Q&A With Mike Moffat, Schiit Audio

Schiit Audio is well-known among audiophiles as a company that delivers well-liked, well-loved?, products at affordable prices. Mike Moffat along with Jason Stoddard are the men behind Schiit, each having a long and storied hi-fi history. Mike Moffat generously agreed to take part in this Q&A, which I hope will give you a better understanding of the company's products and their attitude, which is anything but...

I read Herb Reichert's review of the Schiit Audio Mani phono stage in Stereophile where he mentioned that the famous Loesch phono preamp was inspired by one of your designs, something I didn't know about you and I'd imagine other people equate you with digital design. Can you tell us about your background in hi-fi?
I had no idea about the Loesch preamp or that it was inspired by an earlier design of mine. My background was building analog designs prior to 1982. My first manufactured product was very different for its time, where all competing designs were utilizing 12AX7 tubes with feedback equalization. My design was 6DJ8 based with passive equalization. This was in the mid 1970s. I had a lot of fun building offbeat for the time tube designs until I wavered into digital in the early 1980s. At the time, early digital sounded so vile that I was interested to see what I could do.

The Theta Digital DS Pre from 1988 was the first DSP-based outboard digital to analog converter. Can you tell us about the DS Pre?
As far as I know, it was the first outboard digital to analog converter, period. The Wadia unit was available a few months later. The Theta units took years to develop, combining a time and frequency domain proprietary filter, running in the first available DSP processors, the TMS320 series. It had high output, another analog input, and a volume control. The DSP Pro was released months later, without the quasi-preamp function.

How did Schitt Audio come about and can you talk about the name as well the general Schiit attitude?
It was named by my partner's wife, as a joke. There was endless debate over what to name it. When there was no solution to the naming dilemma, we decided to name and spell it Schiit. Jason and myself have 70 years combined experience in the audio market. We have seen fads come and go.

"Our attitude is to build the best for less, period."

There is an ever increasing youthful market and demand for high quality audio in support of computer based systems. That was our start point. Our attitude is to build the best for less, period. With the need for, and our subsequent production of stereo digital to analog converters in that market, it becomes obvious that we are headed into more mainstream, 2-channel, audio products. Since we are direct, our performance/price ratio is doubled. Our products are designed to be simple – this means no dealer intervention or installation is required, saving our users big bucks in the process.

We have no nose in the air qualities. We discourage overly serious, hand-wringing, superior attitudes. We know that no matter what we build, there will be some who will not like it. It is up to us to build products that succeed. It is none of our business what others think of us. We know what we can do and just keep doing it. We just keep doing what we have been doing for years and years. Most of the time we are right in product selection and targeting. In the few occasions we have been wrong, we move on. We are not perfect, but we have a hell of a time. Read our FAQs.

Schitt makes DACs ranging in price from $99 to $2299. What should people know about DACs and why do some cost more than others?
Our DACs below $400 and below are Delta Sigma. Audio is the red-headed stepchild of electrical engineering in general, several prestige notches below say, microwave rf or computer hardware design. Delta Sigma DAC chips are the delight of “audio” chip makers. They come complete with cookbook “how to make” data sheets and reference designs to copy. They are also cheap. We could not design a $100, or even $150 digital to analog converters without them. DS designs are so easy to make that the result is that are many of them (not all) on the market designed by digital engineering unsophisticates. The keyword is cheap and easy.

"I do not believe in making non oversampling (NOS) converters. The required NOS analog “brick wall” filter has multiple poles and is very prone to ringing and horrendous phase shift."

All of our converters at $600 and above are multibit and multirate. By multirate, we mean oversampling. I do not believe in making non oversampling (NOS) converters. The required NOS analog “brick wall” filter has multiple poles and is very prone to ringing and horrendous phase shift. Our multibit converters are also upgradable (as well as our $400 DS Bifrost) as technology becomes available. Although our policy is to make no sonic claims for any of our products, my clear preference, not only for headphone systems, but even more critically so for speaker systems is multibit. Perhaps you may agree.

Unfortunately, they are far more expensive to develop and build, particularly when proper instrumentation and weapons grade DACs are used. This is opposed to the cost constrained “audio DAC chips” marketed to the “audio” sector. (Read DS) Multibit DACs are capable of converting a literal copy of the recorded data. DS ones cannot.

