Now, assuming you’re still conscious, DACs are hard to understand. I think they are worth understanding sonically, but because of the unusual nature of digital processing and the high level of performance here, I’m going to take the liberty to go slightly deeper and longer. If you just want a quick answer to “is it good?” then the answer is “yes”.

What Is the Issue?

An important reminder, or maybe new information, is that very important sonic factors in music reproduction often involve “the six major issues of audio believability”. As we’ve outlined them the big 6 are:

• The problem of recording standards
• The problem of visual images
• The problem of spatial imaging
• The problem of bass in real rooms
• The problem of dynamics
• The problem of a-musical digital distortions

The latter is the focus of our DAC Series, which this review is part of. Because these are difficult issues, they tend to get less attention than circuit and speaker refinements, but these 6 issues are probably more important for making progress at this stage in the evolution of audio technology which is why we’ve drawn attention to them.

For those disturbed by the implication of the term “a-musical” digital distortions, as if there were musical distortions, we use that term to distinguish such distortions from traditional harmonic distortion, which can be qualitatively like the harmonics that are central to the character of musical sounds. The digital distortions we are interested in here are not like that. Digital distortions can be, and the most problematic are, unrelated to the harmonic structure of musical sounds.

Digital distortions instead are often the byproduct of mathematical processes that may have no analogous process in nature. It is important to understand this because it suggests why digital, with much lower measured traditional distortion and noise than is typical for analog signals, can sound so problematic. Hearing is largely the result of complex brain processing, and the brain is wired by evolution to attend to certain kinds of signals — surprising or unnatural signals get extra attention. Or, to connect with our point about problems of believability, unusual and unnatural sounds are distractions because they focus attention on the distortion and away from the music.

Bad dog!

To give some examples of what we are talking about, my favorite example of an a-musical digital distortion is something called pre-ringing. Pre-ringing is a byproduct of the mathematical transformations needed to render the original analog signal as a digital signal and then return it to the analog domain necessary to make analog acoustic output you can hear. The “a-ha” moment for a lot of people comes when we point out that pre-ringing creates a distorted version of the music signal that occurs in the output to your stereo before the music signal that caused it happens. Yes, there is a type of digital distortion where signals happen before the “real” signal happens. I hope it is clear that this is completely unnatural. No one ever knocks on your door before they knock on your door. You are not living in a Steven King novel.

Another excitingly special digital distortion is something called ‘aliasing difference errors’. High frequency sounds can generate an artifact called an alias which is a tone not present in the original music that is created when the sampling frequency is lower than needed to capture the musical signal. Filtering is used to reduce these artifacts, but certain non-linearities in the process can lead remaining very high frequency alias signals to be subtracted from one another and generate tones shifted into the midrange where they are easily heard. And, as a reminder, they are signals unlike what was in the original music.

There are more such digital distortions. Some can be dealt with in the recording process, some can be addressed in the playback process and some can’t.

I’m not an expert on digital processing and I’m not trying to turn you into one. I’m simply trying to point out some of the myriad ways that musical digital processing is not just 1’s and 0’s and is vastly more complex than getting the numbers in a spreadsheet over to the accounting department in an email attachment. Timing of data and artifacts from conversion from and to analog are factors that matter for music signals and don’t matter for spreadsheets. Another way of putting this is that the signal analysis familiar from analog work is unlikely to reveal these digital problems, particularly because digital distortions are a-musical and thus may occur at an analog level that seems minor to a distortion analyzer but isn’t minor to your Version 10,000 evolutionary brain.

To put this in musical terms, many recordings heard today on digital have easily heard distortions if you are familiar with the sound of real instruments. Perhaps the best examples of this are the sounds of cymbals. Cymbals are hard to reproduce because they have transient information which contains significant high frequency content. At the extreme, if a cymbal transient were a perfect impulse response, it would contain frequencies up to infinity Hz. Yup, infinity. A cymbal strike isn’t a perfect impulse, but its transient nature yields signals at very high frequencies. It is possible that these are above the frequency or near the frequency manageable by the sampling rate, and thus are subject to “special” digital distortions.

