Re: [Sursound] Noise reduction on Ambisonic files
Hi Garth, An interesting one. certainly got me thinking - trouble is, you don't really want thoughts but measurements. I suspect it depends a lot on what the internal mechanism of the noise reduction system is. Mostly, as far as I can ascertain, there's an analysis filter bank to split the sound into bands which are then subject to some sort of processing, then the bands are re-combined somehow either directly or by resynthesis to produce the output. The most critical thing will usually be the combination of the analysis and resynthesis steps. For instance, a well designed and well implemented FFT/iFFT pair should preserve the phase well. However, since you rarely have access to the internals of these things for analysis, measurement - or just listening with a good pair of ears - is the only way forward. I suspect that processing the B format after conversion from A would be the best - anyone else have any thoughts? Dave PS Of course, you could just always process the speaker feeds, for know, as that would be the least risky but most processing heavy option On 4 August 2014 20:23, Garth Paine ga...@activatedspace.com wrote: Hi everyone I have been doing a lot of ambient Ambisonic A format recordings (sps200 into SD788) and as the environmental levels are so low the self noise of the microphone becomes a bit of an issue on playback - I have RX for stereo noise reduction but have not found a solution for multichannel that would make me relaxed about maintaining the phase for decoding - I want to output B-Format so decoding onto any speaker array rather than just output 5.1 and use a surround noise cleaner. I would appreciate thoughts from the list - I am guessing as the Soundfield mics are know for self noise that others have faced and perhaps solved this issue already? thanks in advance ps. you can hear some of the recordings here http://listen.ame.asu.edu/sonic_events.php Cheers, Garth Paine ga...@activatedspace.com -- next part -- An HTML attachment was scrubbed... URL: https://mail.music.vt.edu/mailman/private/sursound/attachments/20140804/ca2c4e9f/attachment.html ___ 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. -- As of 1st October 2012, I have retired from the University. These are my own views and may or may not be shared by the University Dave Malham Honorary Fellow, Department of Music The University of York York YO10 5DD UK 'Ambisonics - Component Imaging for Audio' -- next part -- An HTML attachment was scrubbed... URL: https://mail.music.vt.edu/mailman/private/sursound/attachments/20140805/d6cb70e1/attachment.html ___ 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.
Re: [Sursound] Noise reduction on Ambisonic files
--On 05 August 2014 12:09 +0100 Dave Malham dave.mal...@york.ac.uk wrote: I suspect that processing the B format after conversion from A would be the best - anyone else have any thoughts? I have used RX on two-channel pairs of the B-format, with no obvious breaking of the reconstruction. Last time I had a noise problem, it was a single capsule, so I processed that channel of the A-format alone (using RX) before making the B-format, and the results seemed OK. If you want to process four channels together, Audition can be persuaded to open and rewrite a four-channel file correctly (even though it can't create one). Paul -- Paul Hodges ___ 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.
Re: [Sursound] Noise reduction on Ambisonic files
Hello Garth, Dave, I'd advise converting from B-format to A-format, then doing all de-noising, compression, etc, in A-format. Followed by, re-encoding back to B-format. This is how I've done all my work (both acousmatic involving field-recordings and musical location recordings). Depending on your de-noiser / gating settings, you can get image distortion effects... but these are similar to working with stereo. If you really want to be precious, instead of just 4 channels of A-format (for FOA), you can decode to more (say 6 channels as vertices of an octahedron, or 8 channels, in a cube), do your processing and then re-encode back to B-format. Also, you can adjust the polar patterns of your decode. Using 'controlled-opposites' decode can give you a smoother result, depending on your material. If you use a more custom version of A-format (octahedron or cube), make sure your re-encoding matrix is correctly scaled, so you get the correct balance of W vs X,Y,Z on re-encoding. Hope this helps! My best, Joseph Anderson j.ander...@ambisonictoolkit.net http://www.ambisonictoolkit.net On 5 Aug 2014, at 4:09 am, Dave Malham dave.mal...@york.ac.uk wrote: Hi Garth, An interesting one. certainly got me thinking - trouble is, you don't really want thoughts but measurements. I suspect it depends a lot on what the internal mechanism of the noise reduction system is. Mostly, as far as I can ascertain, there's an analysis filter bank to split the sound into bands which are then subject to some sort of processing, then the bands are re-combined somehow either directly or by resynthesis to produce the output. The most critical thing will usually be the combination of the analysis and resynthesis steps. For instance, a well designed and well implemented FFT/iFFT pair should preserve the phase well. However, since you rarely have access to the internals of these things for analysis, measurement - or just listening with a good pair of ears - is the only way forward. I suspect that processing the B format after conversion from A would be the best - anyone else have any thoughts? Dave PS Of course, you could just always process the speaker feeds, for know, as that would be the least risky but most processing heavy option On 4 August 2014 20:23, Garth Paine ga...@activatedspace.com wrote: Hi everyone I have been doing a lot of ambient Ambisonic A format recordings (sps200 into SD788) and as the environmental levels are so low the self noise of the microphone becomes a bit of an issue on playback - I have RX for stereo noise reduction but have not found a solution for multichannel that would make me relaxed about maintaining the phase for decoding - I want to output B-Format so decoding onto any speaker array rather than just output 5.1 and use a surround noise cleaner. I would appreciate thoughts from the list - I am guessing as the Soundfield mics are know for self noise that others have faced and perhaps solved this issue already? thanks in advance ps. you can hear some of the recordings here http://listen.ame.asu.edu/sonic_events.php Cheers, Garth Paine ga...@activatedspace.com -- next part -- An HTML attachment was scrubbed... URL: https://mail.music.vt.edu/mailman/private/sursound/attachments/20140804/ca2c4e9f/attachment.html ___ 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. -- As of 1st October 2012, I have retired from the University. These are my own views and may or may not be shared by the University Dave Malham Honorary Fellow, Department of Music The University of York York YO10 5DD UK 'Ambisonics - Component Imaging for Audio' -- next part -- An HTML attachment was scrubbed... URL: https://mail.music.vt.edu/mailman/private/sursound/attachments/20140805/d6cb70e1/attachment.html ___ 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. -- next part -- An HTML attachment was scrubbed... URL: https://mail.music.vt.edu/mailman/private/sursound/attachments/20140805/26eef386/attachment.html ___ 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.