Can you provide some details on the Closed-Form Digital Filter used in the Yggdrasil?
The below are the claims of the Digital Filter/Interpolator in the Yggy/MB Gungnir/MB Bifrost:

  1. The filter is absolutely proprietary.
  2. The development tools and coefficient calculator to derive the above filters are also proprietary.
  3. The math involved in developing the filter and calculating has a closed form solution. It is not an approximation, as all other filters I have studied. Therefore, all of the original samples are output. This could be referred to fairly as bit perfect; what comes in goes out.
  4. Oversimplified, however essentially correct: The filter is also time domain optimized which means the phase info in the original samples are averaged in the time domain with the filter generated interpolated samples to for corrected minimum phase shift as a function of frequency from DC to the percentage of Nyquist - in our case .985. Time domain is well defined at DC - the playback device behaves as a window fan at DC - it either blows (in phase) or sucks (out). It is our time domain optimization that gives the uncanny sonic hologram that only Thetas and MB Schiit converters do. (It also allows the filter to disappear. Has to be heard to understand.) Since lower frequency wavelengths are measured in tens of feet, placement in image gets increasingly wrong as a function of decreasing frequency in non time domain optimized recordings - these keep the listener's ability to hear the venue - not to mention the sum of all of the phase errors in the microphones, mixing boards, eq, etc on the record side. An absolute phase switch is of little to no value in a non time domain optimized, stochastic time domain replay system. It makes a huge difference with our MB units. This switch is incorporated in the Yggy.
  5. This is combined with a frequency domain optimization which does not otherwise affect the phase optimization. The 0.985 of Nyquist also gives us an advantage that none of the off-the shelf FIR filters (0.907) provide: frequency response out to 21.71925KHz, 43.4385KHz, 86.877KHz, and 173.754KHz bandwidth for native 1,2,4, and 8x 44.1KHz SR multiple recordings - the 48KHz table is 23.64, 47.28, 94.56, and 189.12KHz respectively for 1,2,4, and 8x. This was the portion of the filter that had the divide by zero workaround which John Lediaev worked out in 1983, to combine with #4 above AND retain the original samples.
The Loki is the only Schiit DAC that supports DSD. What's the deal with DSD?
DSD is an incompatible with prior art audio record(or mastering)/playback system. In this sense, it has much in common with Quadrophonic records, 4-track and 8-track tape players, Elcaset tape players, open reel tape players, cassette tape players, DATs, mini discs, and some others which I may have missed. These are marketed as either technically superior, (DSD) or more convenient (cassette tapes). I was never a technical fan of DSD (a further extension of DS DAC tech – further away from reproducible numbers).

"My experience of the last few dozen years or so told me that yet another scheme to convert recording studios and resell the music lover his music collection once again was almost certainly doomed to fail."

My experience of the last few dozen years or so told me that yet another scheme to convert recording studios and resell the music lover his music collection once again was almost certainly doomed to fail. Further, with DSD, the huge size of the files made it a virtual certainty it would neither be streamed by any major providers nor its music sold by the likes of Apple or Amazon, necessary to guarantee the success of the format. The problem was that I had a zillion requests to make something for the DSD format. So I caved and built the Loki, a low risk product for a niche I adjudged to be stillborn.

I was wrong in the sense that it was not stillborn, just very low survival potential. It has been helped by the fact that the many of the DS DAC chips mentioned above in question #3 have incorporated DSD as a buzzword “feature”. We actually took a flyer and built 1000 Lokis which had been (sales-wise) slowing until we recently ran out of them. We had a Loki funeral party at Schiit and will make no more DSD products – at least unless the format ever recovers from its condition of being almost on life support as well as being distributed by the majors.

There's been some controversy over the Schiit Wyrd, which is a "Audiophile USB Hub". From a technical perspective, why does the Wyrd make a difference in terms of a system's sound quality?
That is an interesting concept, particularly given Schiit's policy of making no sonic claims for any of our products. We do that not so much out of sonic agnosticism – we well know that products sound different; rather, we prefer to let our customers determine for themselves what pleases them sonically. I also strive to find measurable differences between sonic differences. I have several projects to that end in process now.

Returning to the Wyrd, it was initially designed as a fix for the very widely variable performance of USB sockets on computers in general, and PCs in particular. Since PCs come from many makers, the quality variance over a large sample of PCs varies significantly according to cost. The connector quality varies as well. What this means in the real world is that one USB socket on one computer does not equal another on different machines, nor even on the same one.

"What this means in the real world is that one USB socket on one computer does not equal another on different machines, nor even on the same one."

A significant cost for Schiit, one that becomes critical when we sell tens of thousands of $100, $150, and $400 digital to analog converters is that of customer service.

