What defines the quality of a DAC?

The difference in quality between the cheapest DAC and the most perfect DAC is enormous, provided you use a high quality playback system. But what makes the most perfect DAC the most perfect? A digital to analog converter has a job that is easily described: it has to render digital samples in corresponding voltages at precise intervals. But that’s perhaps somewhat abstract and some explanation might be in order. Let’s imagine a short piece of analog music signal by chance being a straight line like this. On digitising, amplitude samples - tech speak for measuring the voltage - are taken at the given sampling frequency, like 44.100 times for CD quality.

00:54 - These measurements are stored in a table and sent to a transport or storage medium like a hard disk, cd or thumb drive. On playback the table is read and then rendered in discrete voltages. This would have resulted in a staircase pattern if not for the reconstruction filter that ‘slows down’ the the signal so the original straight line is regained - if all went well. If the interval at which the voltages are plotted is not very constant, there will be amplitude errors as can be clearly seen when a white line is plotted behind the resulting red line. If the jitter is caused by the interference of a single frequency, this creates side bands at the modulation frequency of the clock and lead to what is commonly called ‘the digital sound’.

01:43 - In layman’s terms, it will cause tones in the signal that originally where not there. But there are several signals that can cause jitter, including noise. The jitter then is less easy to identify but still ruining the soul quality. Sound is variation in sound pressure over time. The sound pressure is converted during recording to voltages and then coded in digital information while the time factor is stored in the sampling frequency.

02:16 - The time factor can be distorted by jitter but the amplitude of the voltages can be off too, of course. We call this poor linearity. If the digital signal says the output voltage of a given sample should by 0.123 volts and the real output voltage is 0.156 volts, the converter is not linear. We measure this by applying digital signal levels to a DAC and measuring the output voltage. This is such a measurement. You see that -30 dB digital input generates a voltage that is 30 dB below the maximum output when you look at the turquoise line.

02:56 - The same goes for -40 dB, -50 dB and so on. The green line shows the deviation from full linearity and you can see that this DAC is very linear down to -100 dB with only minor deviations down to -120 dB. -120 dB equals a resolution of 20 bits, which is about the best I have ever measured, regardless the price of the DAC. A real world 24 bit DAC does not exists. DAC’s that claim to be 24 bits will except 24 bit long digital signals but the lower four bits will be irrelevant unless the digital filtering uses it for dithering and rounding of. The audibility of a DAC’s non-linear behaviour is questionable, at least in my experience.

03:50 - In the mid eighties we measured a lot of cd-players, many of which didn’t even managed to perform as 16 bit converters. But it was always jitter that was the main factor - although we didn’t know back then it was jitter. We learned that some years later. In the animation on how digital audio works, we have seen that without the reconstruction filter the output of a DAC would be a staircase. I must admit that a staircase as in the animation in most unlikely since that would need analog electronics with unlimited bandwidth. It rather is a staircase like signal. The reconstruction filter is the solution for this.

04:35 - It slows down the signal to a degree that it is just fast enough to connect two samples in a natural way. But there is no such thing as free lunch. Filters have not only influence on the amplitude - the wanted roll off - but also on time behaviour. There have been many attempts to make the perfect reconstruction filter but very good filters often are very expensive. Most often the digital signal is upsampled and filtered during upsampling, then sent to the digital to analog converter circuit. Since the upsampling is done to a high sampling rate, the analog filtering can be done at a higher filter frequency of half that sampling rate.

05:17 - This causes the artefacts to happen at higher frequencies too, mostly outside the audio band. Digital audio works with square wave like signals that require steep voltage changes and thus instant current. This is where switch mode power supply excels. But these use switching at very high frequencies and therefore generate a lot of noise on the power lines and ground plane. They have the advantage of being very efficient - very little energy gets converted to useless heath. Basic designs are also very cost effective. But if you need to have very low noise figures, as you would in DAC’s, you need to filter the output of the power supply and do that without slowing it down as simple filtering will do. So there are four main factors that influence the sound quality of a DAC: jitter, linearity, reconstruction filter and power supply. It’s safe to conclude that low jitter is the most important factor to the sound quality of a DAC. But a poor power supply will cause poorer jitter behaviour and an import signal with a lot of jitter might ruin the audio quality as well.

06:36 - In my setup 1 I have used the same DAC for some years now, the Mytek Brooklyn. Initially the first generation and currently the Brooklyn Bridge that is almost identical but has a network bridge integrated. Sound wise there is little to no difference between the DAC function of both. I used them with the sBooster BOTW P&P ECO MK II, 12-13 volts, so bypassing the internal switch mode power supply. That can be easily done since there is an external power connector on the rear.