Re: [Sursound] Noise reduction on Ambisonic files
Hi Dave Nice to hear from you and thanks for your input - it seems strange to me, given the known self noise of the most prevalent ambisonic microphones that there is not a solution out there already. Indeed your summations of the process aligns with mine, but I am somewhat nervous about looking to apply this over all 4 channels as there is so little across the 4 channels that I could use as a common measure of phase accuracy after processing, and to be honest I am not looking to write code for this as DSP is not my strong point - but I would see a use for this across the community. I would of course be happy to apply the noise reduction to the B-Format file. The idea for the Listen(n) project is to provide a wind range of listening outcomes from mobile devices with headphones to surround sound setups - so the decoding would need to be simple and be applicable across domestic platforms - so I am imagining that the noise reduction would therefore need to happen pre decoding to the listening format? Would love to find a solution - It has been suggested for instance that I use single instances of Izotopes RX on each of the 4 channels for the A-Format file and load a noise template in each - still I am concerned about any phase variation pre to decoding. Am I being over concerned? Cheers, Garth On Aug 5, 2014, at 4:09 AM, Dave Malham dave.mal...@york.ac.uk wrote: Hi Garth, An interesting one. certainly got me thinking - trouble is, you don't really want thoughts but measurements. I suspect it depends a lot on what the internal mechanism of the noise reduction system is. Mostly, as far as I can ascertain, there's an analysis filter bank to split the sound into bands which are then subject to some sort of processing, then the bands are re-combined somehow either directly or by resynthesis to produce the output. The most critical thing will usually be the combination of the analysis and resynthesis steps. For instance, a well designed and well implemented FFT/iFFT pair should preserve the phase well. However, since you rarely have access to the internals of these things for analysis, measurement - or just listening with a good pair of ears - is the only way forward. I suspect that processing the B format after conversion from A would be the best - anyone else have any thoughts? Dave PS Of course, you could just always process the speaker feeds, for know, as that would be the least risky but most processing heavy option On 4 August 2014 20:23, Garth Paine ga...@activatedspace.com wrote: Hi everyone I have been doing a lot of ambient Ambisonic A format recordings (sps200 into SD788) and as the environmental levels are so low the self noise of the microphone becomes a bit of an issue on playback - I have RX for stereo noise reduction but have not found a solution for multichannel that would make me relaxed about maintaining the phase for decoding - I want to output B-Format so decoding onto any speaker array rather than just output 5.1 and use a surround noise cleaner. I would appreciate thoughts from the list - I am guessing as the Soundfield mics are know for self noise that others have faced and perhaps solved this issue already? thanks in advance ps. you can hear some of the recordings here http://listen.ame.asu.edu/sonic_events.php Cheers, Garth Paine ga...@activatedspace.com -- next part -- An HTML attachment was scrubbed... URL: https://mail.music.vt.edu/mailman/private/sursound/attachments/20140804/ca2c4e9f/attachment.html ___ 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. -- As of 1st October 2012, I have retired from the University. These are my own views and may or may not be shared by the University Dave Malham Honorary Fellow, Department of Music The University of York York YO10 5DD UK 'Ambisonics - Component Imaging for Audio' -- next part -- An HTML attachment was scrubbed... URL: https://mail.music.vt.edu/mailman/private/sursound/attachments/20140805/d6cb70e1/attachment.html ___ 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. ___ 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.
Re: [Sursound] Noise reduction on Ambisonic files
At 20:56 05-08-14, Joseph Anderson wrote: I'd advise converting from B-format to A-format, then doing all de-noising, compression, etc, in A-format. Followed by, re-encoding back to B-format. What's the theory that predicts that the results will be any different than doing it on B-format, given that the transform is a linear matrix? David ___ 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.