Hi Sampo, There was a bit silence on this thread, so I thought I would briefly chime in.
I’ll note that DirAC has indeed already been extended to arbitrary ambisonic orders in [1]. There’s also an alternative decoding method described in [2], which instead adopts a different sound-field model and rendering strategy (but also supports arbitrary order as input). Both methods operate in the spherical harmonic domain, much like signal-independent HOA decoders do. I would say that these solutions are also both rather scalable, since higher input orders allows for more simultaneous sources/reflections to be resolved and spatialized appropriately within the same time-frequency index. Higher orders also allow for better modelling and reproduction of the directional energy distribution of the captured ambience. Regarding the mixed order query, I didn’t fully follow, as I was under the impression that mixed order Ambisonics referred more to the act of oversampling the sphere on the horizontal; i.e., placing more microphones/loudspeakers on the equator (aiming for, say, 5th order resolution on the horizontal plane), and fewer microphones/loudspeakers elsewhere (aiming for maybe 3rd order resolution for elevated sources). One might do this for both practical and perceptual reasons. On the other hand, mixed order rendering may also refer to the act of taking into account the frequency-dependent limitations of the microphone array being used. For example, while the Eigenmike is a 4th order array - it will only provide these 4th order components between around 2.5-5kHz (above which we run into spatial aliasing, and below which, we have too much noise). Therefore, in theory, one should conduct the decoding using all 25 channels only within this bandwidth. The dipoles produced by the Eigenmike, on the hand, are probably good up to around 8kHz, in which case these standard/active decoding methods can revert to their first-order forms to render these other frequencies etc…. Anyway, if you’re curious, I’ll note that I have been maintaining free VST implementations of both [1] and [2], if you would like to listen to these rendering methods using your own HOA material. More information regarding recent developments of these active Ambisonics decoding (and encoding) method may also be found here: https://leomccormack.github.io/sparta-site/ All the best, Leo [1] Politis, A., Vilkamo, J. and Pulkki, V., 2015. Sector-based parametric sound field reproduction in the spherical harmonic domain.<https://leomccormack.github.io/sparta-site/docs/help/related-publications/politis2015sector.pdf> IEEE Journal of Selected Topics in Signal Processing, 9(5), pp.852-866. [2] Politis, A., Tervo, S. and Pulkki, V., 2018, April. COMPASS: Coding and multidirectional parameterization of ambisonic sound scenes.<https://leomccormack.github.io/sparta-site/docs/help/related-publications/politis2018compass.pdf> In 2018 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) (pp. 6802-6806). IEEE. P.S. HUT became Aalto in 2012 (and also the email extensions). From: Sursound <sursound-boun...@music.vt.edu> on behalf of Sampo Syreeni <de...@iki.fi> Date: Thursday, 11. May 2023 at 5.00 To: sursound-list <sursound@music.vt.edu> Cc: Ville Pulkki <ville.pul...@hut.fi> Subject: [Sursound] two ideas, with questions of active decoding For as long as ambisonics has been around, there were also ideas of active decoding. Even "infinite order decoding", actively. There's been Harpex for two specular sources, derived from the theory leading upto Dolby Pro, and then there's been DirAC, which additionally tries to separate the reverberant field from the specular sources, which it then at "infinite order" renders onto a rig via VBAP. So it's nice in all and might work, but I'd say this work is not too principled or generalizable. I mean, none of this theory or technique actually seems to take the ambisonic framework seriously. For example, none of it generalizes to arbitrary order, or works with the harmonical functions. None of it really works with spherical harmonics as HOA of does. For some reason this kind of analysis leaves me...lacking. So tell me, researchers, professors on-list, and your post-docs, might you maybe delve into a couple of my amateur ideas and questions? 1) Express any active decoding solution in an infinite series of spherical harmonics, utilizing only them. Becaause this kind of solution would necessarily solve the active solution problem for any distribution of sources, instead of just one, two, or in periphony even, four. 2) The mixed order problem. Years back I thought it'd be optimal to just sum first and higher order order ambisonic signal sets together, and decode them en masse, using a single decoder. I was wrong: the optimal decoding for first and second order sets is different. So, since people obviously mix first and higher order signal sets, the optimal decoder needs to separate the contribution of each successive order from each other, in order to optimally decode them. All of the orders, overlapping slightly, to a degree. This all can be systematically implemented via some variant of non-negative matrix factorization, or a couple of related methods. I believe. Please try it out, or make someone try out. -- Sampo Syreeni, aka decoy - de...@iki.fi, https://eur01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fdecoy.iki.fi%2Ffront&data=05%7C01%7Cleo.mccormack%40aalto.fi%7Cbc9c2bc7479e4ddfaee408db51d43fb5%7Cae1a772440414462a6dc538cb199707e%7C1%7C0%7C638193744268420244%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=A8%2BF3AgmAQZEXfb0Un8zbfbCPVi5PerwyxlIDldji8k%3D&reserved=0<http://decoy.iki.fi/front> +358-40-3751464, 025E D175 ABE5 027C 9494 EEB0 E090 8BA9 0509 85C2 _______________________________________________ Sursound mailing list Sursound@music.vt.edu https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fmail.music.vt.edu%2Fmailman%2Flistinfo%2Fsursound&data=05%7C01%7Cleo.mccormack%40aalto.fi%7Cbc9c2bc7479e4ddfaee408db51d43fb5%7Cae1a772440414462a6dc538cb199707e%7C1%7C0%7C638193744268420244%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=cJz6PuDWv0TCP9NKBjHGJN8U%2FpWIXq0CnfZp0t6m4Zs%3D&reserved=0<https://mail.music.vt.edu/mailman/listinfo/sursound> - unsubscribe here, edit account or options, view archives and so on. -------------- next part -------------- An HTML attachment was scrubbed... URL: <https://mail.music.vt.edu/mailman/private/sursound/attachments/20230516/e329112f/attachment.htm> _______________________________________________ Sursound mailing list Sursound@music.vt.edu https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit account or options, view archives and so on.