Hence the Wyrd – it was built as an inexpensive, quality USB Hub for the purpose of streaming reliability. It was the first product of its kind, and is well suited for that purpose. We are well aware of the fact that many have noted sonic differences, even though there is little in the way of consensus. That may make sense in system sample to sample variability, given the differences in source USB, and target converters. This gives rise to an exponential rise in YMMV, and makes it even more important with this product to stress we make no sonic guarantees. It is first and foremost a reliability enhancer.

What are the top five aspects of a DACs design that directly influence how it sounds?
Multibit, multibit, multibit, multibit, and finally – multibit. Multirate with a proper filter is required for best reproduction of Redbook material. To clarify, “proper” means time and frequency domain optimized.

If you could describe your design philosophy in one sentence, what would it be?
To sonically delight as many people as possible as inexpensively as possible with technologically sound products.

What can we look forward to from Schiit Audio?
I am always looking for ways to reduce the cost on Multibit design. At $600, the Bifrost MB is the lowest end of that scale. I hope to make Multibit tech more affordable as time goes on, without resorting to major cop-outs such as NOS. As I attend more and more headphone meet-ups, I see an enthusiastic group of young audiophiles who are much like I was at that age. They are very limited in their funds available, and very unlimited in their enthusiasm to listen to music on better equipment. The more we bring in, the better. This is required for a legacy, not just mine, but that of our hobby. If we collapse into a series of products which are priced above cars, our pastime is doomed. We must remember to play to our customers, not other manufacturers. We must remember not to take ourselves so seriously that we deny ourselves the fun of the hobby. I digress.

"We must remember not to take ourselves so seriously that we deny ourselves the fun of the hobby."

There is one other product that I am working on which is a new and completely different product category. It is not a D/A nor A/D converter, amplifier, DAP, nor anything resembling any product ever built in support of audio. The seeds for it were inadvertently planted when I played in a bluegrass group 50 years ago and wondered why expensive, old instruments sounded far better than newer ones. We call it “The Manhattan Project” and I have a prototype “Gadget” running which is absolutely hair raising. It operates only in the digital domain.

COMMENTS
Bill Leebens's picture

I believe PS Audio, Arcam and Musical Fidelity all had outboard DACs on the market before Theta.

Of course, in digital terms, that might as well have been the Pleistocene Era--thirty years ago....

judmarc's picture

Mike's was the first *good* outboard DAC. ;-)

(ducks and runs...)

jim tavegia's picture

Keeping in mind the project you are working on and this would really peak my interest as I don't believe that one needs to spend $1K on a single channel mic pre to capture what acoustic instruments have to offer.

I hesitated to complain about the quality of the music I bought as once I started recording I realized how difficult it can be to capture a great performance. I no longer do computer recording as it had become the most untrustworthy medium and now just use stand alone Flash Recorders from Tascam with SDHC cards at up to 24/192.

draenor94's picture

If someone would dial back the big billboards I might actually be able to read the article. I had to dig out my Aviators. They fogged over so much, I feel I like I fell into a sceptic tank (pun intended).

Michael Lavorgna's picture
...because they provide a nice illustration of their attitude.

Sorry you have an issue reading things with pictures.

dysonapr's picture

Nice to see Mr. Moffat is still living in the real world (that's the one where median household incomes are not keeping up with inflation).

CarterB's picture

I don't have any Schiit DACs due to lack of DSD but I like to keep up with their products as they seem to be a bargain. I hope the new product hinted at in that last paragraph is different and fun as it sounds like it.

dougmon's picture

Mr. Moffat says:

'I do not believe in making non oversampling (NOS) converters. The required NOS analog “brick wall” filter has multiple poles and is very prone to ringing and horrendous phase shift.'

I've been reading up on NOS DACs and one thing they all claim to have in common is no filter whatsoever. So what about the brickwall filter is required?

judmarc's picture

NOS DACs don't have upsampling/oversampling/interpolation filters. What they do have, unavoidably, is a final filter to turn the digits into music. (You're hearing music, not digits, right? If that's so, there's a filter.) This final filter is sometimes what's referred to as an "analog filter" like capacitors and resistors, or even just a transformer (like some Audio Note kits).

All filters have a problem to solve, brought on by math and physics. It's this: The better "time domain" (ringing distortion) behavior the filter has, the worse "frequency domain" (aliasing distortion) behavior it has. The best way to minimize this problem is to give the filter a lot of headroom to operate. That is, if the filter's task is to knock down all the stuff from 20KHz to 40KHz that could alias into the audio band, starting the cut at, for example, 352.8KHz, is a lot easier than starting it at 44.1KHz. That's why CD players and DACs converted to 8x oversampling (moving the signal to 352.8 or 384KHz before final filtering) as an industry standard by the early 1990s.

