For this review, I received the silver option because Vincent and I were agreed that it would photograph better against a black background. The product arrived in two boxes where the main gear in a smaller box was put inside a bigger carton. What arrived with it was Vincent's 0.25m USB filter, 2.0m Ethernet cable and plastic remote. The latter is standard. To get the other additions, one pays €360 and €390 respectively. The d1-integral-headphone has a very nice OLED display for all the necessary intel like current digital input, file density and volume setting between -70 to 0dB. One can also invert absolute phase, set ground connected/lifted or apply FIR compensation. Vincent likes non-oversampling DACs but not their 3dB treble loss at 20kHz. Against that he developed a compensation filter for just the high frequencies. One thing to be stressed from the go is that Vincent's product worked flawlessly. I had no hiccups but very good response to RC commands and no noticeable delay when switching between various key functions. From its UI perspective, this deck was very solid.

The rear is generously appointed with variable XLR and RCA outputs on the left, both at 3Vrms. These XLR make the d1-integral-headphone more versatile but are not true balanced. Attenuation is handled by FPGA at 69-bit resolution. It'd be nice if the RCA/XLR socketry were marked 'left' and 'right'. Next come the headphone socket (3Vrms, 1.2W peak power @ 15Ω) and four digital inputs of coaxial S/PDIF, AES/EBU, TOSLINK and USB. A bit to the right sits the AC socket (the power supply has a regular IEC on its back, a diode and an on/off switch) and a server module above. This consists of two type A USB ports, eSATA, RJ45, HDMI and a microSD slot. Total power consumption is 22 watts.

Additional commentary should go toward how to configure the d1-integral-headphone and tap the Roon magic. I won't do this better than Vincent so his on-site tutorial is the destination. Let's get at the innards. To some degree, all TotalDAC decks are similar. Not only do they share a chassis, they also get the same conversion method. Vincent eschews conventional chips even though a number of vintage R2R solutions remain available. Our French engineer however decided to bake his own from scratch. It took him many attempts, iterations and years to pull this off; long enough to be convinced that this was the best possible approach. The result is no small converter board but rather a large PCB that here consists of 96 x 0.01% VAR bulk metal foil resistors by Vishay and an FPGA for glue logic. That's a beautiful sight though it's easy to spot also that only 1/3rd of all resistor spots are occupied - nowhere near the declared 100 units. In the pricier models, these empty rows are filled with the same resistors. As for our count, one sees four rows of 12 resistors each. That nets 48 Vishays, times two though because the other half hides on the board's underside. Mystery solved.

The lower board is home to discrete voltage regulators, filter capacitors and the output stage. The latter consists of two discrete class A modules (voltage gain and current buffering) developed by Vincent and also used in his d1-six DAC. Not much is known about it. He only explains that his headphone output taps into some of the same circuit. If one still feels salty because of the 96 not 100 Vishays, the last four hover around the transformers. A CuBox platform handles the server part and the USB transceiver is by XMOS. Surprisingly, iFi Audio's USB drivers worked like a charm. The core innards mount to a solid copper stabilizer and heat sink. Internal damping material on the upper chassis parallels lots of visible RF shielding. To fight jitter, the FPGA's FIFO memory runs a 10ms audio buffer whilst a local oscillator handles the clocking. To exploit that properly, Vincent suggested to use his USB filter when the d1-integral-headphone was fed via RJ45.

Sound. To assess performance, I needed to do several things. As most people into audio at this calibre know, this machine is first off a D/A converter. That had to be assessed above all else. To hone in on its sonic signature, I used my main system where Vincent's machine went up against LampizatOr's balanced Golden Gate. The price difference here was rather significant yet thanks to such comparisons as a reality check, the new product's value could be established. I also had to control the quality of the TotalDAC's inputs. This unit was designed to work in the Roon environment via RJ45. Yet I had to also test at least one other scenario because not everyone has opportunity to connect a NAS-fed router. The obvious choice was the USB input. Finally I wanted to test the quality of the French digital volume control. To confirm initial observations in the big setup of Trilogy 925 and Boenicke Audio W8, I also used a FirstWatt F7 and Xavian's Perla monitors with the same sources. After this it was time to move the tester to a smaller rack. My desktop rig consists of W5 monitors by Sven Boenicke, the FirstWatt F7 and a Xonar Essence III DAC. Vincent's replaced the last one. Because the Asus is both DAC and headphone amplifier, it was a perfect opportunity to check what the French would do with headphones. Well, that was the grand plan.

Then just in the nick of time, something special arrived by way of the Polish Divaldi AMP 01. This cute little thing is a discrete Class A headphone amp with phono stage. Its topology promised something special, hence my choice to include it. HifiMan HE-1000 and Meze 99 Classics headphones helped me execute this part. That's the overview. Now onto how things went. In the so-called big rig, it was rather obvious that sonically, I had on hand two completely different performers. Certain from the very beginning was that in many aspects, this was a battle between different philosophies and goals. From a purely subjective perspective, not every day do I enjoy such an educational comparison. Often it's a case of better versus worse. This time around, I had two vastly different flavours. Here it's worthwhile knowing that since the Golden Gate's DSD engine accepts only USB, I connected the TotalDAC to a laptop in the same way. Afterwards, the tester moved into the environment suggested by Vincent, namely router + NAS + Roon + iPad-based remote. His heavily filtered Ethernet and USB cables made appearances, too.

To keep it short, things were best without PC-based transport. The sound was more resolving and a bit denser without any negative side effects. In comparison to USB, the outcome was more cheerful and detailed. It might seem unfair to run the Golden Gate via laptop and the d1-integral-headphone as described. The goal was simply to squeeze every bit of performance from these two and there was no other way to do it. If the Golden Gate could have been fed from the same signal, it would have been extremely convenient. But that would have bypassed its DSD engine and with that, the deck's entire reason for being. Moving on, many times before I've pointed out that LampizatOr DACs tend to sound rather direct. That's when the very first row of instruments appears quite close to the listener and clarity is very high.