It doesn’t take long to find transients that on a lot of digital playback equipment sound odd and unnatural. And it isn’t just with cymbals but also voice and piano and wind instruments. I listen to about 250 audio systems per year, and a common pattern you begin to notice is that vinyl-sourced systems rarely have these unnatural artifacts. Digital-sourced systems often do. Before you write to the King and ask for my head, this isn’t to say that vinyl is better than digital. It is to point out that we have ways to understand what “the absolute sound” sounds like and what are playback-related distortions of it (this includes me spending hours with pro drummers understanding the sounds of different cymbals and different techniques).

Sound Quality

Okay, Tom, when are you going to get to the Tidal Camira?

How ‘bout now, but I’m going to cover its sound quality first and then we’ll go into just a bit of the underlying technology. If you care about sound quality, with this one I think you’ll want to know what it does before I stumble around the room pointing at some of the circuitry that might make it work.

Those of you who have followed my speaker reviews will know that I like to use a sort of factor analysis of products. This is a convenient method for describing what a product does and allows the viewer some perspective on what the possible sonic factors are as well as describing how a given product performs on them.

An important bit of context here is that I find in many discussions with audiophiles that they tend to lead with a voicing and frequency response perspective. This makes sense because frequency response is readily heard. Plenty of music will have a moderately wide range of frequencies on display and you can readily hear bass roll off or an upper midrange dip or treble peaks. This is more problematic when we come to DACs and preamps and the like because they mostly can be designed to avoid emphasis or de-emphasis of various frequency bands. So, I’m not going to talk directly about frequency voicing here.

The real problem comes when you have equipment, like a DAC, that has distortions with certain kinds of signals and not with others. You have to use the right signals and wait for them in the right music to hear what is going on. This may be helpful to you in auditioning DACs.

I find that DACs differ in five key areas, at least four of which have this problem of “hiding until you look in the right place”, whence they are in plain sight.

Treble Transient Unpacking

My first DAC sonic differentiator can be called Treble Transient Unpacking. This is the main problem with DACs and believability. It is an occasional distortion, but a very big and obvious one. That’s why it distracts.

What is treble transient unpacking? Well, let’s start with the distortion you often hear with digital playback. On cymbals for example, you will often hear something that sounds like a crash, even when you doubt that the cymbal is of that type or has been struck with big force. If you listen carefully, you will notice that even these crashes don’t sound right. They come across as more like a burst of noise, than a big cymbal hit. It sounds to me as if all the transients of the strike got jumbled in time, leading to something less like real instrumental output and more like unnatural noise.

The Tidal Camira “unpacks” what was previously noise. It becomes a transient followed by ringing and shimmer. The transient isn’t over-spiked and the decay is spread out like the real thing. The transient is often followed by notable tones and waves of tones and different types of tones, not by grit, undifferentiated except by level.

I noted this unpacking with the Berkeley Audio Alpha DAC Reference Series 3 as well. The Camira takes this idea farther, largely by unpacking these treble transients more frequently. The unpacking on the Camira also seems somewhat more complete. So far, the Camira has the best performance I have heard in this difficult and important sonic department. Of course, I will continue to look for better performance or lower cost or both.

I do want to say that the Camira crosses some threshold of listenability that is new ground in my experience. It just requires less “bracing for impact” when treble transients are part of the music. Two qualifications are needed here, however.

Some digital distortions seem to be in the recording (vinyl check is warranted) and the Camira doesn’t mitigate these. There are older (e.g. 1990s) digital recordings where vocals have an edge that seems to be encoded in the file. Maybe someone can fix this, but we often forget that recordings and decimation (converting bit streams to different data rates) have potential distortions before they get anywhere near your equipment.

The second qualification about the Camira is that it unpacks treble transients, but the result isn’t as natural sounding as with the best vinyl rigs unless you use a very good streamer. The pain of digital transients is mitigated by the Camira with lesser streamers, but it seemed to me still possible to make the result more realistic sounding. Indeed, treble transient unpacking improved significantly when I used the Tidal Arkas streamer or the Antipodes Kala 50 streamer/server. I would say the difference the streamer makes is 75% as big as the difference the DAC makes. Maybe as big. With the combination of the Camira and an excellent streamer, we are very close to having our digital cake and eating our vinyl too, at least in terms of treble transients.

