Contents

Talk:Ambisonics: Difference between revisions

Revision as of 01:08, 28 December 2013

Professional sound production Start‑class Mid‑importance

This article is within the scope of WikiProject Professional sound production, a collaborative effort to improve the coverage of sound recording and reproduction on Wikipedia. If you would like to participate, please visit the project page, where you can join the discussion and see a list of open tasks.Professional sound productionWikipedia:WikiProject Professional sound productionTemplate:WikiProject Professional sound productionProfessional sound production articles Start This article has been rated as Start-class on Wikipedia's content assessment scale. Mid This article has been rated as Mid-importance on the project's importance scale.

This article is written in British English, which has its own spelling conventions (colour, travelled, centre, defence, artefact, analyse) and some terms that are used in it may be different or absent from other varieties of English. According to the relevant style guide, this should not be changed without broad consensus.

Until recently, the article included, "(in general, the more speakers, the higher the accuracy of the reconstructed soundfield)". This was removed by User:Nettings with the edit summary, 'removed "the more loudspeakers, the better". daniel has shown that rE suffers from too many loudspeakers.'

It is still true, however, that six speakers are better than four, and that eight are better than six. The problem Daniel's theoretical work exposed was only with very large numbers of speakers. The two points are not in conflict, so can we find a form of words that accommodates both? HairyWombat 15:17, 16 December 2013 (UTC)[reply]

It's not exactly a theoretical problem. And eight speakers for first-order horizontal is already stretching it - I doubt you will find them preferrable to six, really. The more speakers, the higher the accuracy (for a given order) is definitely a myth. Besides, this whole article includes waaay too much information that is not really relevant to an introductory article in an encyclopaedia. Nettings (talk) 15:51, 16 December 2013 (UTC)[reply]

Here is a very good elaboration on the problem with too many speakers: Solvang, Audun, "Spectral Impairment for Two-Dimensional Higher Order Ambisonics", JAES Vol.56/4, 2008. If you cannot access this, I can send you an excerpt by mail if you're interested, just get in touch. Nettings (talk) 09:22, 20 December 2013 (UTC)[reply]

Six speakers are better than four. This has never been questioned. I have not read this JAES paper, but the results from listening tests are mixed even with many speakers. HairyWombat 17:31, 20 December 2013 (UTC)[reply]

Time permitting, I will add some elaboration on this inspired by Wiggins, who made a very nice rE plot for the corner case of three speakers (where horrible speaker detention is evident), and shows the improvement in the rE plot with four and six speakers. So at the low end of the scale, "some more speakers than the absolute minimum == better localisation". But the generalisation "more == better" is demonstrably wrong, see Daniel and Solvang. As the speaker number increases more, spectral issues and phasing will dominate the perception, so that there is a subjective loss in quality for ${\displaystyle N_{speakers}\gg 2\ell +1}$. From my own experience, I'd be vary of subjective listening reports based on just a single system experience, because reverberant room acoustics tend to mask phasing issues that are very evident (and obnoxious) in dry acoustics. Whereas the objections from Solvang and Daniel are quantifiable. - Btw, thanks for your helpful hints on proper archival and indentation. If you find other places where my edits do not adhere to WP best practises, I'd love to learn about them.Nettings (talk) 11:17, 21 December 2013 (UTC)[reply]

Hi, I just added a statement to the initial section stating renewed interest in Ambisonics from research institutions and media companies. Before I add a looong list of recent papers from said parties to support this claim, does anyone know of a single source that will corroborate this fact more concisely? Nettings (talk) 16:59, 16 December 2013 (UTC)[reply]

Hi *! I have taken the liberty of ruthlessly editing the article, to transform it into a first-class wikipedia citizen eventually. I realize that I've removed content that contains valid and valuable information, and which represents long and hard work by other people. The content is not lost, merely commented out, and can be cherry-picked at leisure. (Besides, Wikipedia does not lose anything and you can always revert my changes if you disagree). However, I felt that the previous state of the article required a heavy-handed clean-up effort. I hope we can arrive at a sub-page structure that leaves room for the interesting and diverse historical information on ambisonics while keeping the main article short and to the point, while transforming the anecdotal style to something that's more in keeping with encyclopaedic writing. (I'm picking this up as I go, so feel free to shorten my contributions even more!) I realize that there is currently too much "original research" (aka. stuff everyone knows) which needs backing up by citations, and that we are totally missing information on psycho-acoustic criteria and optimisations vs. physically correct sound field reconstruction. Time permitting, I will try to add some missing bits in the near future, and I hope that other Ambisonics enthusiasts will join in. Nettings (talk) 22:12, 16 December 2013 (UTC)[reply]

A horizontal parametric microphone aiming in direction ${\displaystyle \Theta }$ with pattern ${\displaystyle 0\leq p\leq 1}$, where 0 is a figure-of-eight, 0.5 is a cardioid and 1 an omni:

${\displaystyle M(\Theta ,p)=p{\sqrt {2}}W+(1-p)(\cos \Theta X+\sin \Theta Y)}$,

free-field normalised, i.e. constant gain on-axis, hence linear interpolation between W and (X,Y).

I'm working on a gnuplot animation to show a couple of mics morphing into one another, showing the current gains for WXY for each pattern.--Nettings (talk) 01:08, 28 December 2013 (UTC)[reply]