Contents
Chapter 1: Overview
Interface conventions
Special buttons
The Sample view
User-definable Edit regions
Sample view keyboard shortcuts
| Chapter 2: File options | |
| Loading samples | |
| Saving samples | |
| MIDI sample transfer | |
| Sample analysis | |
| Chapter 3: Clipboard editing | |
| Copy | |
| Paste | |
| Cut out | |
| Auto-crop | |
| Overlay | |
| Undo | |
| Chapter 4: Processing a Sample | |
| Re-synthesis tools | |
| Audio and Image Filters | |
| Studio Effects | |
| Chapter 5: Additional tools | |
| Stereo <-> Mono converting | |
| Set playback frequency | |
| Drum Split | |
| Stereo Cross-fade | |
In programming Squash it!, we've tried to make the program as flexible and fully-featured as possible, whilst still retaining a sense of ease-of-use and enjoyment. Squash it! works in 100% GEM environment, so it's possible to use desk accessories (memory permitting) alongside it, and may also work under some multi-tasking operating systems. Since it is GEM-based, Squash it! works in an almost identical way to the desktop itself, with lots of windows, and a menu bar, but there are things we've added to the program that you might not be aware of, hence this section.
| [control]+[d] | Toggle sample draw |
| [l] | Toggle left channel edit (stereo samples only) |
| [r] | Toggle right channel edit (stereo samples only) |
| [m] | Toggle editable area between maximum and marked regions |
| [p] | play sample |
| [+](+[shift]) | zoom in (10 x factor) |
| [-](+[shift]) | zoom out (10x factor) |
New
Open
[control]+[o]
Squash it! will effectively load in any data of any size (up to the maximum amount of free memory), although it has direct support for AVR, AIFF, and Intel-WAV formats (in eight or sixteen-bits, mono or stereo). Even if the extension is wrong, Squash it! will recognise the sample's format.
Formats such as SD1/SD2 (Sound Designer), IFF, or DVSM, and so on, can be loaded, although you must manually set the sample's properties in the raw sample dialogue box (which will automatically appear if the sample is not recognised). These properties include the sample's bit depth (8/16-bit), its channel status (mono/stereo), and sampling frequency in Hertz (from 8 to 50KHz). To edit the raw sample frequency setting, press and hold the left or right arrow buttons to lower or raise the sampling frequency. It must be noted that, raw sample data, if 16-bit, is assumed to be Motorola format.
If sample load is successful, the sample window is updated and all editing processes are activated. Note that, since raw sample import ignores any previously saved header information, it's likely that the first one hundred or so bytes of information in the sample will sound like noise (this is the header). This can easily be deleted using the Sample Cut Out feature (detailed below).
Load Clipboard
[alternate]+[l]
Sample data can be loaded directly into the clipboard, effectively allowing you to have two samples in memory. However, you cannot load raw sample data into the clipboard directly; the sample format must be AVR, AIFF, or WAV, else it won't load.
Save as...
Save block
[control]+[b]
User-defined sections of any sample can be saved as individual samples. When selecting Save block, please make sure that the desired area is either encapsulated by the user-definable flags, or in between the left and right sample window positions, else the whole sample will be saved. The sample is automatically saved as an AVR format file, in 16-bits.
Save Clipboard
[alternate]+[s]
Although you cannot edit the clipboard directly, the need to save it to disk may arise at some point (particularly when using it as a scratch-pad or alternative undo buffer), and so this option is provided. As with the block save, the clipboard is automatically saved as an AVR format file, in 16-bits.
Copy
Paste
Cut Out
Auto-crop
Insert
[control]+[i]
This mode works in a similar fashion to Paste, copying the clipboard data to a user-defined section of the sample. However, unlike Paste, Insert does not overwrite the data, rather it moves it to the right, allowing the clipboard data to be 'inserted' into the sample. Choosing Insert from the Edit menu with the [shift] key held down, or pressing [control]+[shift]+[i] will insert the entire clipboard, starting at the left or first marker, and ending when the clipboard finishes. Otherwise, Insert will copy into the desired region only, ignoring the remainder of the clipboard data if the region is smaller. If there is no sample in memory, Insert will work in exactly the same way as Paste.
Ring Modulate ([2])
Ring modulating a sound increases the number of side-bands in a sound's spectrum, and also their amplitudes relative to the fundamental frequency. Usually, this results in a metallic sounding sample, and lots of synthesizers will allow you to ring modulate with a fixed-frequency sine wave or similar. With Squash it! though, it's possible to use the clipboard as the modulator, for some truly expressive sounds.
Amplitude modulate ([3])
Amplitude modulation is similar to ring modulation, except that the sample's loudness is affected rather than its frequency. The result is similar to an amplitude LFO except of course that the clipboard contains many more amplitude nuances than a simple waveform, and so can be very interesting.
