Well, I've made some progress but I am definitely seeing some very inconsitent behaviour from ImpulePrep.
First of all I did a few things to improve my signal to noise ratio: 1. I set Audacity to use the 'Stereo Mix' rather than 'Line' input of my SB Live! (internal) card. This seems to give a higher gain, though not sure if would just be amplifying any noise too. Thoughts? 2. I set my RSSPL to use the 90 scale (previously was on the 80 scale). Just recording silence you can see a lower background noise level. 3. I turned the gain on my sub down 2db (this whole exercise seems to be suggesting my sub is too loud anyway), which has enabled me to jack the volume of the sweep up (so compensating for the use of the 90 scale on the SPL. After making many recordings all of which appeared good in Audacity, and all with the same basic settings (ie gain in Audacity, SPL setting, volume and sub gain), the resulting Impulses seem to vary widely. Sometimes the left impulse looks good (low noise floor), but the right impulse bad (high noise), and sometimes the other way around. I did manage to get a good L/R pair, but only by analysing every resulting impulse in REW and then selecting the best L and best R recording I had made. One thing I noticed was that I got rubbish results if I trim too much off the beginning and end of the sweep recordings. I'm using a Meridian controller and Meridian DSP (ie digital) speakers. I know that the speakers introduce at least a 7.5ms delay, and probably more. Looking at the trace in Audacity there does seem to be a short lag between the sweep and mic inputs. Perhaps by trimming the start and end too short I confuse things. It would be useful if there was a longer period of silence between sweeps (say double). I have attached a good L impulse and bad R impulse from REW. I have the measurement files too if they would help analyse further. What is the differenct between Impulse_Response_Measured and Filtered, ie what is being filtered? When I have a bad impulse response the measured trace in Audacity has noise before the spike and after, but the filtered trace has less noise after the spike (we are talking about 0.05 on the Audacity Y scale). +-------------------------------------------------------------------+ |Filename: r.jpg | |Download: http://forums.slimdevices.com/attachment.php?attachmentid=3293| +-------------------------------------------------------------------+ -- mortod ------------------------------------------------------------------------ mortod's Profile: http://forums.slimdevices.com/member.php?userid=4651 View this thread: http://forums.slimdevices.com/showthread.php?t=26005 _______________________________________________ plugins mailing list [email protected] http://lists.slimdevices.com/lists/listinfo/plugins
