15 USB/SPDIF converters shootout

[Blake]

Nice review Blake. It would have been interesting if you could have use the ST Fiber connection with Reflink/uLink, as stated in the TAS Review, where both units shine is with ST Fiber, RefLink was better than OR5 and uLink was closer to RefLink.

 

Yeah, I agree. I wish my DAC2.5 had an ST input, but no dice. You have to move up to the Bel Canto DAC3.5 to get the ST input. If someone wants to send a DAC3.5 my way......

[christopher3393]

The kind of subjective correspondence indicated by 4est's reply to Blake is, to me, important and interesting. It indicates a kind of "elective affinity" for lack of a better term, a mutual recognition that confirms an experience of listening that may elude quantification.

[Blake]

..also, I should have clarified one item- on high rez material (Patricia Barber's "Smash"), the REFLink & uLink using the iFi iUSB, with aes/ebu or s/pdif, bettered the Legato.

 

Finally, I think this was a bit of an unfair playing field for the REFLink, as I was not able to use the REFLink's optimal connection method (ST Fiber). I suspect that the REFLink using ST Fiber may have given the BADA-aes/ebu a run for its money, but who knows. If I were able to add more converters to the review, I would not be surprised if both Bel Canto units beat most of their competitors at their respective price points (the uLink trounced the Musical Fidelity V-Link 192). I posted a very glowing review of the uLink some time ago and my opinion of the uLink has not changed.

[miab]

The one converter I was hoping to hear a little more about is the Yellowtec PUC2. By some accounts it has bettered the BADA as a USB/AES interface but also acts as a A/D for digitizing vinyl/tape. It is quite difficult to get where I am so I was hoping to have a few comparisons to the above mentioned converters before jumping through hoops to buy it.

[Elberoth]

Blake - thanks for your informative review. Quite recently, forum member guydebord also came to similar conclusions, having compared BelCanto REFLink vs Berkeley vs Audio GD DI-V3s TCXO + DI-PSU in another thread:

 

Bel Canto REFlink in the house

 

Berkely Alpha USB is really a brilliant piece of digital audio engineering.

[guydebord]

From a pure playback standpoint what makes the plugin special. I tried Wave tube saturate plugin but it has weird popping/distortion sounds and the CPU drain is overkill.

 

I only use IK Multimedia ARC2 System plugin and so far I have no interest in any other. My CPU load playing 24/192 is 3% at most, a very efficient and well designed plugin that does what its supposed to do better than the Lyngdorf DPA-1 it replaced and the DEQX I have tested in several occasions.

[Swansong]

I only use IK Multimedia ARC2 System plugin and so far I have no interest in any other. My CPU load playing 24/192 is 3% at most, a very efficient and well designed plugin that does what its supposed to do better than the Lyngdorf DPA-1 it replaced and the DEQX I have tested in several occasions.

Thanks Guydebord,

I will look into it, how about the audio gd v3 comparison I asked about. My next purchase will be a USB interface so can you give me more info on your thoughts regarding Audio gd di v3 / psu combo? It sounds like a sleeper if ones not a brand/location snob. That describes me because I love rolling tubes out of my little tubedac-11. I was leaning toward the JK mk3 mod, any thoughts on the two?

[DanRubin]

Somewhat off topic, but I think this might be the place to get an answer. I've been evaluating the new Benchmark DAC2 UGC, using its USB input from Mac Mini running latest Audirvana. Works fine (quite nice, in fact). I decided to try its S/PDIF input by way of the SoTM USB-S/PDIF converter. Audirvana recognizes it and good sound is coming through, but I'm getting a fair bit of popping and crackling. And the Benchmark's word length indicator is switching occasionally between 16- and 24-bit, though that doesn't seem to be in sync with the popping and crackling. Has anyone encountered this issue with any of the converters? Know what going on? Suggestions?

 

Thanks.

