I guess I will add some comments here, as a Legacy Audio Valor owner, given that someone asked for Legacy Audio owner input. The version I have is the Wavelet 2, which uses a 64 bit audio processing engine. The version here looks to be pretty old. The Wavelet 2 manual does not seem to be available on the company web site.
Something I noticed was that the test was done at volume 97. I have heard from another guy, who also has Valors, that unity gain is at 85. While I simply use a Lumen U1 directly connected via USB (as well as turntable front-end), he has Esoteric SACD, DAC, and pre-amp as front-end and leaves the Wavelet volume set to about 84. There are different modes for different volume level ranges, and he set the mode to remain in a good mode (where noise levels are kept very low). What he did for the selection was start with the volume at 85, where he could hear noise out of the high frequency speaker drivers, then reduce the volume to when it goes silent, and step it up but remains silent. A noise level test would probably be more representative at a volume level of 84, rather than 97. With this setting, his digital front-end (probably over $90,000) is getting what he wants. He was originally skeptical that the Wavelet would support the sound quality (given that it also uses the input ADC), but his concerns were alleviated once he heard the results.
It is my sense that the Wavelet is most logical in a Legacy Audio speaker system. In my case, the speakers are active with over 2,000 watts per speaker to drive the various drivers. 3 channels are used for tweeter/upper-mid-range, lower-midrange/bass, and subwoofers, and the forth channel drives ambient drivers (which fire out the back/top and sides of the speaker). I did originally have the Legacy Audio Aeris in full 3-way active custom configuration, but upgraded back in January. I use a 20 watt per channel integrated tube-amp (as power amp) with 60 year old NOS tubes to drive the tweeter/upper-midrange section. The tube flavor comes from the pre-amp section of the integrated amp, which is why I am not using a tube power amp. The tube amp is pretty noisy, so I changed the Wavelet driver gain settings to have the gain for the tube-amp be 10dB above the other channels. This way, I could reduce the gain of the tube amp by 10dB and so reduce the relative noise level. Tube amp THD can be as high as 1%. However, I will say this tube-amp has orders of magnitude greater resolution than the Proceed AMP 2, which I used to own, which was THX certified. The room correction/speaker adaptation calibration is pretty much automatic. ie. Set microphones at 4 feet in front of the speaker, do the frequency sweep, and upload results to the server, which updates the processing settings. Given that it is so automated, I would think that it would not be easy to use as a general processor. It was originally focused on Legacy Audio speakers, but they extended it be more generic.
Someone on this forum asked what this product was about. It is a digital pre-amp, with (EQ and time-based) room correction, 4 way digital active crossover, with output DACs. I believe someone mentioned 8 DACs. This would likely be up to 4 for each speaker, specifically targeting the Valor speakers. It also includes input ADCs to take in analog inputs. It allows one to enable/disable processing, and does support a 6-channel equalizer (ie. tone controls). Tone controls are performed in the digital domain so that no phase issues come into play.
I will say that the Wavelet takes a long time to boot up (before it gets the IP address). As such, I don't bother switching it off. I only switch off the tube amp when I am not using it. The ethernet connection is used for both getting to the server for room-correction model generation as well as provide a web-interface that can be accessed by a web browser on a PC, iPhone, etc. I use Safari on my mini-iPad to connect to this interface to control the Wavelet. I connect directly to the IP address that is displayed on boot up (rather than having the public server route me to my device). While a Raspberry PI might be used to support the web server interface and provide the control, the processing is done on a 64-bit DSP.
While one can use a microscope to focus on specific components, sound quality comes from the synergy of all the components. The component which has the greatest distortion are the speakers. The interaction of the speakers with the room has a large impart on sound quality. I did see some comments about people feeling "sorry" for people getting the Wavelet. I found this comment funny, given that it is being used in the $86,000 Legacy Audio Valor system. One really does need to experience the system to understand why some people are willing to spend this kind of money for 2-channel audio (due to how realistic the music sounds). Audio shows tend to be less than ideal. I do use this as part of my home theater system. However, for home theater, I would say that I probably would not notice the difference between the Legacy Audio Aeris (at 1/3 the price), and the Valors.
I will say that I am planning to demo the Linn Klimax DSM streaming pre-amplifier in my system. If I find that the sound quality improvement is significant, I may go this route, and simply set the Wavelet volume level to 84.