Before you can know HOW to sample, you will need to know WHAT sampling is.

"Sampling" is derived from the word "sample". A sample is like a single snapshot of a sound at a given point in time. The better your "camera" (in other words the A/D converter built into your sound card) the better the "snapshot" and the more accurate the sample. A/D is short for Analog to Digital. The analog voltage or wave should be a fair representation of a sound wave in real life. There are 3 basic factors which will determine the quality of the sample. The first is number of bits. An 8 bit sampler can only record 256 different VOLUMES or Voltages. Because of that you will be able to see very distinct STEPS if you look at the wave. This will only record sound that is as good as the sound on a telephone at best. Currently 16 bit is the standard for CD's. With 16 bits you can record up to 256 X 256 different LEVELS or volumes. (over 65,000)

The second sampling factor is sampling RATE. This rate is expressed in how OFTEN you "snapsot" or sample in a single SECOND of time. This will affect the BANDWIDTH or FREQUENCY range of the sample recorded. According to some physics and math calculation, the highest FREQUENCY will be 1/2 the sampling rate. In other words CD quality samples at 44,100 samples per second can reproduce frequencies over 22,000~ (kilocycles or 1000 X 22 cycles (waves) per second). Most humans can hear "HIGHS" between 16,000~ and 20,000~, so 44,100~ is more than adequate for a sampling RATE. If you pick a sampling rate of 11khz the sound recorded will be "muddy" sounding at best.

The other factor determining the quality of the sample will be the FILTER QUALITY. This is the internal filter located within the sound card. This filter will prevent the sound card A/D from recording any frequency above the SAMPLING rate. If the the A/D records anything higher that 23,000 (for an example with 44,100 sampling rate) you will get sampling ALIAS NOISE. This noise is UGLY and sounds like buzzing GARBAGE. If the filter is very poorly designed however it might clip TOO many of the HIGHS. So now when you go to buy that new sound card, remember BITS is not the only factor determining sound quality!

Ok. Now that you know a little bit of theory let's jump into sampling!

For simplicity sake I'll assume that you all have at least a 16 bit soundcard and a Intel based computer with VGA. You can download a demo version of Cool Edit from http://www.syntrillium.com

Step #1 Getting the Sound In

The wave is inserted into your sound card A/D by plugging a microphone into the MIC input or plugging your cassette, CD player etc. into the LINE input of your sound card. This is part of recording the wave. You will also need software such as COOL EDIT or Wave SE running at the same time to capture, view and edit the sound.

Important! You must properly CALIBRATE your sound card for the VOLUME/Impedence/DC Offset you will be using. I will have diagrams illustrating this more fully, but for now lets say that some systems will express volume in terms of "db" or have a "db VU meter".Some systems will specify Hi or Low level inputs on a CONTROL PANEL. (which looks like a picture of a bunch of buttons and switches. The DC offset calibration may not even be mentioned on cheaper sound cards. It is a common thing with the TBeach Monterey. (skip the next paragraph if you want to wait for the illustrations) To understand the importance of DC offset picture in your minds eye a rectangle which is longer than it is tall. Remember that your wave must fit into that rectangle and the more of that rectange that your wave fills up, the louder the captured sound will be. Since your wave is usually symmetrical and moves up and down from a ZERO line, you will be able to record the loudest volume if you get the zero line to cut the rectangle EXACTLY in half (along its length). That is what DC offset does. It adjusts the position of the zero line. (or makes GROUND=ZERO)

After you get a waveform (.wav) into the computer you can make it into a patch (instrument sound)in many different ways. The simplest way is by making a SHORT LOOP patch and most of the things that you learn here you will be able to apply to other sound patches.

Start by looking at the illustration. The area hi-lighted with CYAN is the area of the wave known as a single CYCLE. This is because a simple cycle starts at the ZERO CROSSING LINE,goes up in a POSITIVE direction, then down PAST the zero crossing line ,then back up to zero and then repeats for the next cycle. Although the illustration shows a more complex wave you can use the pattern of repetition when finding loop points. The hi-lighted area adheres to the rule below.


The first rule is this. LOOP POINTS SHOULD BE AT ZERO CROSSING POINTS. This is because there is SILENCE at the zero crossing point and by connecting silence to silence you get NO CLICKS, BUZZES etc. WHAT IS A LOOP POINT?

Imagine that SAMPLE that you recorded was about 1 second long. Compare this to a piece of string as long as your arm. We only want a piece of string 1millimeter long!We cut out this piece and connect the end to the beginning. (very hard to do with real string) This is called a LOOP. Now you know why. A LOOP is formed when the beginning of a string is connected to the end. With AUDIO we tell the sound to play over and over from the beginning of the cycle (START LOOP POINT) to the end of the cycle (END LOOP POINT). In other words our CYCLE can be compared to that 1mm piece of string.

You can hi-light the area between the loop points by holding the left mouse button down as you drag the mouse from left to right. (As I have done in the illustration using Cool Edit) Then to get a quick idea what looping sounds like COPY the hi-lighted area to the CLIPBOARD. (Like with all other Windows applications) You can OPEN up a NEW INSTANCE (run another copy of) Cool Edit by clicking on the FILE menu item. Then OPEN NEW from the same menu.

I don't used built in LOOPING FUNCTIONS. What I do instead is PASTE, PASTE, PASTE, PASTE etc. I paste about 8 times (the beginning of the cycle to the end of the cycle over and over)Then I COPY that 8 cycle wave and paste it over 8 more times etc. etc. When we have about 1000 or more cycles,then we will have a usable wave. (This function is sometimes difficult with COOL EDIT though, because you must constantly move the curser (pointer) from the beginning to the end of the wave to paste. If you have an old version of the SB16 "Creative Wave Studio" select the RAP function from the menu instead and "RAP" "rap" "rap" ETC.) Loop points themselves will be used with sound card synths but for the purpose of this lesson ,I'm showing you another way that will create more interesting and dynamic sounds.


Use the COOL EDIT PULLDOWN MENU "GENERATE" AND then select TONES then MONO 16BIT. Set up the controls exactly as I have in the PHOTO ALBUM PIC CALLED "generate tone" and hit "OK". This will sort of sound like a an old hammond organ sound 1.5 seconds long. Compare your wave with the close up called "hammond.gif"

After you have made a string of cycles 1000 cycles or more long, then play it. Sounds like a cheap 70's ORGAN or a telephone tone right? That is why we will SCULPT the wave in the next lesson.

Check out my review again if you wish.


Ask The Musician How it Was Done