This is a type of chip which converts the electrical signals from the input sockets into digital information which the computer can understand.
The initials A-D actually stand for Analogue-to-Digital. This means that the chip converts 'Real-World' information into computer data.
Aliasing is a peculiar effect, which becomes more evident in samples as the sample rate is reduced. The actual effect is caused by the frequencies above twice the sampling rate becoming wrapped into lower frequencies resulting in frequencies in the sound which shouldn't be present.
For the technically minded, this is because the sample rate can't adequately represent the frequencies, which the sound contains. They appear off the end of the sampling frequency spectrum and their energies are dissipated by wrapping back round into the lower end of the spectrum again, cropping up as the frequencies which are not evident in the original (the frequencies become evident as aliases in the original).
The level of the sound measured from the zero line. Amplitude can be measured as the actual sample value, which depends on the sampling resolution, or the most command form is in decibels.
See File Format.
Low frequency sound usually below 800 Hz.
The sound which reaches your ear at any given time will often consist of not one, but a number of different frequencies all at once. The human ear can detect frequencies from as low as 25 Hz to about 16-17,000 Hz. This is know as the bandwidth of the ear.
Of course, the term bandwidth can be applied equally to any range of frequencies such as filtering bandwidth (the range of frequencies a filter removes or boosts) or reception bandwidth (the range of frequencies a radio or TV can receive).
The beat of a sound is usually related to the tempo. This is usually noticeable from the drum beat.
See Resolution.
Within Soundprobe a buffer refers to a section of memory into which Soundprobe stores sound. This is always held in RAM for speed.
A byte is a single storage location with the computer's memory. In principle, each byte can hold one character of a letter or other document, or one value of an 8 bit sound sample.
See Resolution.
A section of RAM which is used to hold data for a short period while it is changed.
Small sudden changes which occur as the result of digital editing, CD errors or sometimes small pops from vinyl. These can quickly be found using either automatic region creation, or via the 2D Frequency Graph within Soundprobe.
The Windows clipboard is section of memory which holds data that can be shared by applications. In the case of Soundprobe this holds the sound that you wish to share between Soundprobe and another audio application, or another application that can use audio and supports the Windows clipboard.
This occurs when the level of the sound exceeds the maximum level that can be stored in the selected resolution. This is normally set by your sound card when recording, making sure the input level is low enough to prevent clipping is important in ensuring good quality sound recording.
It can also occur during editing. If you apply an effect, which amplifies the sound level, then the maximum level can exceed the maximum level, which can be stored at the set resolution. If this occurs, it is best to undo the operation, reduce the level of the sound, and retry the effect.
Clipping can easily be spotted where the sound level is at the top/bottom of the graph, and is a noticeable flat section of sound all at the maximum level.
The ability to continually record sound until the user stops the recording, the sound recorded is the determined to be the previous time just before recording was stop the length of which is determined by the time set to record.
This process reduces the dynamic range of the sound. Some sounds can change dramatically from low levels to high levels. This can cause damage to speakers if the audio level on the amplifier has been increased to hear the low sound level, which then suddenly changes to a high level. A compressor reduces the high levels to give a more constant sound level throughout the sound. Soundprobe offers this process within its Expander effect.
In the case of Soundprobe, this means that two sounds are similar. How strongly the correlate determines how similar they are.
The CPU is the heart of the computer that you are using. The letters CPU stand for Central Processing Unit. Without this, the computer couldn't even run programs! The CPU, also know as the Processor, actually executes the instructions which form the programs that you use.
The volume of two sounds is faded such that the first sound fades out as the second sound fades in.
To loop continuously.
See Decibel.
Most sound cards record with a slight offset from the centre line. This can be seen as the waveform sitting above or below the exact centre line, and can cause distortion if processed repeatedly. Soundprobe has the ability to correct this either during recording, or from the Enhance > DC Adjust menu item.
This is the exact opposite of the A-D converter.
The decibel is a logarithmic unit normally used for signal levels. The ear responds to sound levels logarithmically, which makes decibels a more natural unit when using sounds.
It is measured from a reference level, which is normally the quietist level you wish to measure or can measure. 0 dB represents the maximum amplitude, and it becomes negative as the level decreases.
The lowest level depends upon the resolution of the sound with a sample value of one being the smallest measured value. Zero sound level is normally represented by negative infinity (- INF), but even a level of zero with digital sound has some output.
This simply means that any editing or effects you apply are done directly to the file you are editing. Soundprobe uses non-destructive editing, which means that when you open a file, Soundprobe loads this into its own virtual memory and doesn't save the changes until you select save.
A device is usually associated with a piece of hardware connect to your computer. Windows uses a small piece of software known as a device driver to communicate with that hardware.
A window in which you usually have OK and Cancel buttons.
