Computers & Music:

Using the SB Soundcards

by Kai Pihl

The SoundFonts

The synthesizer of AWE, EMU8000 is designed by E-mu Sytems, Inc. It utilizes the so called wavetable technology. The samples are recorded and digitized from real instrument sounds. This means that they are not really synthetic. These sounds are stored in a 1 megabyte read-only-memory (ROM).

On AWE32 and SB32 there was also 512 kilobytes of dynamic RAM, which could loaded with user sounds. This RAM was expandable upto 28 megabytes with SIMM-modules. Note that the value edition of AWE had not this expanding possibility. The AWE64 has 512 kilobytes and AWE64 Gold has 4 megabytes of RAM. Both are expandable upto 28 MB with special RAM-expansion picky-back cards from Creative. User sounds can be made by recording them as WAV-files and using Vienna SoundFont editor to combine them as SoundFont banks. These banks are then read into the RAM of the card and utilized by the synthetizer.

SoundFont is a trademark of E-mu Systems. The first SF specification was specially made for AWE and was not public. It had a file-extension SBK. A new SoundFont 2.0 specification was released in the beginning of 1996. This spec is introduced to become a new standard and is public. It uses the SF2-extension. SoundFonts have such a name because they behave much like the character fonts. Character fonts contain usually the same characters in the same order, but there are exceptions. Here in Scandinavia for example, we have the letters Å, Ä and Ö in place of certain english characters. Many nations have their own special characters. However, the main purpose of a character font is to determine how the individual characters are drawn: Are they italics, bold, or have they some other special appearance?

In SoundFonts there are instruments instead of characters. As with alphabets the names and order of the instruments can vary. And just as the shape of a character, so the "soundshape" of an instrument helps us to regognize it. We can distinguish a solo violin from an alto violin, etc. In a SoundFont bank, the shape of the sound is determined with the sound parameters.

Separating the parameters from the samples makes it possible to use the same samples for other banks by "shaping" them again in a different way with the parameters. Also the logical order of the instruments in a bank can be changed. This explains how it is possible to get three different soundbanks from 1 meg ROM on the AWE. The General MIDI (GM) definition is a try to standardize the assortment. It contains 128 melodic instruments and 47 drum-instruments. 'General MIDI' bank needs none of the DRAM. The two others, GS and MT-32 where originally determined by Roland, a synthesizer manufacturer. They need a small amount of the DRAM.

Good samples versus poor samples

There is a lot of talk about the samples, whether they are good or not. The ROM samples of AWE are very often considered as poor. I would like to see this discussion to be in it's end. On my opinion it's not useful to blaim samples, not even the AWE. Much more useful is to check out what possibilities there is to get AWE behave in the way you want. In fact, there is quite a lot what you can do. Even the ROM samples can be used very nicely. The same way a trained player can get quite good sound out of a low-priced violin, a poor player can not get even a Stradivarius to sound good.

The microphone problem

Before we can go forward we must check out the microphone. A test with the shipped microphone unveiled that it could be suitable for voice controlling experiments and netphone applications but not for sampling. I think Creative has not even meaned it to serious sampling. I tried another microphone which I knew to be quite good. No sound at all! I thought that my plug was not compatible with the input jack. When I checked out the plug of the shipped microphone I saw three contacts. This little cutie couldn't be a stereo microphone. Then I realized that the third contact was for the operating voltage, that is needed for condenser microphones. Usually they get it from a battery in the microphone housing. For the shipped microphone it was taken from the card via the cable, a solution called the phantom voltage. Our microphone was dynamic, so we couldn't use it without an adapter.

Condenser microphones are good but expensive and if used with AWE they need to be checked for the compatibility against the phantom voltage and/or the possibility not to use it. A better solution to this problem is to buy a microphone pre-amplifier and connect the output of amplifier to the line-input of AWE. This solution is quite cheep, but if you want to spend a little more money, you can buy a full-featured mixer, that gives you the possibility to connect more devices to your AWE. We are now satisfied with our 6 channel audio mixer and a uni-directional dynamic microphone.

Sampling at home

Recording an instrument for creating a synthetizer sound of it, is called sampling. Sampling at home is quite the same as in the studio. You should be aware of environmental noise. The windows of the room must be closed. Airconditioning, computer itself, neighbours, dish- and washmachines etc. are all sources of disturbing noise. Your own ears are the best meter to check if there's audible environmental noise on your recordings or not.

When sampling you don't play a whole song but invidual notes or sounds, depending on the instrument. There are melodic insruments that play notes and non-melodic instruments that generate noise-like sounds (e.g. drums). Many times the instrument can be played in different ways. A violin can be bowed with different force, plucked, etc. A snare drum can be struck hardly to it's center, gently with the stick to its frame keeping the other end of the stick on the membrane (side stick), and so on. You should make it clear what you want to record.

When played, the synthetizer lowers and raises the sampled sound according to the key you hit. If the amount of lowering or raising is too big, the sample doesn't sound original anymore. That's why the samples must be recorded at different pitch levels. Vibrato should not be used, because it can be given afterwards with the SoundFont parameters. In all, getting good samples is not so easy. As with many other jobs, practising makes the master also here.

The recording can be made with any program capable to store the sound as a WAV-file. The way of calling recorded instrument sounds as samples is a little bit misleading. A whole WAV-file consists actually of an innumerable amount of very small parts called also samples. They are so small that they can't be heard individually. They represent the state of the sound at a given moment of the time. When put together and played as a series from the beginning to the end they can be heard as sound. An ordinary taperecorder stores the sound continuously without dividing it in small samples. That's why they are called analog recorders. A WAV-file contains the sound in an incontinuous, digital form. The computer, the soundcard and the software form together a digital recorder. Note: DAT-recorders, though using a tape, are digital recorders and make use of similar sampling technique than computers.

AWE comes with a quite versatile audio program, the WaveStudio. It's good enough for most of the purposes. It's capable to record and play back 8-bit and 16-bit audio, mono or stereo, with a speed up to 44100 samples per second. Stereo gives the sound a feeling of ambient. Although AWE is capable to produce stereo, I recommend not to use stereo samples before you understand troughoutly the operation of your AWE.

The sampling speed of 44100 samples per second is said to be the lower limit for high quality sound. For an ordinary user a speed of 22050 is enough. You can record the same sound with both speeds and compare them. There's not so big difference. The difference between 8 and 16 bits can be heard more clearly. So, at least in the beginning of our sampling career, we can use 16-bit resolution and 22050 sampling speed. It saves memory. I'm not saying that sampling with 22050 samples per second gives you high quality sound, but it gives an acceptable quality.

After you've read this article the practise is the best teacher in recording and sampling. There are many good books of the subject. Few of them are listed in the references.


Next: Shaping the sound Prev: First experiences To: The Contents