Humans can hear up to about 20,000 Hz, which is why CDs have a sampling rate of 44,100 Hz, a little more than twice the highest perceivable frequency as explained by the Nyquist Theorem. Yet many producers, engineers, and listeners swear that music recorded and played back at higher sampling rates sounds better. A sampling rate of 192 kHz yields a frequency response up to 96,000 Hz, which is in the realm of dogs and bats. How can recording those frequencies improve the sound that humans can hear? Or is there something else about the ultrasonic range that improves a recording’s sound quality?
These are some of the questions that inspired Meridian to develop a new audio-encoding technology called Master Quality Authenticated (MQA). Scott Wilkinson recently wrote about the December 4 launch of MQA at an event held in London. Earlier this week, I had the opportunity to attend a similar demo at Meridian’s offices in New York City.
Bob Stuart, co-founder of Meridian, gave a presentation in which he explained the science and concepts behind MQA. The presentation began with a graph that illustrated how consumers gravitate toward convenience, even at the expense of quality. For example, the cassette succeeded the vinyl LP despite its lower quality because it offered ease of use. Then came CD, which debuted with the promise of bringing quality back into the equation while offering even more convenience than cassettes.
As Bob said, “When we think about sound quality, most of us think CD is better than cassette. Consumers found cassettes quite convenient, but CD was better because it had a pause button, and you didn’t have to rewind it. These days, downloading is convenient, particularly if you have a good cloud system behind it, and streaming is even more convenient. This tells a story of how consumers behave, but also how the quality has gone down.
“MQA is about trying to put the quality back but keep the convenience. It’s based on a lot of research over the last several years. Key insights from the auditory sciences and digital-sampling theory lead us to a method of capturing audio that is better and clearer sounding than conventional digital audio.” The crux of it is this—the timing data contained in the ultrasonic frequency range has value when it comes to precisely reproducing audible frequencies, but it takes up far too much bandwidth using conventional PCM sampling.
Meridian’s solution is packaging the audible spectrum with the timing data from the inaudible spectrum as metadata. The result is a file that plays like standard CD-quality PCM audio on a regular player, but on an MQA-enabled player, it has the precision of hi-res audio. And by “precision,” Bob is talking about improved impulse response—reproducing a sound at the exact moment you are supposed to hear it.
In this graphic, Bob Stuart points to the difference in impulse response between CD audio and MQA.
In addition, the technology authenticates what is digitized at the studio as well as the resulting output from your DAC, making sure they sound the same. Bob says the result is an exact and verified copy of what the artist approved, unaltered by variations in timing between different DACs. Furthermore, when you play an MQA file in a player that doesn’t have an MQA decoder, it still offers CD quality sound and backward compatibility.
Because MQA does not encode all the ultrasonic information directly, but rather as metadata, the resulting file size and bitrate are substantially less than high-resolution PCM files. How much less? According to Bob, MQA files are about 5% the size of uncompressed 24/384 PCM files. As a result, they can be easily downloaded and even streamed without requiring lots of storage capacity or bandwidth, making them much more convenient that conventional high-res audio.
After the technical presentation, it was time to listen to music. Meridian used a pair of its self-powered, MQA-ready DSP7200 speakers. The DSP7200s are very high-end and revealing, and the MQA tracks Bob played sounded fantastic. However, there was no comparison to non MQA-encoded music, so it was not clear exactly how much of the fidelity I heard was due to the MQA format and how much was due to the capable speakers and the intrinsic quality of the recordings.
One of the tracks was “Get Lucky” by Daft Punk. Bob joked that most of the audience probably only ever heard that track playing in a bar. I’m not in that group—I use “Get Lucky” as a reference track at high-end audio shows and when I review speakers and headphones; I’m deeply familiar with it. Despite all that, I can’t say I could hear exactly what MQA did for the track—was it tighter? Did it have a better soundstage? Was the bass more impactful? There was no way to tell without hearing a non-MQA version. I do know that “Get Lucky” still sounded fantastic when I play a non-MQA version at home on speakers that cost a fraction of what the DSP7200s go for.
Aside from the DSP7200 demo, Meridian had a room equipped with several laptops and high-quality headphones, including models from Sennheiser, Grado, and Audeze. Several Macbooks had a copy of Audirvana running, with a playlist that included a number of songs in 128 kbps MP3 and 1 Mbps MQA format. Unsurprisingly, it was possible to tell the two files apart, but because of the amount of ambient noise at the event, it wasn’t exactly easy. I’d like to perform a similar test in a totally quiet environment using 256 kbps iTunes files vs. MQA’s 1 Mbps streaming format, as well as CD-quality audio versus MQA. Another interesting demo would be to see if MQA can be distinguished from high-res PCM—theoretically, they should sound the same, even at very different bitrates.
I could tell 128 kbps MP3 apart from 1 Mbps MQA audio using this rig.
I’ve reliably heard the difference between 128 kbps and 256 kbps MP3s in double-blind tests, so the MP3 versus MQA via headphones comparison offered me no real insight about how good MQA audio really sounds. It’ll take a much more thorough demo before I can understand what MQA brings to the table in terms of real-world, audible sonic benefits. I look forward to that—perhaps it’ll happen at CES 2015. Meridian did promise that there’s a lot more info on its way in the coming months.
Near the end of the presentation, I asked Bob why he thought MQA would succeed in a market where a lot of good ideas do not. He said that Meridian would not have embarked on this project if it did not seem worth it, and that he believed it would be widely adopted. Meridian has deep roots in digital audio—the company is responsible for dither algorithms used in CD mastering as well as the lossless compression known as Meridian Lossless Packing, which is the basis of Dolby TrueHD. Thanks to that heritage, I think Meridian has a reasonably good chance of succeeding in its quest to make MQA into a mainstream audio format. Now, I just need to find a demo that reveals a real difference between CD-quality and MQA-encoded audio.