Here are some useful ideas and techniques that I use in my MIDI productions:
#1) Turning Mono Samples into Stereo Samples: You can turn any mono sample into a stereo sample by using the SP's Stereo Merge function. Let's say you have an interesting, evolving, Wavestation-type sound ( in mono) and you want a wide, stereo version of it. Copy that wave to a different wave location, then audition the two waves on the Dual Wave audition page. With identical mono waves, the resulting stereo wave will also be mono. Now go to the left wave's Wave Start/End page and start advancing the wave's start point. The farther you advance the start point of the wave, the wider the stereo image becomes. You can use small amounts for a narrow, chorused effect or longer times for a widely separated sound. When you have the optimum amount of width, use the Stereo Merge function to combine the two Mono waves into a Stereo Wave. The SP takes the master start time and loop points of the new Stereo wave from the left mono wave's settings.
#2) Part Doubling and the Haaus Effect: I often double electric rhythm guitar tracks, fingerpicked or arppegiated acoustic guitar parts, or tabla & hand drum loops to give the effect of two players playing almost exactly, but not quite together. If it's a guitar type part, I'll record and edit the entire track first, then copy it to a second track, assign a slightly different sound to the 2nd track, then pan the two tracks hard left and hard right. With MIDI, if two parts are playing exactly together, they will sound like one part. This is a good technique to use if you want a wide, upfront, solo type sound. If you want to emulate the sound of two musicians playing almost together, try delaying one track ( I usually delay the right side) by a few clicks. My sequencer ( an old but trusty Yamaha QX-3) has a resolution of quarter note=96, so depending on the tempo, I will delay the right track from 2-6 clocks. When you do this, you will notice that the delayed track seems quieter than the "on time" track. This is called the Haaus effect, named after the German scientist who studied this phenomenon. The sound that arrives at your ears first will slightly mask the second sound, so the ear perceives the first sound as being louder. To compensate for this, I usually give the delayed track a slightly higher MIDI Volume (controller #7) amount than the "on time" track. If the "on time" track is at Vel. 70, then the delayed track will need to be at Vel. 80 to have equal volume. For doubling tabla or hand drum parts: Let's say you have 6 loops you are using for your tune. Arrange the loops on a Map on 6 adjacent keys. Go through the tune and sequence the loops in the order they are to be played. Next, pan all those loops hard left, and add to your Map the same 6 loops, in the same order, on the next 6 highest keys; panned hard right. Copy the hand drum MIDI track to another track, transpose it up 6 semitones, and now you should hear both left and right loops playing together. Delay the right track a few clocks to slightly open up the spacing between the parts, increase the right hand track velocity by around 10 to balance the volumes and voila' : instant part doubling!
#3) Using Chorus and Flanging on Basses: Many Electric and Synth Bass sounds can be a little dull and lifeless in their original form, so adding some stereo chorus or flanging can bring them back to life. I generally use Chorus on Electric bass, and Flanging on Synth Bass. Stereo Chorus on an Electric Bass can slightly fatten up the sound, give it a wider stereo spread, and by subtly changing over time, make the sound more interesting. Using Stereo Flanging on Synth Bass will give an even more noticeable effect; making the sound " harder", more upfront, and slowly evolving/less static. I usually use fairly slow speeds, with not too much depth on Electric Bass, and more depth/pitch change on Synth Bass.
#4) Musical Dynamics and Quantization: Probably no MIDI-related subject has been discussed more than the detrimental effects or overuse of quantization. Here's my take:
I'm a drummer (please-no jokes!), and therefore rhythm is something I'm aquatinted with. I personally like quantization-I like hearing everything hit together cleanly and rhythms played accurately, so I quantize everything I program. This doesn't mean that rhythm tracks have to sound sterile and machinelike. Actually, I've had more comments on how much my tracks groove than otherwise. The key is velocity: Parts that are dynamic and that breathe in a musical way can groove hard, even when fully quantized. If I'm working on a Pop-Rock type tune, I will have the 2 & 4 hits on the snare at a constant velocity throughout the tune; maybe vl.=110 or 120. Any in-between snare notes will of course be at a lower velocity; around vl. 50-70. The same goes for bass drum.The real key is hi-hat: I do all my programming from a keyboard, so when I program hi-hat (say 16th notes) I do it in real time, and accent the quarters more than the other notes. Once in a while, I might throw in an accent on the "e" or "a". I never compress the hi-hat's velocity once its' recorded, as I want to keep its' full dynamic contrasts intact. I might add or subtract velocity to the hi-hat track as a whole, but I never make all the hi-hat notes the same velocity.