The effects usually associated with having the gain too high usually causing some form of clipping of the waveform.
When using a low sampling resolution, a quiet sound can only fill some of the lowest values in the range. This usually occurs with 8-bit sound where the total range is only 256 levels, so a quiet sound may only occupy 1 or 2 levels, giving poor sound quality, which results in white noise. To mask this, adding a small amount of random noise (dither), can force the sound to occupy more of the levels, and also gives a more pleasing sound to the ear rather than the odd pop from the level change between 0 and 1 occasionally.
In Soundprobe a document is where the sound and all associate information is kept internally.
A device driver is a small piece of software that is used to communicate with a piece of hardware.
See Device.
This is the range of variation of the sounds amplitude. A piece of music with many low and high sections has a greater dynamic range than a voice, which has a narrower range.
Envelope of usually associate with the ADSR (attack decay sustain release), which changes how the amplitude of a sound varies over time. In Soundprobe this can also apply to other parameters and not just the gain (used to create a volume envelope). An envelope in this case is the shape of how the parameter changes over time.
This is the opposite of a compressor in which the sound level is boosted to increase the dynamic range. In Soundprobe, however, the Expander effect actually combines a compressor and expander into one, allowing you to do both, or either, and even do a noise gate, you are able to dynamically change how the sounds dynamic range changes.
A technique used to split a sound into its frequencies. The result can then be processed in complex ways and then transformed back.
See Fast Fourier Transform.
A file containing information and how the data within it is stored. In the case of a sound file format, the file contains information about the sound, such as the sampling rate and length, plus the sound data itself.
When many small files have been saved onto a hard disk, it is likely that they have spaces between them (free disk space). When you then save a larger file to your hard disk, the free space is spread over many smaller areas, degrading the speed at which the hard drive can write the data. This is called fragmentation. Windows offers a simple utility to defragment your hard drive (found in the System Tools).
As the word implies, this is the number of times something happens within a given time period.
In the case of sound sampling, frequency is the term used to define the speed of a set of sound waves and is measured as the number of sound waves per second (Hertz).
The ability to directly edit a spectrogram or other frequency display.
A filter is something that will prevent a specific element from passing through it. For example, a sieve is a form of filter. You pour flour or sugar into it and the small holes allow the fine grains to pass through, retaining any lumps.
It is possible to selectively remove a range (or band) of frequencies from a sound once it is recorded. The frequencies, which are allowed to stay, are known as the pass band, those, which are to be removed, are said to be in the stop band. The difference between the upper and lower frequency limits of a filter are known as the bandwidth.
Heinrich Hertz was a German physicist who lived in the late-1800s. In the course of his work he discovered that he could generate electromagnetic waves causing electricity to arc between two metallic spheres. The electromagnetic waves that he discovered are used today to send radio and TV transmissions around the world (and beyond).
Subsequently, scientists have used his surname to express the number of time that some form of regular or wave-like event occurs within one second. Hence 50 to 100 Hertz could define the frequency range (or bandwidth) of a hi-fi system's bass speaker. In this case the speaker can reproduce sound waves which repeat from and 100 times per second.
A kilo is a measure of one thousand. One kilogram (kg) is one thousand grams.
Computers count in multiples of 2 (where as we normally count in multiples of 10) so, to make calculations easier, the computer kilo is 2^10 or 1024.
Therefore a computer which uses 4 kB storage is actually using 4 x 1024 = 4096 bytes of storage space.
The computer's memory is the part of the machine where programs and information such as letters, documents and sound samples are held. This part of the machine is active only when the machine is switched on. Unfortunately, when the computer is switched off, this data is normally lost. Information and programs, which need to be kept, must be saved onto a hard disk for safe keeping before the power to the computer is removed.
A note is a single tone, which is generated by playing an instrument. Notes are grouped into sets known as octaves, where they are allocated names starting from A to G and back to A.
A noise gate is similar to a compressor/expander. It eliminates low level sound by compressing the sections so they become inaudible. This can be used to remove background noise during quiet areas of sound, or remove noise from gap in speech.
This is the maximum frequency, which can be produced at a given rate. This is half the recorded sampling rate. It is best to record just over twice the maximum frequency you wish to reproduce to prevent aliasing effects.
On a musical keyboard, the traditional white and black keys play musical notes. If you look carefully at the in which these keys are organised, you should notice a repeating pattern.
For every seven white keys, there are a group of five black keys. These twelve keys together form what is known as an octave. Each repeating pattern of keys represents a repeating order of notes. The white keys run from note A to note G and back again. The black keys represent incidental tones which are known as Sharps or Flats (hence 'C sharp' etc.)
The purest form of what is known as 'International A' is defined as the tone represented by a 440 Hz sine wave. Each time you increase the notes by one Octave, you double the frequency, so the 'A' one octave higher is in fact a tone of 880 Hz and the one below is 220 Hz.
