Reducing smearing and other artefacts with FFT

Reducing smearing and other artefacts with FFT

Date	2015-05-15 13:17
From	"F. Silvain"
Subject	Reducing smearing and other artefacts with FFT
	Hey hey everyone, I wonder, what techniques there are in Csound to further reduce smearing and other artefacts of FFT analysis? I'm mainly interested in mainpulating frequencies not time stretching. I am using pvslock and balancing the resynthesized audio signal with a delayed version of the input. That does help. But especially voices sound as if there was a short delay or chorus effect on them. I've read that up-to-date commercial software uses multiple techniques and wonder, if there are some clever thoughts around here. Thank you! Ta-ta ---- Ffanci * Homepage: https://freeshell.de/~silvain * Twitter: http://twitter.com/ffanci_silvain * GitHub: https://github.com/fsilvain ------------------------------------------------------------------------------ One dashboard for servers and applications across Physical-Virtual-Cloud Widest out-of-the-box monitoring support with 50+ applications Performance metrics, stats and reports that give you Actionable Insights Deep dive visibility with transaction tracing using APM Insight. http://ad.doubleclick.net/ddm/clk/290420510;117567292;y _______________________________________________ Csound-users mailing list Csound-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/csound-users Send bugs reports to https://github.com/csound/csound/issues Discussions of bugs and features can be posted here

Date	2015-05-15 13:36
From	Victor Lazzarini
Subject	Re: Reducing smearing and other artefacts with FFT
	pvslock is very experimental and it does not really work well yet. The best solution would be to use either temposcal or mincer, which do phase locking. Not 100% better but it helps. ======================== Dr Victor Lazzarini Dean of Arts, Celtic Studies and Philosophy, Maynooth University, Maynooth, Co Kildare, Ireland Tel: 00 353 7086936 Fax: 00 353 1 7086952 > On 15 May 2015, at 13:17, F. Silvain wrote: > > Hey hey everyone, > I wonder, what techniques there are in Csound to further reduce smearing and > other artefacts of FFT analysis? > > I'm mainly interested in mainpulating frequencies not time stretching. > > I am using pvslock and balancing the resynthesized audio signal with a delayed > version of the input. That does help. But especially voices sound as if there > was a short delay or chorus effect on them. I've read that up-to-date > commercial software uses multiple techniques and wonder, if there are some > clever thoughts around here. > > Thank you! > > Ta-ta > ---- > Ffanci > * Homepage: https://freeshell.de/~silvain > * Twitter: http://twitter.com/ffanci_silvain > * GitHub: https://github.com/fsilvain > > ------------------------------------------------------------------------------ > One dashboard for servers and applications across Physical-Virtual-Cloud > Widest out-of-the-box monitoring support with 50+ applications > Performance metrics, stats and reports that give you Actionable Insights > Deep dive visibility with transaction tracing using APM Insight. > http://ad.doubleclick.net/ddm/clk/290420510;117567292;y > _______________________________________________ > Csound-users mailing list > Csound-users@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/csound-users > Send bugs reports to > https://github.com/csound/csound/issues > Discussions of bugs and features can be posted here ------------------------------------------------------------------------------ One dashboard for servers and applications across Physical-Virtual-Cloud Widest out-of-the-box monitoring support with 50+ applications Performance metrics, stats and reports that give you Actionable Insights Deep dive visibility with transaction tracing using APM Insight. http://ad.doubleclick.net/ddm/clk/290420510;117567292;y _______________________________________________ Csound-users mailing list Csound-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/csound-users Send bugs reports to https://github.com/csound/csound/issues Discussions of bugs and features can be posted here

Date	2015-05-15 19:42
From	David Mooney
Subject	Re: Reducing smearing and other artefacts with FFT
Attachments	None None
	There's always a trade off with FFT between accurate frequency and accurate timing. Play around with the window size. --David Mooney On Fri, May 15, 2015 at 8:17 AM, F. Silvain <silvain@freeshell.de> wrote: Hey hey everyone, I wonder, what techniques there are in Csound to further reduce smearing and other artefacts of FFT analysis? I'm mainly interested in mainpulating frequencies not time stretching. I am using pvslock and balancing the resynthesized audio signal with a delayed version of the input. That does help. But especially voices sound as if there was a short delay or chorus effect on them. I've read that up-to-date commercial software uses multiple techniques and wonder, if there are some clever thoughts around here. Thank you! Ta-ta ---- Ffanci * Homepage: https://freeshell.de/~silvain * Twitter: http://twitter.com/ffanci_silvain * GitHub: https://github.com/fsilvain ------------------------------------------------------------------------------ One dashboard for servers and applications across Physical-Virtual-Cloud Widest out-of-the-box monitoring support with 50+ applications Performance metrics, stats and reports that give you Actionable Insights Deep dive visibility with transaction tracing using APM Insight. http://ad.doubleclick.net/ddm/clk/290420510;117567292;y _______________________________________________ Csound-users mailing list Csound-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/csound-users Send bugs reports to https://github.com/csound/csound/issues Discussions of bugs and features can be posted here -- Opaque Melodies http://opaquemelodies.com Works in sound and fiber

