Ambisonics holds some wonderful nostalgic feelings and memories for me. It was during my studies at Guildford that I had the option to experiment, not only with stereo analog and digital recordings, but also with a four channel analog recorder with surround sound. In that time the theory of Prof Felgett and Michael Gurzon made a big impact on me about the possibilities to produce a much enhanced result when playing back recorded music(over the existing conventional stereo). At the university I was lucky to be able to experiment with a brand new Calrec soundfield microphone as well as an UHJ encoder. I chose to do my final project on the subject of Ambisonic recording and playback, including the possible marketing options.
Two things really stick till today: Surround recording and playback offer (potential and real) advantages over 2 channel stereo recordings, and if you offer a good playback to the general public, they can easily distinguish and favour the surround.
I did many test recording in the B format, using an omni directional microphone and 2 cross pair fig of 8 mikes, coincidentally placed, as well as recording with the soundfield mike that used 4 capsules, and can produce B format through the mixing of the output of the capsules.
The mathematical theory was quite impressive, but I am not at all convinced that the psycho-acoustical theory and explanation is very correct. After having had another opportunity to extensively experiment with surround sound recording in the early 1990's and then even more, during the development support time for SA-CD for Philips and Sony, experiencing the great advantage of the 5 channel ITU speaker set up, compared to the old fashioned 'square' quadraphonic speaker set up, and managing to dramatically improve on my former 'surround' microphone placements, it is really only the nostalgic feeling that I have left over. I would never consider using the technique for music recording any more, as I feel very lucky to have been able to do recordings in 5 channel surround, that can be realistically played back through a thoroughly tested and used consumer format, over the last 12 years.
I am writing this, just coming out of a finished recording period, recording Prokofiev violin sonatas, and having balanced the stereo and the surround to my best abilities, and thankfully being lucky enough to monitor the whole project in surround.
Now for the next week we are starting a 4 (sa)cd project with Tchaikovsky and Rimsky orchestral music and the prospect of hearing the orchestra back through the 5 loudspeakers is still, and again, very, very exiting.
By the way, I do think that if you consider the option of using many more speakers for playing back surround sound recordings, the wave field synthesis has a much greater viability, both theoretically as well as practically, than ambisonics has. It is with fond memories that I think back on ambisonics.