07:13 - The initial source was an Apple Mac Mini running Roon and connected over a USB Audio Class 2 connection. That was followed up by the Sonore MicroRendu network bridge, which gave a clear sound improvement. Resolution was better, there was less stress and lows had more texture. When the MicroRendu had to be returned, it was replaced by an SOtM sMS-200 network bridge that gave just a very small improvement over the MicroRendu. A real big step was taken when the sMS-200 was replaced by the sMS-200 Ultra. It was like having a new stereo all together. Remember that we are talking digital signals only and I am sure there are people that can’t imagine the impact this change had. Then I heard of the Syntaxx power supply, a hard wired, hand built 6.2 amps linear power supply, specially developed for the Mytek Brooklyn. It appeared to be an 8 amps designed, fused at 6.2 amps and did wonders for the Brooklyn.

08:18 - The switch to the ICON version again gave some improvement. And at € 1200 it should. The latest change was replacing the sMS-200 Ultra network bridge by the Auralic Aries G2 digital transport, as Auralic calls its network player. I use it mainly as network bridge for Roon. Again the sound quality went up a notch or two. Or four, depending on the size of your notches. I know, they are significant jumps, monetary wise. And, of course, being a company I usually pay a different price, no VAT - the European sales tax - and the remaining amount is tax deductible. But the sound improvement between the sMS-200 Ultra and the Aries G2 again is impressive. I can’t judge your financial position and other personal factors that define how you value money and how sound quality. But I can judge the sound quality difference, at least in my perception.

09:21 - We have seen in the beginning of this video that in a DAC there are four factors important for the sound quality: jitter, linearity, power supply and reconstruction filter. I have stated that linearity only will have a minor effect on the sound quality. At least within the real world variations. The reconstruction filter of the Mytek leans on the MQA philosophy, also when non-MQA material is played. Other reconstruction filters can be chosen when MQA is switched off, but I like the MQA based filter. And that was a given throughout the upgrade process. Only two factors did change and had a profound effect on the sound quality: the digital input signal and the power supply. The technical reasons behind this I will cover in a future video. I have been asked several times if it wasn’t possible to build a DAC with a perfect power supply. And, of course, that’s possible but it will have consequences on the price and the size of the housing.

10:28 - The same kind of question is why the internal streamer in the Brooklyn Bridge can’t be made as good as the external network bridges I used. Again the answer is yes and again it will have impact on the price and probably the cabinet size. I have not reviewed the Mytek Manhattan DAC II that, with the same features, costs 2.5 times the price of the Brooklyn Bridge. I guesstimate that it will sound clearly better with it’s 300 watt linear power supply and the € 995 costing network bridge card. And I am sure the digital input circuits are better too.

11:08 - So yes, that can all be done but can or will you pay the higher price? If not, the upgrade path like I described earlier might be nice. In the end you will have payed more but you also enjoy quality improvements earlier. I used the Mytek Brooklyn series DAC’s as base for this video, simply since that is my reference. The same will go for all kinds of other DAC’s, although not all have a provision for an external power supply. Using a network bridge is not the exclusive right of Roon.

11:45 - Both Sonore and SOtM offer working with DLNA, Squeezebox emulation based on the Logitech Media Server, Music Player Daemon and Airplay emulation. See the reviews of Sonore and SOtM products. Using for instance Audirvana 2 or JRiver Media Center you can send music to a network bridge that supports DLNA. Or you can send music from iTunes over Airplay to a network bridge set to Shairpoint. A quality streamer is, of course, also an option. If you don’t already own a DAC, you could opt for a streamer with integrated DAC. And if you do already have a quality DAC, go for a streamer or network bridge that only has digital outputs. In general it is easier to get a USB Audio Class 2 output better sounding than a SPDIF or AES/EBU output. Although when these are very well designed on both sending and receiving side, there will be no sound quality difference. But these devices are few and far between.

12:48 - O, yes, the Manhattan DAC II uses a ESS9038 DAC chip while the Brooklyn II and Bridge use the 9028. Something you might be able to measure but, all other things being equal, I would be surprised if there is an audible difference between these two DAC chips. And on that bombshell we’re at the end of this show. I love to see you back next Friday at 5 PM central European time, in a new show. If you don’t want to miss that, subscribe to this channel or follow me on the social media so you will be informed when new videos are out.

13:23 - If you liked this video, give it a thumbs up. Many thanks to those viewers that support this channel financially, it keeps me independent and thus trustworthy. If you like to support my work too, the links are in the comments below this video on Youtube. I am Hans Beekhuyzen, thank you for watching and see you in the next show or on theHBproject.com. And whatever you do, enjoy the music. .