I should also mention that the treble of the Camira seems less bright than the treble of some other DACs, but I don’t think this is a frequency response phenomenon. If you unpack transients, they are more spread out and less attention-getting and you perceive this as “less bright”. Which is why, say, violin tone isn’t darkened or muffled on the Camira – it isn’t darkening tonal balance, it is avoiding over-amped transients and conversion artifacts.

Soundstage Width and Depth

Now I was in for another surprise with the Camira. I simply didn’t expect the large improvement in soundstage possible with a switch from a very good DAC to the Camira. Just for perspective, I’d say the soundstage width and depth with the Camira were 25% bigger than very good chip-based DACs. If you do the math, that’s a 50% increase in soundstage size, not that this is precision measurement. But I want you to know what I mean by “large improvement”. It wouldn’t be an improvement if performer placement didn’t seem more natural and simply less generated by the speakers. But it was advanced in this area.

I should add that I have noted some of this phenomenon going from a Topping ESS 9038 based DAC to a BluSound 9039 with QRONO to the dCS Lina to the Mola Mola Tambaqui to the Berkeley Alpha DAC Reference. What is impressive about the Camira is that the improvement is almost equal to all the other steps combined. If imaging information is based on phase information in the recording, then it might be that the unpacking of treble is something of a timing phenomenon, as it would be with phase information. That’s pure speculation on my part, but imaging may be simply another signal that reveals an underlying digital issue, or not, depending on the DAC. And, sorry to say, I have to note that the high-end streamers I used enhanced this phenomenon.

Bass Transient Response

Another feature of the Camira is the snap and drive it lends to bass instruments. The leading edge of notes simply has more detail and power. Again, this doesn’t sound like elevated bass, it sounds like better transient handling. Maybe the same transient unpacking happens, but in the case of bass it sounds more like bass time alignment. In any case, with reference to the absolute sound, it seems more natural and real, and I was surprised and impressed. The combination of the Tidal Arkas streamer with the Camira was especially impressive in this regard, and the dynamic capability of this combination extended above the bass region.

Tonal Density

In one sense, the Tidal Camira has excellent tonal density. This is in the domain we can call ‘continuousness’ which is simply the positive aspect of a lack of grain. The music sounds of a piece, as it does in real life. The Camira delivers continuousness very well, especially if you know real sounds, because some instruments like violins are not themselves grainless, having a texture as the bow slides over the strings.

The other aspect of tonal density derives in my listening from whether the equipment (the Camira in this case) sounds ‘bottom-up’ or ‘top down’ in orientation. This may simply be a byproduct of very slight frequency-shaping, but it sounds to me more like the presence of tiny transient errors, either softening transients in the case of a ‘bottom-up’ sound or highlighting transients in the case of a ‘top-down’ sound. Whatever the case in terms of signal processing, the Camira has a slightly ‘top-down’ sound, though this seems for the Camira to stem from its extra transparency and image definition in the upper midrange and treble region, which simply causes attention to go there in contrast with other equipment that is fractionally more opaque in these difficult bands. I also thought the use of the Tidal Arkas streamer addressed this very small top-down orientation and made things about as neutral as I’ve heard.

Organicness

My final parameter for DAC evaluation can be called ‘organicness’. In an attempt to avoid too high a ‘Woo Factor’ in this description, there is a lot of research in A/D and D/A conversion suggesting that timing errors, and not just jitter but also interpolation errors, are a factor in perceived sound quality. I have spent some time listening to time-corrected vs standard conversion techniques in an effort to explain what work on this dimension sounds like. I’ve landed on the term ‘organicness’ mainly to indicate the time-managed digital signals and digital conversions sound more natural and relaxed and whole. If your reference is the absolute sound, and you have experience with real instruments, you can notice this right away.

I also think the feedback from some audiophiles that time-managed digital signals are inferior may come from using stereo systems as a reference and becoming accustomed to the sound of timing distortions. In any event, I and others on The Absolute Sound editorial team have experienced a distinct preference for organicness in digitally processed music since it better approximates our reference, much as many of us have noticed the success of vinyl playback and tape in certain important dimensions, of which this is one. This is, I think, more a matter of goals and methods than it is a matter of taste.