Dan

[Swansong]

For what I can remember, I would categorize them in the following order, using the GD as reference:

 

100 - Audio GD DI-V3 TCXO + DI-PSU. As I said, to my ears and in my system (with a pair of ultra revealing pro-monitors), it had a slightly less clinical presence than AP, more liquid, soundstage was also slightly better. For those of you that think that spending more = better, if your translate Chinese R&D + costs of production into dollars, if this unit was made in the USA or any western city, it would easily cost above $1k. I have been more than 20 years in this hobby to understand that in hi-end, price is very relative and many times misleading. Audio GD is a genuine hi-fi lab with a very creative Engineer/Designer in charge and it just happens to be in China.

99 - AP1+PP. Great if you like a more clinical/technical sound without the usual harshness, I can understand why many would prefer this, especially at the first listen.

90 - BelCanto Ulink

85 - Stello U3

85 - M2Tech hiFace Evo (no PSU)

85 - Wavelenght WaveLink HS

80 - Musical Fidelity V192

 

 

I had the chance to audition Audio GD's top DAC as part of the summit group from HeadFi, at that time (2 years ago) I compared it with the Naim DAC, Lyndorf DPA-1, Transporter and MBL 1611F, all of them were better than the Audio GD. Saying this, Im not claiming that the Audio GD DI-V3 is the best there is, Im sure it can be bested, plus we all have different systems and hear differently. I already spoke to a Bel Canto and BADA dealer and Im arranging to audition both soon, can't wait! I have a good feeling about the REFLink.

 

Do you know what firmware your using? They seem to have #3,4,5 and each has different sound signature. How do you order this item?

[The Computer Audiophile]

Hi Guys - I just posted my review of the Computer Audiophile - Bel Canto Design uLink USB to S/PDIF Converter Review

[drez]

@Blake Thanks so much for this comparison, while it pains me that I bought the Reflink (upgrade from purepower Audiophilleo 2) when the BADA was only a little more, it is good to know that there are still further improvements to be had from my computer transport. i will try to keep both transports side by side so that I can at least comment further on the Reflink/BADA performance comparison. I might also try an AES cable, which I have avoided up to this point.

[gtener]

Has anyone seen or heard iFi Audio's iLink?

[squidly]

Has anyone seen or heard iFi Audio's iLink?

I have one on order and it should ship this week. I'm curious how it will fare against my V-Link 192. I'll definitely post my findings. I also have one of their Gemini cables on order, but it might be a few weeks as they're having manufacturing problems with the connectors.

[Blake]

@drez: I really liked the REFLink and I was actually set to buy it until I made my impulse purchase of the BADA. With redbook material, using s/pdif, the BADA, REFLink and Legato were all pretty darn close in performance and I would not be surprised if some listeners would rank the units differently than I did. You may want to keep the REFLink.

 

The additional $400 investment in the BADA may or may not be worth it to you. It is only with aes that the BADA really jumped ahead of the others for me.

 

If you consider the cost of a good aes cable, and add that to the $400 price difference between the REFLink and BADA, it does require some thought as to whether the BADA is a worthwhile upgrade to your REFLink (and to be clear, you might not actually find the BADA to be an upgrade). Hopefully you can demo the BADA before buying it because you might come to a different conclusion than I did.

 

I highly recommend all 4 of the converters in my review. It just boils down to price points and where you want to jump in.

[Clemmaster]

Wow, that's some fantastic shootout! Many thanks for that.

 

On a side note, the BADA looks like perfect engineering to me! All the (basic) clever ideas are implemented with - what appears to be - top notch parts.

 

I have two questions for you:

 

- You mentioned the HEX in the review but could not see any comment on its built-in USB module. Did you get a chance to compare it to some of the DDC? The reviews from John Darko and others at 6moons suggested that it performs right up with the best of the best. The direct (I²S) connection might play a big role here, so does the fact the USB module is a OEM Hiface with dedicated power supply, just like the JKSPDIF which 1) has great ranking in your shootout, 2) was said to be the favorite pairing with the Metrum Octave (again, Darko but also in the 6Moons review, IIRC). Audio-GD DAC also happen to have both the dedicated power supply and the I²S connection.

 

- Did you get a chance to hear the Concero as a USB-SPDIF converter? I own a Metrum Quad and used to use it with the Stello U3 with very good result (off a Squeezebox Touch + EDO). I then got the Concero and, technically, it is clearly better than the U3. It is also brighter. The Quad is an interesting beast: it clearly gets the sound signature of the transport feeding it (that's what one could call "transparency"!) but removes most of the digital nastiness (glare and harshness) and adds the R2R magic to it.

