TOUJOUR PAS DE TEST !!!!!!

Hi Everyone, here are a few thoughts from the team

Preston Whiting here…I’m the Chief Operations Officer at Black Lion Audio.

First, we would like to thank everyone who’s given the Revolution 2X2 a shot in your studio. Your overwhelmingly positive calls, comments and reviews have been graciously received. Your support means the world to us and we do this for you.

When we decided to release our own interface offering we did so fully recognizing that we audio enthusiasts are spoiled for choice. Having dissected and modified more interfaces than anyone else on the planet, we could not in good conscience follow the general conventions or design patterns we’ve worked so hard to improve throughout our history as a company.

Our obsessive experiments over the last 15 years of modifying interfaces has wired us differently than other pro audio companies. While we fully agree with publishing technical specifications of our products – we are in strong opposition to using numerical data measured by a machine as the ultimate gauge of success.

Our unwavering integrity demands complete transparency in publishing our measured specs. To be frank, we’ve found many of the specs generally touted to reveal the quality of an audio interface to be missing important context or not matter at all. We’ll continue to think outside the box, achieve superior results and quantify new ways of measuring performance that helps us hit the mark. What we won’t do is hold back a phenomenal product because it doesn’t fit neatly into the crowd.

As a final point – I’d advise against being dissuaded from experiencing the Revolution 2X2 based on a review rigidly rooted in test equipment results. We must use our ears to make our audio decisions. We wouldn’t judge a pair of shoes based on how well they fit on our hands.

We were going to leave our response to the review at that but our lead designer Jesus Ortiz was chomping at the bit to approach this from a more “technical perspective.”

The first thing that caught my eye is his graph showing our noise floor. He gave an FFT graph of the noise floor relative to the tone and a few of it's harmonics. The interesting thing about that graph is that the THD without the noise taken from that graph is close to around -114dB. Not bad at all.

Later he mentions that our DAC can't even deliver 16bit transparency. The issue is that using the word "transparency" is very misleading. By transparency one would assume he's talking about quality. In reality I'd guess he's referring to a similar spec that is used in Comm called Equivalent Number of Bits (ENOB). It's basically a "real world" estimate as to how many actual bits of useable range you get if you deduct the bits that are just noise floor. Again, this is noise floor related, although normally SINAD is used. Using this spec without context is a little misleading because he's technically correct in that an ideal ENOB 16b measurement would be for about 98.08 which we're just below. That's pretty common since everything has noise. If he's testing this at 16bit resolutions then to say that we're below the 16b ideal ENOB just means that our interface is not a simulation and contains real devices. For perspective, a SINAD for an ideal ENOB of a 24bit system is about 146dB. No interface hits that. 24bit or not. Going further down, an ideal ENOB for an 18 bit system would have a SINAD of around 110dB. Very few interfaces hit that. Even when weighted. Put differently, a 24bit systems with a SINAD of around 110dB has an effective number of bits of about 18.

When he refers to his ideal output level being 4Vrms while ours is about 3.2Vrms, there's very little context as it's one of the few times he refers to signal in volts. Most of the rest are in decibels. For perspective, the difference with our level and his is under 2 decibels relative to the dBu scale.

Later when he talks about IMD/Noise Floor at max volume, there's no real reference between the devices being compared. "Max volume" is assumed to be the volume know turned all the way clockwise but that's not the same for every device. The reader wouldn't know if one device has 10dB of gain at their output at that setting while another interface may have -6dB of gain. This should be spec'd in some way like the inputs have Equivalent Input Noise where the gain stages are compensated for gain or attenuation. The idea behind it is that if my converter puts out a decent level, but the noise floor is say -96dB, then if I cut 6dB then the noise floor where he's measuring is -102dB. Not an accurate value as one can argue that the noise floor value was manipulated. On the other hand, a noise floor of -110dB fresh out of the converter but that receives 20dB of gain along it's way to the output would then read a noise floor of about -90dB. Again, not completely accurate. This actually goes for many of his noise floor tests. They're all done at the output knob set to max and our unit actually has positive gain at that point which brings up the noise floor. Even the output level is a little misleading because he's testing it at (in our case) 3.2Vrms. This is around +12dBu. While not out of the realm of user levels, it may make more sense to test at a more traditionally used line level of around +4dBu or so. That'd be around 1.228Vrms.

One of the important thing to keep in mind is that we're arguing two different things. He doesn't see our perspective and we don't see his so there's likely to be no agreement from him. One of the things we focused on was internal jitter from the master clock. During the process we saw the difference it made. I'd say "night and day" would be an understatement. The problem is that he specs jitter, but I'm not particularly clear as to what he means. I suspect it's SPDIF which isn't a particularly great jitter source for any device. SPDIF encodes the clock into the stream but the standard process to do this process is usually less than ideal. What's more, there's no detailed distortion or THD measurements without noise so it's hard to distinguish between the two and frankly they're not completely related in terms of sound quality vs pure functionality. Most of his THD related tests are single channel. The fact is that one of the biggest differences we've seen in jitter has to do with the stereo image meaning that correlation should be taken into account. Recently I did a quick preliminary test where I used standard WAV audio files and generated random jitter at different levels. I superimposed the jitter into the digital WAV file and took the distortion measurements. Jittered vs jitter free samples always showed higher distortion but that's not my point. Using different types of jitter and different treatments on the jitter I was able to show that the variations on the same jitter sample embedded onto WAV file showed audible difference in the image, spacing, and placement of the mix, but at the same time, the distortion caused by jitter calculation for every one of the WAV files was exactly the same. The phase response polar plots definitely showed different phase correlation between the two channels.

All this to say that while his measurements are sound and well done, they can't be taken completely as definitive. They're taken as a strictly R&D/QC Engineer perspective which can often times be a rather narrow way of seeing things. On top of that, measurements and interpretations that don't match these specs are often automatically dismissed as irrelevant when in reality we should be concerned as to why these measurements and interpretations that seem beyond the limited industry approved standards do point to other things going on in the system.

And a little context from myself on our design process

our industry is in a never ending loop where interface manufacturers are trying to one up each other on specs as the singular goal and hope that the interface sounds good. We have now spent 15 years making exactly those interfaces sound better.

Our design process is exactly the opposite. We spend countless days obsessing over every single component that goes into our mods and even more so the Revolution 2X2. We listen to every single component and hold double blind listening tests to decide on which circuit implementation and component choices sound best. once we have come to an unanimous decision on what is the the single best combination available, it is then that we measure the specs. Whatever the specs are... they are.

We firmly believe that designing to numbers on a piece of paper and not the sound quality and overall experience is counterintuitive to the artistic process. We believe that designing to a piece of paper results in an interface that looks good on paper and that rarely translates to a tool that professionals prefer. Because of our obsession over sound quality as well as the process that we have honed over 15 years, we are confident that users will be absolutely floored by the way the Revolution 2X2 sounds.