Date	2015-05-15 20:42
From	"F. Silvain"
Subject	Re: Reducing smearing and other artefacts with FFT
	David Mooney, May 15 2015: > There's always a trade off with FFT between accurate frequency and accurate > timing. Play around with the window size. I already did and found the best possible values for my current tasks. But compared to dedicated applications performing the same task, there's still a way to go. I'm especially looking at pitch correction (autotune style). ... Ta-ta ---- Ffanci * Homepage: https://freeshell.de/~silvain * Twitter: http://twitter.com/ffanci_silvain * GitHub: https://github.com/fsilvain ------------------------------------------------------------------------------ One dashboard for servers and applications across Physical-Virtual-Cloud Widest out-of-the-box monitoring support with 50+ applications Performance metrics, stats and reports that give you Actionable Insights Deep dive visibility with transaction tracing using APM Insight. http://ad.doubleclick.net/ddm/clk/290420510;117567292;y _______________________________________________ Csound-users mailing list Csound-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/csound-users Send bugs reports to https://github.com/csound/csound/issues Discussions of bugs and features can be posted here

Date	2015-05-16 07:59
From	"F. Silvain"
Subject	Re: Reducing smearing and other artefacts with FFT
	Victor Lazzarini, May 15 2015: > pvslock is very experimental and it does not really work well yet. The best solution would be to > use either temposcal or mincer, which do phase locking. Not 100% better but it helps. Mincer's phase locking sounds MUCH better, but its pitch shifting doesn't sound too good. Is the phase locking implemented in mincer so specific, that it can't be incorporated in pvslock? Certainly both use FFT analysis. ... Ta-ta ---- Ffanci * Homepage: https://freeshell.de/~silvain * Twitter: http://twitter.com/ffanci_silvain * GitHub: https://github.com/fsilvain ------------------------------------------------------------------------------ One dashboard for servers and applications across Physical-Virtual-Cloud Widest out-of-the-box monitoring support with 50+ applications Performance metrics, stats and reports that give you Actionable Insights Deep dive visibility with transaction tracing using APM Insight. http://ad.doubleclick.net/ddm/clk/290420510;117567292;y _______________________________________________ Csound-users mailing list Csound-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/csound-users Send bugs reports to https://github.com/csound/csound/issues Discussions of bugs and features can be posted here

Date	2015-05-16 09:44
From	Victor Lazzarini
Subject	Re: Reducing smearing and other artefacts with FFT
	Not it can’t. The reason is that once we do PV analysis, we loose the phase information, and that is what mincer does. Unfortunately mincer does not do formant preservation, and that is where pvscale wins. Adding formant preservation to mincer is actually possible (it just depends on my having time to code it) You can possibly combine the two. In my example, I use mincer to do timescale modification and then pvsanal/pvscale/pvsynth for pitch modification. More CPU cycles, but nothing absurd. regards ======================== Dr Victor Lazzarini Dean of Arts, Celtic Studies and Philosophy, Maynooth University, Maynooth, Co Kildare, Ireland Tel: 00 353 7086936 Fax: 00 353 1 7086952 > On 16 May 2015, at 07:59, F. Silvain wrote: > > Victor Lazzarini, May 15 2015: > >> pvslock is very experimental and it does not really work well yet. The best solution would be to >> use either temposcal or mincer, which do phase locking. Not 100% better but it helps. > > Mincer's phase locking sounds MUCH better, but its pitch shifting doesn't sound too good. Is the phase locking implemented in mincer so specific, that it can't be incorporated in pvslock? Certainly both use FFT analysis. > ... > > Ta-ta > ---- > Ffanci > * Homepage: https://freeshell.de/~silvain > * Twitter: http://twitter.com/ffanci_silvain > * GitHub: https://github.com/fsilvain > > ------------------------------------------------------------------------------ > One dashboard for servers and applications across Physical-Virtual-Cloud > Widest out-of-the-box monitoring support with 50+ applications > Performance metrics, stats and reports that give you Actionable Insights > Deep dive visibility with transaction tracing using APM Insight. > http://ad.doubleclick.net/ddm/clk/290420510;117567292;y > _______________________________________________ > Csound-users mailing list > Csound-users@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/csound-users > Send bugs reports to > https://github.com/csound/csound/issues > Discussions of bugs and features can be posted here ------------------------------------------------------------------------------ One dashboard for servers and applications across Physical-Virtual-Cloud Widest out-of-the-box monitoring support with 50+ applications Performance metrics, stats and reports that give you Actionable Insights Deep dive visibility with transaction tracing using APM Insight. http://ad.doubleclick.net/ddm/clk/290420510;117567292;y _______________________________________________ Csound-users mailing list Csound-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/csound-users Send bugs reports to https://github.com/csound/csound/issues Discussions of bugs and features can be posted here

Date	2015-05-16 11:29
From	"F. Silvain"
Subject	Re: Reducing smearing and other artefacts with FFT
	Victor Lazzarini, May 16 2015: ... > You can possibly combine the two. In my example, I use mincer to do timescale modification > and then pvsanal/pvscale/pvsynth for pitch modification. More CPU cycles, but nothing absurd. Victor, I mostly need frequency processing, not only pitch shifting, though that is a VERY good start. Thank you for enlightening me on these points. ... Ta-ta ---- Ffanci * Homepage: https://freeshell.de/~silvain * Twitter: http://twitter.com/ffanci_silvain * GitHub: https://github.com/fsilvain ------------------------------------------------------------------------------ One dashboard for servers and applications across Physical-Virtual-Cloud Widest out-of-the-box monitoring support with 50+ applications Performance metrics, stats and reports that give you Actionable Insights Deep dive visibility with transaction tracing using APM Insight. http://ad.doubleclick.net/ddm/clk/290420510;117567292;y _______________________________________________ Csound-users mailing list Csound-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/csound-users Send bugs reports to https://github.com/csound/csound/issues Discussions of bugs and features can be posted here