NOS DACs go back to the "bad old days" of the first CD players that many people at the time thought sounded horrible and gave the phrase "digital sound" a bad name. In these DACs, the final filter starts (at least with RedBook input) at 44.1KHz, making it very difficult for the filtering to do a good job with time domain and frequency domain distortion (see Mike Moffat's description).

judmarc's picture

Sorry, just to relate my last note to the "brickwall" issue: Because the final filter in an NOS DAC has to cut so abruptly (it starts with a 44.1KHz signal and has to be down to pretty much nothing by 40KHz), what you would see on a scope trace of the filter response is essentially a "brick wall" where the signal goes from full strength to nothing relatively suddenly. That's how these filters got their name.

dougmon's picture

Thanks, that's helpful. So it seems when it is claimed that an NOS DAC has no filter, the claim refers only to digital filters?

judmarc's picture

Exactly - and specifically, digital interpolation (a/k/a "upsampling" or "oversampling") filters used prior to the final filter that converts to analog.

Michael Lavorgna's picture
You may be interested in reading Bruno Putzeys' approach in his Mola Mola DAC where he talks about how he deals with exactly this issue:

Bruno Putzeys Talks DACs

judmarc's picture

Happy holidays, Michael.

You in turn may be interested in an Open Hardware project where you, too, can have a shift-register-based DAC like Bruno, for ~$400 worth of parts: http://www.signalyst.com/hardware.html

Unlike Bruno's, this is designed for PDM rather than PWM input. I would be interested to hear from the designer what he thinks about Bruno's remarks regarding PWM versus PDM.

Meanwhile, Mike Moffat's multibit DAC designs (which I heard and enjoyed very much at RMAF) don't need no steenkin' PDM *or* PWM. They're pure mainline PCM only.

Lotta different ways to skin a cat, audio-wise (and a lotta different letters you can stick in between P and M, apparently).

Michael Lavorgna's picture
"In listening tests I just found that the classical transition band, which is literally 10% of the Nyquist bandwidth, is just too narrow. It's just enough to be audible. Make it a tiny bit wider and immediately the signature of the brick wall disappears."

I'm most interesting in hearing music ;-)

Happy Holidays!

judmarc's picture

Essentially what Bruno is saying in the bit you quoted is that the steepness of the filter should be relaxed for better sound. There are many filters available that have relatively relaxed slopes, among them Ayre's (including the Pono) and some of the filters offered by software players. When Mike Moffat says he starts the cut in his filter design earlier than most others, and that the filter is optimized in both frequency and time domains, that makes me think the slope of his filter is likely somewhat relaxed as well.

I haven't read anyone else saying what Bruno did in your interview about PDM vs. PWM (in fact I'm unaware of another DAC using PWM), so that's the bit that interested me.

Michael Lavorgna's picture
It also dispels the notion that all NOS DACs employ a decades old brick wall filter. Someone reading that statement might very well make the incorrect assumption that all NOS DACs suffer from the same problem.

Happy Holidays!

dougmon's picture

That's interesting. I recently purchased a NOS DAC, and I really don't hear any of the effects of what I think of as brick walling -- so this made me wonder.

Michael Lavorgna's picture
...are not created equal. If you read our reviews of NOS DACs like the Metrum Pavane and the totaldac, you'll see very different approaches. I also happen to have enjoyed both of these DACs...a lot. The totaldac is, and this is no secret, the DAC I've most enjoyed, to date.

To put a finer point on my point, saying that all NOS DACs suffer from XYZ is like saying all delta-sigma DACs suffer from ABC. While this may be true in a very general sense, it has absolutely nothing to do with how every DAC sounds or, of greater importance, how much we enjoy listening to music through them.

Reed's picture

I like that they are genuine and approachable. They are the kind guys you would like to sit down and have a bee with and talk audio. Also, when I was there, at least three audio headphone designers brought prototypes of headphones for them to listen to. Each time they stopped what they were doing, talked to them, listened to their headphones and then offered an opinion and encouragement.

BTW...the polar opposite of Richard Vandersteen. I really liked the Treo speakers I heard in several booths at the show and wanted to pass on praise. What I got was the "you are far below my pedestal" treatment, so no Treos for me. I was recently at a local audio store event and Totem was there. The Totem guy treated every person there like they were the only person in the room. I bought a pair of Forests on the spot because I like the sound and found an attitude I want to support.

We need more of that attitude in the industry.

X