Sampling Hints The digital audio section of the AWE32 is not great by today's standards. Consequently when sampling audio from external sources, it is best to capture it digitally if possible (for example by "ripping" the audio directly from CD), or use a newer, better quality soundcard if sampling an analogue source. However, if you must record audio through the analogue line-in of the AWE32, there are a few tips that will help you achieve the best possible result. The first concerns the input-gain control on the AWE32 - make sure this gain control is always set to x1*. With the gain control set at x1, the signal to noise ratio (SNR) of the resulting recording is approximately 75db. However with gain settings of x2 and x4, the SNR drops to 70db and 65db respectively, basically resulting in a lot more background noise in your samples. Also try to avoid using the microphone input on the AWE32...it's VERY noisy. It's better to use an external microphone pre-amp, and run this through the line-in. *note that with certain versions of the AWE32 driver, the gain control setting is not available. Looping and Crossfading Some professional samplers have a built in crossfade feature, where part of the end of a loop is automatically mixed with the start, to make it smoother. Unfortunately the AWE32 has no such feature, meaning that it can be frustratingly difficult to produce good loops. This short tutorial shows you how to crossfade your samples before they are loaded into the AWE32's RAM, using a .wav file editor. 1) We start with a sample of a crowd cheering. As it is, it will loop without clicking, but it doesn't sound very convincing (i.e. from listening to it, the points at which the loop occurs are quite distinct). Our first step is to select approximately the first 20% of the waveform, and cut it out. Then separate this smaller section into it's own file. 2) Then fade this smaller section in, from 50% to 100% of the original volume (make sure it's a linear fade). 3) Now select a section at the end of the larger wave, that is exactly the same length as the smaller bit you cut out (and finishes at exactly the end of the larger waveform), and fade this out, from 100% to 50%. 4) The final step is to paste the smaller section into the larger, at the same point where you started the fade out. Make sure you mix the two waves, and don't replace what's already there. Now the loop is much more convincing. It takes many iterations before the listener can detect that it's a loop, and when mixed with other sounds/instruments, it may well be impossible for them to distinguish. Click here to listen to the original loop. Click here to listen to the crossfaded loop. Calculating BPM's Do you often find a great drum loop that is at the wrong tempo to fit in with your music? There are expensive programs out these days, that will stretch the loop automatically, but using a simple .wav file editor capable of time-stretching waveforms, and a bit of math, you can do it yourself. Take the following drum loop as an example. It consists of 4 beats, and goes for 2.47 seconds. From this we can calculate that one beat will take 2.47/4 = 0.6175 seconds. Hence this loop has a tempo of 60/0.6175 = 97.17 bpm. Now say we want the loop to fit into a 115 bpm composition. First we have to calculate how long this loop should go for, to fit into 115bpm. This is basically the reverse process of the above. First we find out how long 1 beat at 115bpm will take, i.e. 60/115 = 0.5217 seconds. Now multiply this number by 4 (0.5217x4), and we get the time that 4 beats at 115bpm will take i.e. 2.087 seconds. Now all you need to do is use the time stretch/compress feature of your .wav editor, to compress the loop to 2.087 seconds. Click here to listen to the original drum loop (97 bpm). Click here to listen to the drum loop at 115 bpm. Expanding the 'Size' of Samples Many of the synthesiser samples that I use in my own work, are created using a simple, subtractive soft-synthesiser called 'Analogic' (discussed later in this page). Whilst this program accurately simulates many of the features of hardware synthesisers, its major shortcoming is the quality of its built-in effects. As such, I usually export samples from the program dry (ie without any effects), but this often means the resulting instruments sound quite thin and frail. Using the simple techniques discussed below, you can give the samples (and their resulting instruments) a much more depth, and make them sound a lot 'bigger'. 1) Firstly make a backup copy of the original sample... it will be used later. The first technique is to make the single sample actually sound like it is multiple samples occurring at once. One way to do this is to make a copy of the original sound, and paste this copy back in, so that it starts very slightly after the original. Using a .wav editor, this can be achieved by simply copying the waveform to the clipboard, then placing the cursor 3ms from the start of the original waveform, and pasting the copy back so it mixes over the top of the original. 2) The second step (which also makes the sample sound like it consists of multiple instruments sounding simultaneously) is to create a chorus-type effect. Make a second copy of the waveform, and use the editor's time stretch function to make this copy 0.25% longer (and consequently slightly lower in pitch). Similarly, use the time stretch function to make the first waveform 0.25% shorter (and consequently slightly higher in pitch). Then paste-mix the two waveforms together. 3) The next technique is to give the sample some stereo separation... to get it to sound like it's emanating from the left and right of the listener, rather than from directly in front. The simplest way to achieve this is to convert the sample to stereo, and apply a phase shift to one of the channels. This means that one side of the stereo pair starts a fraction of a second after the other. In this example, I applied a phase shift of 4ms to the left side. The left channel in the waveform was phase shifted forward by 4ms 4) Now apply reverb to the sample. I could apply the EMU-8000's reverb effect once I had created a soundfont bank from the sample, but the AWE32's reverb is a little short. You can often create deeper, larger reverb effects using the reverb function in your .wav editor. In this example I used a reverb time of about 1600ms. 5) Now paste-mix this chorused, phase shifted, reverberated copy of the sample together with the backup copy you put aside in step one. In this example I used a 60:40 ratio of original sample to processed sample, but the exact amounts will vary in different cases. Click here to listen to the original sample (based on a sawtooth wave). Click here to listen to a riff using the original sample. Click here to listen to the same riff using the expanded sample. Analogic Although it's not directly related to the AWE32, I thought I had to give mention to a fantastic software synthesiser called Analogic. It has two oscillators, which can produce saw, square, triangle, noise and sine waves. It also incorporates pitch and volume envelopes, an LFO for each oscillator, and four types of resonant filter (high-pass, low-pass, band-pass and band-reject). It has an excellent and intuitive graphical interface, and takes under 300K of disk space. The best part is that you can export the sounds it creates directly to .wav files, and then use these in your soundfont banks. The built in effects are not very good, but you can export the sounds dry, and then apply high quality effects to them in a decent .wav file editor. The program requires you have a genuine AWE32 or SB16 (so unfortunately SB16 emulators will not do), and is a little picky about video cards. It also only works in true DOS (so it won't work under Windows 9x), but if your system meets these requirements, I highly recommend you have a look at it. It is great for creating sounds for use in electronic music. I have personally used it for a large number of different sounds in my own compositions. A big thanks to Newstyle software for providing such an excellent (and free) product. Click here to download Analogic.