Dramatically improved performance with DirectStream DAC

Ever since I got my DAC I’d been using either the network bridge, what I started with, then the USB input for several years via the USB output on my Roon Rock Nucleus and an AQ Jitterbug inserted. I was never very happy with this sound and I was mistakenly blaming other system components. Only recently did I become aware of how much I didn’t know about how to optimize the performance. I was reminded about the I2S input and how it is the preferred input for sound quality. Some research guided me to Small Green Computer and the Sonore product line. I read about separating the noisy network side from the “good” side for streaming via optical Ethernet. Coincidentally, Paul posted on this topic recently.

After a little stumbling with just what to assemble for my solution, I ended up with a new Trendnet switch with SFP converter, fiber to an opticalRendu, fed by a Sonore LPS, USB to ultraDigital, and hdmi to I2S. Immediately I noted a big improvement in the bass, and much cleaner and detailed presentation. After some time I became aware of a feeling of relaxation. After running this setup for a day to break it in,
I ran a very interesting and revealing test. I hooked my direct USB connection from my NUC with AQ Jitterbug inserted back into the USB input of my DAC. Then I grouped the DAC and Sonore inputs in Roon. Then I switched back and forth on the PS Audio remote while playing the pink noise track on the Stereophile test CD. Clearly I could hear a layer of upper midrange brightness or glare on USB that disappeared on Sonore. Additionally, as I had mentioned before, the bass was substantially, obviously fuller via Sonore (and I2S). It’s such a black and white test that I recommend it highly for any doubters. That midrange brightness had me frustrated for years! As I kept improving components, it still wasn’t right. Thankfully I have an excellent analog rig.

Now some may question how much of this improvement is due to the Sonore chain and how much is due to the I2S input. That’s a good question. Obviously I can hook SPDIF out of the UD and give it a whirl, but I’m in no hurry. Maybe the I2S, an underlying architecture that has been around since day one, helps explain why some still prefer playing Cds over streaming over the network with no optical conversion to remove noise before it gets to the I2S internals. But before I ended up with what I have, I started with optical Ethernet from that Trendnet switch to an FMC, to my Roon NUC, and usb to DAC. This yielded I think a blacker background and overall improvement, but still lacking in bass. Wanting to be able to stream DSD 256, I quickly moved up to the UD with SGC LPS. Then when the optical Rendu and Sonore LPS arrived yesterday, I finished the setup, leaving the NUC on the network side.


Those discoveries just make you want to smile and listen to the music.


You would really like the Rendu Signature SE. I have tried many other endpoints, including the opticalRendu. The Signature is even better.


I’m using the following setup:
SGC SonicTransporter i5 Roon Server
UpTone Etheregen switch w/ FARAD3 PSU, connected via SONORE Optical Module to the internet
SoTM sms200ULTRA NEO streamer w/ FARAD3 PSU and w/ a PinkFaun ethernet isolator at its input
Purist Ultimate USB cable into the DS DAC

I wonder whether I should invest further in improving the USB connection considering that:

  1. the SoTM streamer USB output is NOT galvanized, but they claim they use some proprietary low-noise technology.
  2. the Purist USB cable is one of the best out there

My options are:

  1. Move to i2s with a SONORE UltraDigital. But then I need to use another HDMI cable and also this UD draws current from the streamer USB so may affect its performance ? which HDMI cable is recommended ?
  2. Improve the USB with UpTone ISORegen. This sticks out of the DAC back and is not very practical. it also requires a separate AC cable.
  3. Some other USB improvement like SoTM TxUSB, needing a 2nd USB cable, PSU etc…
  4. Keep it the way it is…

BTW, SONORE claim that if you use their Signature Rendu, then its USB output is so good that using the UD with i2s is redundant (they even offered me a trial with refund if I but the Rendu). So apparently, to their claim, a good USB implementation is all it takes…

Any insights ?

When I mentioned that the bass was improved with I2S, I mean it was as if I cranked the bass knob up, if I had one. The bass was so much punchier and fuller that I wondered if I was going to have to redo the equalization settings on my Vandersteen 5A Carbons. I’m even now wondering if my USB input is somehow broken on the DAC. There’s just no bass in comparison. Nearly as dramatic was the removal of a clear layer of upper midrange glare. Maybe @vkennedy61 can chime on how much better the Signature Rendu would be, but I have my doubts that it would make that much of a difference. Maybe @tedsmith can chime in? Why is the bass so anemic through USB?

@pcullen1, thx for your input.
Could you compare between USB with the Optical Rendu vs. i2S with Optical Rendu and UltraDigital ?

In general the bass isn’t different between various inputs. One of my normal tests with each release is to hook the same source to all inputs and randomly and quickly switch back and forth between inputs. Usually you can’t tell when the input selected changes at all. If there’s an audible difference between inputs then usually there’s a difference in the conducted electrical noise from the source (or more rarely the radiated noise.) If your inputs are all coming from the same source unit then often the differences between inputs is minimized because they are sharing the same goundloops, etc. When the sources that are being compared are different physical devices there’s a lot more room for differences in groundloops, radiated noise, etc. Especially with USB, the source is often a computer and with I2S it’s rarely a computer. Computers are very noisy unless extreme care is taken when building them.


@orionn @tedsmith opticalRendu to USB is much closer to opticalRendu to ultraDigital and I2S, with the latter giving a slightly better sense of air, delineation and depth with music. The pink noise test in this case is also similar, but because of how close it was, I had to resort to repeat plays of Vocalise by Rachmaninov to reach this conclusion. It’s the direct connection from RUN NUC to USB which is bass shy and glaring and easily revealed by the pink noise test, I suppose due at least in part to conducted noise as suggested. @orionn as far as I2S cable, I’m using the one supplied with the ultraDigital, which is substantial and has barrel noise filters on either end. I’m not ready to go down the cable rabbit hike here just yet.

thx @pcullen1
I think you are missing out reg the cables
they make a dramatic difference, both usb and i2s

Maybe, but there’s no way they explain the big difference I was hearing between the two setups.