Head-Fi.org › Forums › Misc.-Category Forums › Members' Lounge (General Discussion) › Ambisonics - We had the Holy Grail of audio, and we lost it.
New Posts  All Forums:Forum Nav:

Ambisonics - We had the Holy Grail of audio, and we lost it.

post #1 of 5
Thread Starter 
(Mods: this would look to be part of the tech forum, but I don't really care about discussion on the technical part of it. I would like to stimulate some ambisonics discussion, while providing some tech background. Still, if you feel this doesn't belong here, feel free to move it!)

OK, not quite, but close.

<technobabble>

What is Ambisonics, you say? Well, a bunch of scientists and researchers in the UK discovered that you could describe a 3D soundfield with 4 audio feeds. One would describe the overall intensity, the rest of them would describe the difference between the overall intensity and the axis of reference.

Then you could use a decoder, which would take the feeds, do some math operations on them (mainly they separated the axis components and applied some geometry equations, depending on the spatial position of the speakers), and send the outputs to an amplifier.

Think about it. You could use any amount of speakers in any configuration you wanted (say, 7 speakers around you, 4 on top of you, 3 below you), and Ambisonics would re-create the original sound field. Sounds a lot more advanced than Dolby Digital or DTS, doesn't it?

Well, I thought the same thing you are thinking now: Horsey-poo-poo! So I decided to try out for myself.

I have 2 Ambisonics UHJ-encoded CDs (UHJ means the signal was encoded in two stereo channels -B format means the signal is presented in a WXYZ format). I decided on decoding "Stereotomy" by The Alan Parsons Project. I used some impulse responses (which i found at www.ambisonics.net), convoluted the stereo tracks into mono B-format signals, and then used a software Ambisonics decoder to extract 5 channels of audio, specifically tailored to my home rig (which means, I got the spatial coordinates of my mains and surrounds, and fed the decoder with this information). And then I encoded the resulting 5 channel signal into Dolby Digital.

</end of technobabble>

The results? Oh... My... GOD. I don't feel stating that the resulting disc kicks DTS behind is overstating it. The soundfield was simply STUNNING. It wasn't as if "instruments sounded as if they came from the rear speakers". It was as if "hey, that guitar is *right there*... right there between the front left and the rear left speaker... and that cymbal is over there... to the right of the right rear speaker... and that echo came over here... right in the middle of the two back speakers..."

Incredibly impressive. More so when you consider the fact that this is 21 year old technology, being conveyed in a 21 year old medium (the CD) without special coding, or whatever.

It's interesting to notice, though, that Ambisonics was never a hit because of way too many factors that don't have anything to do with technology. I wonder how many great things we are missing now because of political issues, bad marketing plans, greed, etc.

Well, I have to go now. I will now decode my "The Final Cut" Pink Floyd CD, which is also in Ambisonics. YAY!
post #2 of 5
If I may express a contrary opinion, I think the unfortunate failing of Ambisonics is that we do not have point-sized heads. Ambi records all there is to record about a soundfield at exactly one point. However, humans have finite-sized heads and shoulders that we use to sample the soundfield, so we need interaural time differences because of the time-of-arrival differences at our ears as well as the interaural amplitude differences Ambi provides. Ambi attempts to synthesize ITD cues, but it's not as good as recording it in the first place.

Ambi is very impressive for its mathematical foundations and tools that make it so scalable and expandible. It's definitely far, far ahead of its time in that respect.

Have you read about AT&T's Perceptual Soundfield Reconstruction system? Look at:

http://www.research.att.com/projects/psr/

--Andre
post #3 of 5
Thread Starter 
Quote:
Originally posted by AndreYew
If I may express a contrary opinion, I think the unfortunate failing of Ambisonics is that we do not have point-sized heads. Ambi records all there is to record about a soundfield at exactly one point.
Hmm. I would have assumed they had considered that point (as binaural recordings have done with dummy torsos and heads), but you are right, the Ambisonics microphone is a simple WXYZ one and it doesn't consider the reflections your body might create.

Still, I am amazed at the quality of the sound. All the DTS recordings I own usually sound pretty "discrete"; that is, sound seems to be coming "from the speakers". Ambisonics is a lot more natural sounding.

Maybe the only surround sound CD that can compare (which I own, of course) is John Ergle's recording of the 1812 Overture in Dolby Digital.

I guess it is a step in the right direction, and that a lot of systems are much better than it. Still, I think it just might be the most advanced surround system we as consumers had access to... and my point is, how come we lose the good stuff so easily?
post #4 of 5
i have no idea why i never bothered researching this before. my brother went to europe when he was in college and did some research for ambisonics. i'll be sure to ask him about it next time i see him, but til then anyone got some links i could read up on?
post #5 of 5
Thread Starter 
Quote:
Originally posted by skippy
i have no idea why i never bothered researching this before. my brother went to europe when he was in college and did some research for ambisonics. i'll be sure to ask him about it next time i see him, but til then anyone got some links i could read up on?
www.ambisonic.net has everything you could ever want to know about Ambisonics. I learned how to decode my recordings there.
New Posts  All Forums:Forum Nav:
  Return Home
Head-Fi.org › Forums › Misc.-Category Forums › Members' Lounge (General Discussion) › Ambisonics - We had the Holy Grail of audio, and we lost it.