Another approach I use is this: Many songs I program have a slightly swung or shuffled feel. Sometimes (depending on tempo) quantizing to 8th or 16th triplets is too "perfect" sounding, and is not how a real drummer would naturally play it. If it's a 16th triplet-type groove, I'll first quantize to straight 16th notes, then extract all the "E's" and "A's" to another track. I'll then delay this track to a value halfway in between 16ths and 16th triplets. For example, my sequencer has 96 ppq. resolution. 16th notes are 24 clocks, and 16 triplets are 16 clocks. The difference is 8 clocks. I'll shift the "E's and A's" track back 4 clocks (half the difference), and the groove will take on a slippery, "New Orleans" type feel that is neither straight nor fully shuffled. The amount of clocks shifted will depend on the tempo-more clocks for slower tempos, less for faster tempos. It's best to experiment with the amount of clocks shifted to see what feels best.
#5) Calculating Sample Loop Times: In a perfect world, all sampled loops would be an exact multiple of sampling rate X tempo. Assuming you are sampling at 44.1k, a 1 second drumloop would be exactly 44,100 samples long. To calculate this, divide the number 60,000 by the tempo of the song, then multiply by the number of beats in the loop (generally 4 for a one bar loop, or 8 for a two bar loop), and multiply the result by the sampling rate; 44.1. For a one bar loop at 120 bpm, it would be: 60,000/120 = 500 x 4 (beats) = 2000 (milliseconds, or 2 seconds) x 44.1= 88,200 samples. Lets say you have a 2 bar loop at 144 bpm, sampled at 22.05k. 60,000/144 = 416.66 x 8 = 3333.33 x 22.05 = 73,500 samples. Of course, slight inaccuracies in the performance of the loop or the timing of the sequencer may necessitate tweaking the loop length via fine tuning the Pitch.
#6) Using the SP's Wave Quantization For Cross-switching Samples: Here's a tip sent to me from Brent McDonald on an interesting use of the SP's Wave Quantization feature.
"Many years ago, I got fed up with trying to use what I was given in drum modules and keyboards to sequence convincing drum tracks. It wasn't that the samples weren't good but rather: play a key soft and you get the same sample only softer (or the reverse). Of course, any drum will sound completely different depending on how hard it's struck. Same for high hat, only that's worse to get around as when a drummer plays, it is always a combination of how hard it's hit as well as what the drummer is doing with his foot. Enter this crazy wave quantization function on the SP, which after I stumbled across it, not only breathed new life into my SP but also my sequences. As an experiment, I borrowed a friends snare and recorded a variety of different hits to the hard drive of a computer. I edited the samples; about ten different strikes from soft to loud. The key to making this work in the SP is: each snare hit in the wave sample has to be exactly the same length. After editing, the wave sample would start at the leading edge of the softest hit and play successively thru hit ten and each successive hit would be 2.432 seconds long. That 2.432 number is just arbitrary, as the point is it may take the hardest hit 2.432 seconds to die out naturally and even though the softest strike may take 2.101 seconds; every hit in the sample must be exactly the same length. Now, just port the sample over to the SP and assign the sample to a zone so it only plays back on one key. Set the wave quantization parameter to how many segments the wave sample has in it; in this case 10. The math was never very important to me, but to find out what velocity values each segment in the snare wave responds to, divide 127 by 10. Here's an approximation: The first or softest hit in the wave sample responds to velocities 1- 12, the second hit in the sample 13-25, third 26-38, all the way up to the last and loudest segment in the wave sample which would be about 115-127. So now the harder you play the note, the SP plays a different and harder struck snare sample. Actually, this isn't a revelation as the SP manual describes this (albeit vaguely); I just didn't realize how to use it until I tried it."
"As far as other setup parameters in the SP, it's easy. The only thing you have to mess with is the decay of envelope 1. You want the envelope to close before the sampler plays the next wave in the sequence, meaning everything at 0 except the decay. In a ten segment snare sample of the length I described, it's probably a value of 35 or so. You just want to play the snare sample by striking a note on the keyboard and pull back the decay until you don't hear the edge of the next hit in the sample. No filtering, no nothing, and although you can play with velocities in the SP, if the sample has been edited well, there's no need. You can use this for kick drums, toms and ride cymbals too. A three or four segment kick usually does the trick, and the same for toms although you may want to play with velocity values a bit more for expression. For ride cymbals, of course, each strike takes a long time to decay but you can get away with a two or three segment sample. I tried this with high hats too; it does work but I had better luck going back to my old habit of just laying out twelve or thirteen strikes on the keyboard so I could get my hands on exactly what I wanted; from soft closed all the way up to a hard "slush" kind of a thing. Including two open hits, a footed hat etc. I usually assign them to one group so if you play an open high hat during a fill or whatever, the next high hat played cuts the decay off, pretty much like real."