Pulse Code Modulation is the standard form used to encode sound digitally. This is the general uncompressed form used by Windows.
See CPU.
This is a form of distortion, which can occur when the sample volume isn't set correctly. In effect, the problem arises because the digitiser doesn't have enough resolution to accurately retrace the outline of the sound wave.
As a result, steps appear in the digitised waveform. These steps become audible and the ear interprets them as extra harmonics, which weren't present in the original sound.
This is another name for the computer's memory. In actual fact, the letters stand for Random Access Memory. This means that its contents can be changed at any time. One small problem with RAM is that when the power is removed from the computer, it develops amnesia!
The sampling rate is the number of times the computer scans the sampler every second. Each time the sampler is read, a new sound value is stored in the computer. For example, a sample rate of 10 kHz would mean that Sound Probe will capture 10,000 values of the sound every second.
The sample rate is one of the factors, which determine the quality of the sound recorded into the computer. To an extent, the faster the sample rate, the higher the quality the reproduced sound will be. This causes a problem.
The faster the sampling rate, the quicker the computer's memory fills up. This leads to shorter recording times. Ultimately a compromise will often need to be struck between the size and quality of the recordings, which you make.
Unfortunately only experience can guide you on what the compromise can be. As a guide to beginners, sounds which are very low frequency (lots of bass) such as a bass drum or bass guitars, can be sampled at a lower rate then a sound of a higher frequency, such as a cymbal or singing voice.
As you may already know, the computer stores information in its memory. The basic measure of storage is known as a BIT (BInary digiT). If you think of your hands, then one bit is equivalent to one of your fingers. One bit on its own can be used to count from 0 to 1.
Unfortunately, this isn't really a very useful quantity on its own, so these bits are grouped together in clusters of 8 to form a byte. This would be analogous to one hand (but with 8 fingers!). With 8 bits in a cluster in this way, it becomes possible to hold numerical values between 0 and 255 (this is a binary count - don't worry, just accept it as fact). Therefore, an 8-bit sample can have up to 256 different voltage levels. This can provide very good quality for loud sounds, but the ear is too sensitive to be fooled as the sound gets quieter, since the steps are quite coarse and are often detectable.
The next size up from a byte is known as a word and is 16 bits in size, or two bytes. This could be similar to having 8 fingers on both hands. With 16 bits, the computer can store numbers from 0 to 65535. This means that a 16 bit sound sample can have over 65,000 different levels. The human ear is not quite sensitive enough to be able to detect the difference between the steps of a 16-bit sound. Consequently, a sound, which is digitised to a 16-bit resolution, will be subject to less quantisation than an 8-bit sound.
Real-time is one of those computer phrases, which donÆt appear to make sense at first.
In the early days of computers, they were slow and ponderous by today's standards. They were good at handling business data such as sets of accounts and processing documents, things which they could be left to process overnight and provide the results the following day. This is known as 'batch' processing because the data was gathered during the day and the results provided the next.
This is all well and good but scientific users wanted to analyse things such as sound and the only way in which they could do this was to use special data recorders to capture the information and then download this into the computer later. Often the transfer of the information into the computer took longer than the time to record it. The problem was compounded by the fact that processing this information took even longer still; the scientists wanted to be able to capture and process the information as it happened. Thus the phrase to record and analyse information from the real world in the time it took to happen (i.e. real-time) was born.
Luckily, computers are faster nowadays and Soundprobe has just such real-time facilities. The main area of Soundprobe which uses real-time processing are the real-time effects. These take the input from your sound card, processes it with an effect and send it straight out of the sound card all within less than a second.
Sample is the name given to a sound once it has been recorded into the computer. Once in the machine, it is held as a series of numbers (digital values) which the computer can store and process in its own way.
This is the act of recording a signal into a computer via an analogue-to-digital (A-D) converter. Computers don't only sampler sound, they are also used to measure other real-world activity such as temperature, movement, light levels and the like.
This is known as sampling because it is the collection of data on a regular basis over a given time, even if that is months or years!
Sound is transferred from one person's mouth to another's ear by a series of pressure waves, which travel through the air. In principle, these pressure wave are very similar to ripples on the surface of your bath water!
Signal-to-Noise ratio is the ratio between the noise level and the actual signal level. Higher ratios give better masking of any noise that may be present.
Sound is normally invisible because it travels through air. However, it travels as a series of pressure waves, similar to those, which you would see on a still pond when something disturbs the surface. If you could take a cross section across the wave you would see the top of the wave (the peak) and the lowest point (known as the trough).
This is the point at which the waveform crosses the zero line. Performing editing at this point reduces the chances of audible clicks due to sudden changes in the waveform.