 

I could well lend the Concero but I'm in France so shipping will be expensive

[junker]

About that AES cable make sure that you either use something >= 1.5m due to internal reflections in the digital cable. This is well known in digital EE circles. You don't want the signal reflections (due to any impedance mismatch between the actual receiver and transmitter) to coincide with the center of the rise-time when the signal is at about 0V. Any variance here directly correlates to increased jitter.

 

I got the same answer using a back-of-the-envelope calculation assuming a 25 ns rise-time and a transmission speed of 0.7c.

 

When I spoke to Michael Ritter he referenced that they use the 1.5m Straightwire cable that is silver / teflon and can be found in that length for about $250.Steve Nugent from Empirical Audio confirmed that you can also go very, very short like maybe <=0.25m as well. I already talked with Nate Mansfield at Kimber about having custom cables made at that length.

 

After taxes I'm planning on going with the the Berkeley USB and DAC, and will be trying to just use a short AES jumper between the stacked units. Let's me get the best cable possible and the lowest price.

 

The additional $400 investment in the BADA may or may not be worth it to you. It is only with aes that the BADA really jumped ahead of the others for me.

 

If you consider the cost of a good aes cable, and add that to the $400 price difference

[Blake]

@junker: Yep, my Siltech aes/ebu cable is 1.5m. As recommended by Berkeley, my USB cable will also be 1.5m. I say "will be" because the iFi Gemini has not yet arrived, but I should receive it within a week or two. I am currently using a shorter usb cable. We had a brief discussion about the Straightwire cable earlier in this thread.

 

Hope you enjoy your BADA and Kimber. Kimber HQ is just a short drive north for me. I am absolutely loving the sound of my system with the BADA. It really makes a world of difference to my ears as compared to not having a separate converter.

[junker]

Nice! Not sure how I feel about the long USB cable, but guess the same rationale could apply. Guess I will start off with my 0.5m Kimber USB Ag, but might also try a 6" USB from Amazon to see if I can hear anything. You might want to chat with Nate @ Kimber about making a short TGDL. It would be 6 strand silver / teflon for a couple hundred... ;p

 

How would you characterize the difference in sound vs. straight in to the bel canto?

 

I've been thinking about trying the Berkeley USB into the NAD M51 AES since I can evaluate that for 60 days and save a few $, but most likely would just go for the BADA.

 

@junker: Yep, my Siltech aes/ebu cable is 1.5m. As recommended by Berkeley, my USB cable will also be 1.5m. I say "will be" because the iFi Gemini has not yet arrived, but I should receive it within a week or two. I am currently using a shorter usb cable.

 

Hope you enjoy your BADA and Kimber. Kimber HQ is just a short drive north for me. I am absolutely loving the sound of my system with the BADA. It really makes a world of difference to my ears as compared to not having a separate converter.

[reverendo]

@junker: Yep, my Siltech aes/ebu cable is 1.5m. As recommended by Berkeley, my USB cable will also be 1.5m. I say "will be" because the iFi Gemini has not yet arrived, but I should receive it within a week or two. I am currently using a shorter usb cable. We had a brief discussion about the Straightwire cable earlier in this thread.

 

Hope you enjoy your BADA and Kimber. Kimber HQ is just a short drive north for me. I am absolutely loving the sound of my system with the BADA. It really makes a world of difference to my ears as compared to not having a separate converter.

@Blake

why would BADA recommend 1.5 meters? I heard that the shorter the USB cables, the better. I'm aware of the controversy of coaxial spdif cables and Steve's opinions regarding the length (1.5m), but I hadn't seen this applied to USB cables. Is it the same rationale?

[Blake]

How would you characterize the difference in sound vs. straight in to the bel canto?

 

It may be that I am more sensitive than most people to the sonic differences associated with a stand alone converter, or perhaps it is the case I am just more enamored with the particular sonic impact of a separate converter (I think each listener will tend to focus and place more importance on various aspects of sound than others, so what is big to me, may not be such a big deal to others), but the difference is huge and transformational. Now that I have heard the difference, I could never go back to a direct connection between my computer and DAC.

