rikkisbears wrote:Wow Rob,
you're increbibly game starting on a project like this. Can't imagine the amount of hours you must be putting into this.
I've been pottering away trying to put together an XG mapped Drum Kit for my PA800. Thanks to Sharp's mention of Sample Robot, I was able to sample the drums on my PSR. Very basic because I'm only using the demo version. There's a bit of editing involved cutting blank bits out of the wav samples. Better if I knew what I was actually doing, still, it's an interesting exercise haahaa
Hi Rikki,
All I did with the samples was to cut the leading silence out, and normalise the amplitude to 100%. Then I listened to the sample, and when the "tail" of the sample became almost inaudible, I cut the rest of the sample at that point. Then I applied the "envelope" function with a linear 45 degree slope at the tail to make sure the sound level reduced to zero. For low notes, I needed a 10 second sample, for high notes I needed a 4 second sample. I think cymbals may need 10 seconds, and drums may need 5 seconds or so. But you have to play it by ear.
The samples I used were all stereo .aif files. After manually topping and tailing them in this way (using WavePad) I then loaded all the samples into Awave Studio, and I assigned the note values (eg B1, F#2, A3 etc) to them by hand. I also used Awave's auto tuning function to work out whether the samples were sharp or flat, and I added tuning offsets. However the auto tuning function in Awave only got it right for 50% of the notes and it made wild errors for some samples which I had to fix by tuining by ear on the Pa2x later on.
You will have to make a similar decision for your percussion samples to decide which note plays which sample.
Then I used the auto split function in Awave to split the stereo samples into two mono-samples, Left and Right. It did this completely automatically. The Pa2x does not seem to load and play samples correctly if they are stereo, it seems to convert them to mono and forget about stereo. However if you load a pair of half samples L and R into adjacent OSC's and pan one left by 63 and the other one right by 63, then it works fine.
Then (still in Awave) I created two multisamples Steinway L and Steinway R, and I dragged and dropped all the corresponding samples. I filled in the dialog boxes that popped up to specify the note range over which each sample was spread. So my G#1 sample was spread from G1 to A1, eaach sample only spread by one semitone each side. In your case with percussion samples, you will have one sample per note and you won't have to spread them at all when you assign them to the multisample.
Then I exported each multisample as a separate Korg Trinity KSC script which also created a subfolder containing the KMP multisample and another folder containing all the KSF samples.
I copied these files to a USB stick, plugged it into the Pa2x and from Sound mode, I pressed RECORD. Then from the drop down menu I selected "Import" and I selected the KSC script file on my USB stick. It then imported everything completely automatically - the multisample and all samples. I then selecetd "Write" from the drop down menu. Then I imported the other KSC script based multisamples.
Then I previewed the sounds by playing them in sample edit mode. They However in sample record mode, the Pa2x/Pa800 seems to use a simplified "extreme low fi" playback engine. But you can at least hear whether the right samples are assigned to the right keys. And you can hear which samples are out of tune. I used an acoustic piano tuner (my ears) to listen to each sample, and I made offest tuning adjustments for each sample. Awave studio had got the tuning completely wrong for 50% of the notes, so it was a lot of work on the Pa2x to fix this. Some types of drums, triangle, cymbals, cowbell etc do have to be tuned in this way as well.
After doing the tuning I selected "write" to save the modified multisamples and I also used the Export function to save the modified .KSC, KMP and KSF files back to my USB device.
Then I exited from Sample edit mode and I started editing the Sound itself by assigning the corresponding samples to each OSC. In my case I used the existing Korg Grand Piano RX Sound as a template, and I recomend that you use an existing Drumkit for your samples.
Then eventually I was able to save my user Sound and samples into a new .SEt file and that is what I've just uploaded to Sharp's file share area.
I see a new PM has just arrived and I expect that's Sharp telling me that the files are ready to publish.
The problem I could foresee with Extreme Sample editor is that I don't think it can generate multisamples in Korg Triton/Trinity format, and with Awave Studio that was a massive time saver for me. Maybe I'm wrong about that though, so please llet me know?
Best regards,
Rob
