Now that the dust has settled and I have had some time to collect my thoughts on the M scaler, I thought I would post some impressions, based on what people heard at the show.
Firstly, some background. I first heard the 512,000 tap WTA filter in late July - and was pretty much knocked out by the changes in performance, so then started work on the full M scaler, which I got to listen too in late November. The final version, once all the audibly changing bugs was fixed, was last Christmas eve. So we are talking about very recent developments.
Now part of the difficulties about designing is appreciating the scale of what one achieves. Its easy to say this improvement has this effect in SQ and sometimes it easy to get tunnel vision and to exaggerate how significant a change actually represents, and I am very aware of these dangers; ones own natural enthusiasm can let one get carried away. But assessing scale of changes is vitally important - just to give you an example, it took 3 months to code verify and test the M scaler - and this is based upon merely modifying from 512,000 to the full 1M taps. So you are making decisions that will have important SQ consequences based upon previous listening tests, so balancing the importance of different tests is crucial. I always find it amusing when a audio companies blurb talks about no compromise this or that; design is always about balancing resources to get the best performance you can; compromise, fine tuning or balancing is essential part of the process. So getting a sense of scale based on listening tests is absolutely crucial, as I need to be able to accurately predict at the design stage what a decision will mean. But appreciating how much of a difference actually represents is difficult - for example, when listening to depth, it often just resolves down to A being deeper than B. But what of listening test conducted 2 years ago where C was deeper than D? How can you relate A improvement to the C improvement? This is where getting a perspective on the scale of a change is really important - and it becomes crucial when designing lower cost devices such as Mojo, where you are limited by cost and resources and important compromises have to be made.
So whenever I get to do a listening test, appreciating the scale of the change is vital, as the results will get logged into my mind and used maybe in design decisions 5 years hence. Now with the M scaler I was already knocked out by the change - I knew that increasing tap length would sound better, but I had not expected it to be transformational - and my listening tests and actually using it to enjoy music was telling me this. Because it was so big, I doubted my own opinion, and wanted others to hear it and get their feedback, so I can get a more accurate measure on the scale of the changes. I had been talking to
@romaz during the process, and I was deliberately being neutral about the change and my intense excitement, so that when he heard it at CES I would get a unbiased view.
Now trying to asses scale is important, but you need to be aware that listening in strange surroundings, with unfamiliar gear reduces ones sensitivity, so listening at home will give a much more profound and accurate view. It's one reason why we tend not to do AB tests at trade only shows, as the dealers and press will get to grips with it in their own homes soon enough.
So Romaz (Roy) was only the second audiophile to hear the M scaler properly - and you can read for yourselves his impressions. And doing the AB was amazing - you could immediately hear the effects of plugging the M scaler in - and absolutely everything improved. So much so that after 1 minute Roy declared "I gotta get one of those."
Next came up was
@jude and I used a 1972 Decca recording of Vaughn Williams Fantasia on a theme by Tomas Tallis. Now this is recorded in London's Kingsway Hall, which has the underground nearby, so you often get to hear the rumble of subway trains. Now when the M scaler was switched on, you could hear a collective gasp by all those present - the rumble sounded
completely different - before it was this vague noise - now you actually perceive the pitch of the rumble. Frankly, it was really weird. Next we noticed small ticks and noises from the musicians, and these little disturbances were really clear and precisely located in space.
So I was now getting pretty excited, and was demoing the differences whenever I could. The reaction was universal - everything changes, and it isn't small. Two people commented that the size (scale) of the change was bigger than Hugo to Dave (not sure about that but I can't disagree as it could be). One guy jokingly said I had distorted the direct feed - and I know what he meant by that, but the direct feed is bit perfect. So what changes? The first thing you notice is just clarity - everything is so much clearer and more transparent. Timbre variations are much better, together with pitch reproduction of bass notes. Rhythm's are much easier to follow, and instrument inner detail is easier to hear. Instrument separation and focus is much better, and its much easier to follow rapid variations. Now all of these are expected; they are the usual stuff from improving timing of transients accuracy. But what is also better is depth perception, which is usually nothing to do with timing but small signal amplitude linearity. Now this is better because I have done some more things than simply improve the tap length. One of the curious things was switching on the HF filter with Dave - with 44.1 it should not sound better - and this immediately told me that I needed to improve the WTA filter stop-band performance, and this was done by increasing the bit depth on the quantised coefficients. This worked; now M scaler sounds better with the HF filter off (exactly as it should do). But one unexpected benefit to this has been better depth perception.
Regrets? Yes - I should have trusted my gut reaction, that the M scaler was a profound musical and SQ change, even though my intellect was telling me it's only two and a half bits more accurate. And when launching the M scaler we should have done an AB test to the press at the launch so they can actually appreciate how big the change is. Chord will be doing AB demos of it at the Bristol Hi-Fi show in the UK in February.
Rob