 

The sound with the stand alone converter in the chain is simply natural and realistic. It does not sound at all like digital. With the converter in place, there is no sense of digital tension/glare/hardness. When I run direct, particularly now that I have heard what a stand alone converter does to my system, the sound is very much digital. Before inserting a stand alone converter, I was very close to pulling the trigger on a turntable because I was not 100% satisfied with the digital sound aspects of my system. That is no longer the case.

[Blake]

@Blake

why would BADA recommend 1.5 meters? I heard that the shorter the USB cables, the better. I'm aware of the controversy of coaxial spdif cables and Steve's opinions regarding the length (1.5m), but I hadn't seen this applied to USB cables. Is it the same rationale?

 

My BADA dealer explained it to me, but I cannot recall everything he said so I am afraid I may not accurately convey BADA's thinking on the USB cable length. My dealer did say that while BADA recommends 1.5m length for each cable, having a 1.5m usb cable length is even more important than having a 1.5m s/pdif or aes cable.

 

Here is a link to some discussion regarding cables for the BADA unit.

 

 

http://www.computeraudiophile.com/f8-general-forum/specific-question-berkeley-audio-design-associates-usb-converter-users-15433/

 

Honestly, I just decided to follow BADA's direction as I defer to their obvious expertise. But other respected designers have a different view (AR-T for example). Perhaps the usb cable length issue is unique to this BADA converter.

[junker]

This is a pretty good explaination of the effect with regards to SPDIF / AES:

 

spdif

 

Not sure if this applies to USB. Does anyone know the rise-time of the USB signal? If it is 25ns, then it could be the same.

 

I can use that the the signal speed through copper or silver and a PE dielectric to calculate what the round-trip cable length would be an issue.

 

My BADA dealer explained it to me, but I cannot recall everything he said so I am afraid I may not accurately convey BADA's thinking on the USB cable length. My dealer did say that while BADA recommends 1.5m length for each cable, having a 1.5m usb cable length is even more important than having a 1.5m s/pdif or aes cable.

 

Here is a link to some discussion regarding cables for the BADA unit.

 

 

http://www.computeraudiophile.com/f8-general-forum/specific-question-berkeley-audio-design-associates-usb-converter-users-15433/

 

Honestly, I just decided to follow BADA's direction as I defer to their obvious expertise. But other respected designers have a different view (AR-T for example). Perhaps the usb cable length issue is unique to this BADA converter.

[junker]

With spdif you basically want to delay the reflection past 1/2 the 25 nS - this is about when the receiver flags the voltage as a '1'. You either want the round-trip to take longer than 12.5nS, or WAY earlier beucase with a short cable there can be a second, third, forth, and even a fifth reflection of diminishing amplitude. If you go short I would keep it in the 6" to 10" range".

 

..........0V threshold...../-----1------| receiver

transmitter |____0___/

 

This is a pretty good explaination of the effect with regards to SPDIF / AES:

 

spdif

 

Not sure if this applies to USB. Does anyone know the rise-time of the USB signal? If it is 25ns, then it could be the same.

 

I can use that the the signal speed through copper or silver and a PE dielectric to calculate what the round-trip cable length would be an issue.

[reverendo]

what I do know is that there is quite a controversy regarding this matter... it was kind of battled out between Steve Nugent and Dan Lavry. here's Lavry's take on it: Lavry Engineering • View topic - Replacing the SPDIF RCA jack with BNC

I just know that I'm not in a condition do understand very well, since this is not my field.

Best regards

[junker]

Yeah, I'll stay out of that one but I will say this... I don't want to spend all this money of a digital front end and blow it by buying an IC that is 0.5m off possibly resulting in "jitter" and working against the excellent clocking and isolation in their products. People keep realizing that little things matter much more than we thought early on. I would have never thought that jitter needed to be down in the pico or femtosecond range. Wow...

 

I'm fine buying a 1.5m or longer cable...or a very short jumper cable. I'll trust Berkeley and Wavelength Audio on this one. They certainly have demonstrated excellent engineering skill, and save me the trouble of having to become a digital EE major... I'm just a chemist haha.