The QBD has 16 elements on the pulse array, not 4 on Hugo, and has balanced outputs. But of course it is seriously rivalled by Hugo, so I am currently working on it's replacement, which will feature all of the improvements made over the last 7 years (QBD76 is 7 years old now) that has gone into Hugo, plus much more taps, plus the benefits that the QBD76 currently has (primarily the 16e pulse array).
Now when QBD 76 was under development it was initially planned to use 4096 taps with the same filter structure as in the DAC64, which was 1024 taps. But on listening to various filter options, I noticed something which at the time seemed very odd - it did not tally with my understanding of how the filter worked - the tap length of secondary stages made a big difference to the sound. So this led to a major overhaul of the WTA filter, which ended up being a flat 8 times over sampled (OS) structure, rather than the previous 2 times WTA followed by a 4 times OS filter. This realisation, that the secondary filters were important, gave much more insight into what the interpolation filter needed to do, and opened up more avenues for improvement. This led to a complete change into the WTA algorithm itself, with loads of listening tests to fine tune the recipe. Eventually it got to 18k taps, with the algorithm (or recipe if you like) that Hugo enjoys today. Note that all production QBD76 had the 18k filter, all this work happened at the prototype stage.
