AES. Are there links you can refer me to?"

 

Have you read the following threads?

 

http://www.computeraudiophile.com/content/classic-digital-inputs-ouputs-Aes-vs-Bnc-vs-Rca-vs-Toslink-Does-it-matter

 

http://www.computeraudiophile.com/content/How-can-influence-SQ

 

I have no experience using BNC for the signal (as opposed to wordclock), simply because I've never had a DAC with a BNC signal input. But if Pat says that BNC can provide a near-perfect transmission medium, then I believe him.

 

But what I do have is a reasonably good (to my ears) analog source and a good ADC/DAC. So it's easy to compare digital playback to the original analog source. And what I've found in double-blind listening tests is that AES/EBU can sound totally transparent. But, it seems to be quite a quite 'delicate' thing to get right. I mean, no matter what I try, I cannot get my Weiss AFI1 to sound transparent. But I can get my RME AES-32 to sound transparent, under the following conditions:

 

- AES/EBU interface slaved to DAC wordclock via BNC cable (75 Ohm termination at interface end)

- AES/EBU cable >6' in length

- no additional 'short' (i.e.

[barrows]

expect Mani's experience here to be correct. What I mean that is to get a computer sound card to really work as well as possible slaving to the DAC clock would be the best way to go. Clock circuitry is very dependent on having a very clean power supply, and it is unlikely that the clock generated on an internal sound card is going to have a very clean supply. Sourcing the clock from the DAC means one is getting a more accurate clock to the sound card (assuming a good clock in the DAC), and the incoming SPDIF/AES stream will be in quite close sync (because it was generated by the same clock, subject to slight variations due to transmission losses) with the clock at the DAC. Because of the close sync, the PLL circuit in the DAC will not have to "move" much, and hence the jitter should be very low.

For the Audio GD, as it has no clock output, the best solution would likely be one of the external converters:

Weiss INT202, Wavelength Wavelink, Hiface Evo, Audiophilleo, etc. As these devices (when properly engineered) have a better chance of generating a really good clock, as they will not be subject to the power supply (and internal RFI) issues that degrade internal computer soundcards. As you have already achieved good results with the Orpheus (Firewire-SPDIF) acting as an external converter, perhaps the Weiss INT202 (Firewire-SPDIF) would work very well with the Audio GD.

 

[The Computer Audiophile]

Hi Mani - I really wish you were coming to RMAF next month. Maier Shadi's room will have a Model II as the DAC and sending clock to a Merging Mykerinos card in a silent server. Tim Marutani is assisting Maier on the room. It should be a good one.