Clicks with SDIF files using CNMAT's [SDIF-tuples], [sinusoids~] and three fates
Hi, I am doing simple tests transferring SDIF files created in SPEAR into maxMSP using the above objects. I notice sharp clicks when a partial begins or ends, but only when there is more than one partial sounding. The click is a square cut-off, so I'm wondering if this in fact what is represented in the data, and the SPEAR DSP smooths it, or if it is a problem with one of the CNMAT externals (or the way I'm using them). I've attached a screen shot of a test file that plays smoothly in SPEAR and grittily in maxMSP, as well as the part of the patch I'm using to play it back. Any assistance would be gratefully received.
----------begin_max5_patcher---------- 695.3ocyWFzaaBCFF9L4WgEZmlRhv.FG1sIMMocdGmplLfIwSfcD1nk0ple6 yXf01DRBUE75EH1Xx2qe7q+9LOrvwMQbfJcAeB7CfiyCKbbLc0zgSWaG2Rxg zBhzLL27BAutzcY6iz+jwKnJyyfcclK3JI6dpoOu0dOMVQspev9c811k5O6o spnI.Dk6RfaBgu0EbW23XYlvKR90JTTe32STo6X7s+rhlpZee+nM5HBfvHys v1V5q+6epQdbRoIdtethQJbadviKVzbY4H4PIUJIao8JQQOXDfqNZ9qQ.ejm 2fPxe.HEeAFAuDiFDK3faikHCUPsrAM4PgS+sVImwjBFmB7VOHOBdE73hdlF 6xvLI71LItwhBBaYBdxYB89LR5wq5Ed4rz67oQT7UlFvfvFo6igl4Ap+J3to aE7iG0qe34yPKYa4Z7NzRXzlQL2CMqgsMlg0vKPEIiWKErL4wQmObRISzHHi mwiiP1kLpcUTZNQoqsLajofIUCyE7s4BDGa3Bbl3RpnrjxUmaYRqDEE.MeD0 a2AxHJBfwAkrB8zglJ3Yx2ds0gxfbshCA9sEGZMIQPKWxrhRxVIU5akqzSjD ZEPabTqkYrbfsqhFAuBnL.xXb5sQSeMz2CGvZDkah8M6e7rrWQUuufJAJVIE 7AHHUTTWxk.HvGDXYmRbvHRy.aO+oOxtm356e4aecUGqJSR9OjDFgFAcZOzE dicKNYfSRcdtNQyrBmKtAK9ZvYSfIYrY+EF+JQiYXlCbexW0YBTS+ujWRQcU Zud6OtE3onkoSDy3DESve9fzIHz645GzNVVFk+7JOkrr8BFW0IhKbTzoWSd1 SSQiQSarplZ1ycaMgsplZxRdSMchvmaMgGilZ9zQapovQnoSf46BMghr69N7 X73VVSvwr1YWONJdt8S5FOt3uHDJvpA -----------end_max5_patcher-----------
SDIF files created manually in SPEAR won’t play back very well in Max/MSP. I noticed this a number of years ago and reported it to the CNMAT team.
I assume that this is due to different handling of spectral data. Analysis files created from prerecorded sound files are not affected, interestingly.
If I were you, I’d create a sound file with two oscillators in Max/MSP and compare the data to those created manually in SPEAR. This might give you an idea of what goes wrong and maybe you can come up with a process to correct this.
Thankyou for the confirmation. However I did notice similar results when using prerecorded files that have had their spectral information edited in SPEAR, an analysis of a voice reduced down to a few partials, for instance.