"I know this sounds like a lot of messing, but for me, I've got it down to a science now. Beyond judicious choosing of samples, a good sample editing program and SMIDI setup makes it a breeze. From the time the sample is ported over to the SP, it takes about 30 seconds to get the new sample playing back satisfactorily."
#7) How To Burn An SP Format CD-ROM: Reprinted from my post at the Yahoo! Peavey DPM SP site:
The first step is to collect all your finished Banks on to one large hard drive. The SP needs to have all its sounds as one large file, as it can't jump around between files like a computer can. A CD will hold up to 700M, so you will want to have a hard drive the size of the CD you want to burn or slightly larger. I decided to buy a SCSI 1G Iomega Jaz drive off ebay; it came with one cartridge and cost me $27.00 U.S. I later got a second cartridge for $8.00. I started saving banks to the Jaz drive, but after the first 150M or so, the SP started slowing down and had problems saving. Tip #1: Always check for compatibility when buying a SCSI hard drive for any unit. Both the SP and the Iomega Jaz drive are somewhat old SCSI, and don't work very well together. Luckily, I also have an SP+, and it had no problem writing to the Jaz. Anthony at Peavey has told me that the SP+ has newer and faster SCSI, so it's more forgiving in this area. What if you don't have a Plus? A gentleman by the name of Jim Puckett has 540M SCSI hard drives for sale, and also has a couple of SP's that he's tested them with for compatibility. He can be reached at: PuckettJp@aol.com.
There are two types of drives that will work for our purposes: those larger than 700M (the size of a CD), and those smaller. If the drive is small than 700M, you can go ahead and disk image (more on this in a second) the whole thing (all sectors). If your drive is larger than 700M, then Tip #2: The SP saves its data backwards when saving to hard drive! I found this out the hard way, after saving 600M of data to the 1G JaZ drive starting on the first bank, imaging 600M of data to my computer, burning it to CD, and only the last 200 M would load into the SP. I then tried making an image from a (full) 100M Zip drive, and it would all load after being burned to CD. I then imaged & burned 50M from that drive, and although all the titles would show up when the CD was accessed, only the last 50M would load into the SP. Solution: first calculate the total size of your finished CD. In my case it was 604.5M. To do this, look under Disk on the SP, and go to “get info”. Subtract the available space from the total space, divide by 1024, and the answer is the total size of all your files. In my case 1044480k (total) - 425472k (available) is 619008k divided by 1024k (1M) = 604.5M of sounds. A 1G Jaz actually holds 1020M, so starting with an empty disk, I first had to save 415M of “junk” samples. These can be anything, even the same file over and over. One important thing to remember is that the Bank# order on the drive is not actually the order of how the data is saved. I had 135 SP banks, so I loaded my SP+ up with 32M of whatever was lying around (even silence samples from the SX will work), and started saving this over and over starting on Bank# 200, until I had 619008k space available on the drive. The SP will start saving this data from the end of the drive forward. Most imaging programs will image the data from the beginning of the drive to the end. By filling up the “back” end of the drive first, this will push the “good” data to the front. I then saved my “good” data starting on Bank# 0 to 135, erased the “junk” samples, set the file image size to 610M (just to be safe), and everything worked!
A disk image is an exact copy of a file from an external drive made on your computer's drive. I bought and used Disk 2 File to do this CD, although Translator will also make disk images from SCSI drives to your computer's drive. After I had pushed the “good” sample data to the front of the Jaz drive, and knew I wanted to image 610M from the Jaz drive to my computer's drive, I entered this info into Disk2File, and it imaged the first 610M of the Jaz drive to a folder on my computer. If your drive is less than 650-700M, you can select “all sectors” in Disk2File, otherwise you have to specify a file size.
After file is imaged to your computer, you need a burning program that will burn disk images without adding any extra “headers” at the beginning. I used the free demo version of Nero to do this, although the free version of CDRWIN will also work. Roxio “Easy CD Creator” adds extra info, and does not work. For Nero, just change the file extension from .iso to .nrg, click on the file, and the burn dialog box will open. Leave all parameters the way they are, and burn the CD.
The next thing I wanted to do was make a Demo tune featuring some of the sounds and include it on the CD. There are two types of CD's that contain data and audio: Mixed-Mode and CDExtra. Mixed-Mode has the data first, and the audio second. Most sampling CD's that have data and audio are this type, because samplers cannot see past the first track, but the CD player can play the audio from the second track. This did not work, however, so I also tried a CDExtra, which has the audio first and the data second. I probably burned 20 coasters trying to get these two formats to work, I even thought about including it as a .wav file on the CD, but in the end, I decided to post this first on this site, and now on my site as an MP3.
So there it is. Of course, if you have a Mac, I've heard that it's a much easier proposition to burn image files. If anyone needs help in doing their own SP CD-ROM, drop me a line, and I'll tell you what I can.