I think the Camira is better here than many other DACs. That said, my experience is that the state of the art here depends somewhat on the recording process. But the openness and continuousness of playback with the Camira and an excellent streamer removes a subtle sense of the sound being processed.  This puts us in new territory where some recordings sound remarkably like the instrument being played live.

Believability

These comments are intended to be understood within a framework that says our aim with audio for music is to get to a level of believability in the reproduction of music that we aren’t distracted by the sound quality, and we are able to hear what the artists and musicians and engineers were trying to do. Most listeners may not think of believability as their goal in evaluating audio gear and recordings, and that’s fine, but I think it is the logical goal of the audio hobby.

Now, different listeners may have different believability concerns and sensitivities. Our reviewers can’t know those, so we take our job to be describing what the equipment does, within the context of what we’ve learned about possible barriers to believability.

The Tidal Camira makes an important step objectively in the right direction for addressing DAC distortions and for rendering music more realistically than you may have thought possible. I played track after track from saxophonist Noah Preminger and percussionist Christopher Clarino and violist Kim Kashkashian that were delivered quite close to the proverbial “they were performing in my room” level. And I found less-perfectly recorded work from Throwing Muses and Indigo Girls and John Coltrane that were still more engaging through the Camira, an important point because you don’t want to be limited to a small library of excellent recordings. But subjectively or financially you may or may not care about this progress.

A Note on Circuit Design

The Camira LC has a level control (the LC in the name) that is designed to minimize impact on sound quality. Behind the volume knob is a motor-driven ALPS potentiometer which can be turned manually or with the remote. The potentiometer simply generates a signal indicating how loud you’d like to have the music. The volume level selection signal is transferred, via an A/D converter, as a digital volume indication to a lossless level controller inside the 4R4 ladder DAC.

The volume control uses a unique process of how to divide the music signal into equally big pieces without losing resolution. Digital level controls can suffer the loss of resolution and therefore never sound really great. Some manufacturers add an analog volume control after the DAC as a compromise, basically building pre-amp in front of the output stages. Which is like adding another filter in front of it. Tidal, instead, directly control the A to D process within the converter.

The sign-magnitude pure discrete 32-bit 4xR4R ladder DAC has many special features. The sign-magnitude circuit uses two ladders, one for the positive signal and one for the negative. This minimizes crossover glitches resulting from the use of two’s complement data in the PCM signal.  This is happening in the digital conversion process, but designer Jorn Janczak says a rough analogy is to the benefit of using class A vs class A/B amplifier circuits.

An FPGA is also used where Tidal renders 44.1 kHz sampled signals very closely back to analog without the typical tricks like upsampling. For example, the Camira uses real fill-sampling, and has no pre-and post-ringing of the digital filters. Tidal also leverages the FPGA software to do timing correction that delivers bits to left and right channels at the same time. Like an analog signal. Class A analog stages are used with careful linear regulation to get signal-to-noise ratios higher than the best instruments can measure. 20 regulated voltage supplies in all are used. Janczak says “the devil is in the details to make all of this work.”

A big change with the new Camira LC has been to shorten signal paths. This effort has made for quite a packaging exercise, reducing the Camira from 3 chassis down to one. Since parts in this approach have to be inserted and carried on the undersides of some boards, the structure is CNC routed.

This just touches the surface, but suffice it to say, Tidal has paid attention to many factors in avoiding digital and analog distortions. You may be bothered by the price of doing this, and if you are, it might be wise to think of the Camira LC as a test bed for new ideas that will make future DACs better. Experiments done by small companies aren’t free, and it might be better to have the experimentation and then figure out how to cost reduce it than to not have the experimentation at all.

Summary

The Tidal Camira is the best DAC I have tested to date. It is also the most expensive. Being high-quality is good, but to me the real win with the Camira is that it crosses a threshold on treble transient quality and adds enticing benefits in soundstaging and bass delivery. I am deeply impressed.