Mix ([4])
Mixing offers a way of combining the clipboard and sample for a hybrid sound. In order to fit both samples into the maximum headroom of 96dB allowable in a 16-bit sample, both the samples are scaled by 50%. If the selected region is longer than the clipboard, the remainder of the sample is left unaffected, except for the initial section which is carefully faded back to 100% to avoid an unpleasant sudden increase in amplitude. If however, the clipboard is longer than the edit region, then the remainder of the clipboard is ignored.
Sum ([5])
Summing the sample and clipboard works in much the same way as mixing, except that rather than scaled, the two samples are subjected to a hybrid filter which, whilst reducing overall volume will brighten the loudest frequencies of both samples. As with Mix, Sum will fade back into unaffected sample data.
Filter modulate ([6])
This last mode can be quite effective when used with the right samples. Filter modulate uses a pseudo-low-pass filter which is then open and closed according to the amplitude of the clipboard. Because the clipboard's amplitude is indeterminable, sometimes the filter can become unstable and begin to accentuate frequencies, but this more interesting than problematic.
Unlike other sample editors, Squash it! has been designed to allow you to completely re-synthesize a sample; with more than forty different tools, it's easy to drastically alter a sound's temporal and spectral characteristics beyond recognition. Generally, most processes are permissible for stereo and mono samples, although a few are limited to mono samples. And most too will allow you to edit regions as well as the entire sample (unless otherwise specified).
Reverse
[control]+[r]
This function will allow you to make a sample play backwards. Since it also works with range and channel markers, there is a great deal more flexibility to it than simply playing backwards! For example, one channel of a stereo sample may be reversed, or multiple sections too.
Tile
[control]+[t]
Tiling a sample allows you to easily repeat a defined section a number of times, making it possible to create whole new sounds from small sections of an existing sample. Since the Tile rack-mount has its own ranges, it does not use the markers set in the sample view (neither will it take into account channel flags).
To define an area to be tiled, first select the offset from zero, then the actual size of the tile. These settings will determine what part and how much of the original sample you want to use as a tile. Next, decide how many times you would like to repeat the sound, bearing in mind the available memory, and whether you would like them to overlap, or not. Tiles may overlap up to 50% of their size, thus giving the sound an increasingly complex tone, as well as making transitions between tiles smoother. Tiles may also be placed in an alternating fashion, or normally, where the former reverses each odd tile, just as you might with black and white kitchen floor tiles.
LFO
[control]+[0]
LFO or low frequency oscillation can be applied to a sample to give it a vibrato or tremolo effect. To choose between amplitude or pitch LFOs, select the appropriate mode in the LFO rack-mount by pressing on the LED (it will light up). Please note that pitch LFO is only available on mono samples.
The oscillating waveform can be set from 1Hz, to 50Hz, the greater the frequency, the more rapidly changes occur in the output sound. Squash it! also has a depth control, as well as attack and decay times, that determine the intensity of the LFO. Greater intensities will produce greater amplitude and pitch changes in the sample, whereas lower settings will produce a more subtle effect.
In addition, three shapes are offered (only applicable to amplitude LFO mode): square, sawtooth, or sinusoidal. Square waveform is currently the only waveform available for pitch modulation.
Envelope Shaper
[control]+[1]
Unlike other sample editors, Squash it! does not have any 'fade' tools, rather it uses a four-pole amplitude envelope generator to produce the same (and more flexible) results. In addition, the envelope shaper not only attenuates volumes, but will also increase them!
Each of the four dials controls the volume of the envelope at each stage of the ADSR (attack-decay-sustain-release) shape, where a setting of half-way is equal to no attenuation; lower settings attenuate, and greater settings actually increase the sample's original volume. Next to the four dials are three sliders labelled attack, decay, and release. These determine the length of time for each envelope section, up to a maximum of 33% of the total length of the defined area. The remainder of the time is given to the sustain portion of the sound, thus if all are set to maximum, there is effectively no sustain.
Squash/Stretch
[control]+[2]
This is a dual algorithm, and is designed for changing the temporal characteristics of the sample. As a by-product, in 'scale' mode, the pitch of the sample will change too (longer samples will sound lower, and shorter, higher). However, with mono samples only, it is possible to alter the temporal characteristics of the sound without affecting its pitch, using the 'factor' mode. Since Squash/Stretch affects the sample's time, it can only be applied to a whole sample.
In 'scale' mode, the dial will control the amount by which the sample is stretched or squashed (according to the mode set with the LEDs), up to a maximum of 200% of the original length with stretch, or 10% using squash. In addition, the squash mode has an extra option of 'stack' or 'obliterate', where the former selection will retain all sample data (stacking it upon itself), whilst the latter discards unwanted samples.
In 'factor' mode, the dial decides whether the sample is to be expanded or compressed, rather than the LED switches themselves. For example, a dial setting of 0 produces the shortest possible sound, whereas 10 produces the maximum expansion possible. In this mode, a setting of around half-way will have the least effect. Factor mode uses a system of time granularisation, whereby intermittent sample windows are repeated or removed, and then smoothed together. This allows the pitch to be largely unaffected (and is perfect for drum loops), although some artefacts do appear as a by-product of the process.
Explode
[control]+[3]
Explode will take a mono or stereo sample (including L/R channel select), and 'explode' it. The effect, depending on the parameters set, can range from a variable amplitude LFO, to a disjointed 'explosion' of sound grains.
First of all, the explosion's epicentre must be set, and this can be any point in the sample (some good effects can be generated using the start or end points). Next, the size of the debris, and the amount of turbulence are determined using their respective dials. Debris controls the size of the sound grain, whilst turbulence determines the pseudo-random amount of the original sample data to mix with the explosion.
At the explosion's epicentre, the amplitude curve is at its loudest, producing a scale of 200% of that of the original sample volume; as the debris dissipates from the centre, so the amplitudes reduce exponentially. Setting the debris size to a small value will result in sound grains being spaced far apart, whereas larger debris will join up.
Granularise
[control]+[4]
Audio grains are small sections of a sample that, when combined, create a whole sound. Squash it!'s Granularise tool allows you to re-order the grains in an existing sample, to create a whole new sound from the same composite parts. in addition to the uniform granularisation (where each grain is a pre-determined size), Granularise also has a gated feature, ideal for re-shuffling drum loops to form new, and exciting beats.
In either uniform or gated modes, there can be up to 100 grains which are either determined by the number of grains setting (for uniform), or by the gate threshold control. Squash it! will then take each of these grains and re-order them randomly, applying an amplitude envelope to each one to make transitions between grains seamless.
The way in which the grains are re-ordered although random, is determinable using a user-defined behaviour pattern. There are three buttons: attraction to sample start and end, and uniform granularisation. The former two modes will mix many grains together, overlapping them to produce a rich hybrid tone, whereas the latter is more useful for atonal sounds and effects (such as percussion loops).
Ring modulation
[control]+[5]
Unlike the ring modulation in the Overlay section, this module self-oscillates the sample using one of two modes. By selecting the option from the menu bar as normal, Squash it! will ring modulate the selected region and/or channel, with a reversed copy of itself. However, when the menu is selected with the [shift] key held down (or pressing [control]+[shift]+[5]), it will ring modulate the sample with a phase-shifted copy. The degree of phase shifting is user-definable too. When the option is selected, the mouse cursor will initially turn to a text marker, indicating that it needs further input. Use the keys [1] through to [9] (not on the keypad), to control the phase shift.
Phase Modulation
[control]+[6]
This function will allow a sample to self-modulate using a time-variable phase-shifted copy of itself. The results can vary from simple sum and cancellation effects to the amplitude (caused by overlapping out-of-phase waveforms), to complex frequency partial additions.
Phase modulation has just three controls: Offset determines the initial amount that the phase-shifting starts, expressed as a percentage of the estimated cycle length. Thus, setting the dial half-way, should produce a phase shifted copy 180° out of phase with the original sample. Depth controls the maximum phase shift of the waveform, and rate controls the speed at which the maximum depth is reached. It is this rate control that is responsible for adding very pronounced harmonics to the original sample (up to around twice and half the fundamental frequency, depending of course on the rate's speed).
Squared Modulation
[control]+[6]
This is an odd modulator, in that whilst it adds harmonics, it actually reduces the differences in amplitude. As with the other modulators, it self-modulates the desired sample, with provisions for ranges and channels. Repeated use of the squared modulator will result in a square wave sound, regardless of the original sample's waveform properties.
Optimise
[alternate]+[o]
This menu option offers a fast-track to volume optimising (making the peak amplitude near or at 100% of the maximum allowable amplitude). Further volume control can be sought in the volume tools rack-mount box.
More volume tools...
[alternate]+[v]
The Volume Tools rack-mount contains four processes for affecting the sound's amplitude, including a duplication of the Optimise option in the Tools menu. Three of the modes are selectable using the keys [1] through [3] (or by pressing the radio button on the rack-mount), whilst the fourth, gain, is the default tool, and is automatically used if no other is selected.
Zero is perhaps the simplest of them all, since all this does is clear the selected region of any sample data. (Of course, this module only has a use when used in conjunction with the range markers, and then small sections can be cleared of noise or glitches.) Overdrive, like Optimise will increase the sound's amplitude content, although it will increase its volume beyond the threshold for 16-bit samples (96dB), eventually hard clipping all sample data. However, with lower settings, Overdrive can be more useful than the Optimise tool for gaining overall optimum volume. Overdrive is used in conjunction with the left-most dial in the Volume Tools rack-mount.
In its simplest form, Gain can be used to amplify or accentuate the sample's volume, by a scale of ±10%. However, in addition to the gain control itself, there are also minimum and maximum threshold controls, which determine the amplitude area to be affected. Using these it is possible to amplify the louder sections, whilst keeping the quieter (that which is likely to be noise) sections un-amplified. Likewise, quieter amplitudes can be boosted, whilst loud sections are kept the same, giving a similar effect as an expander.
Erode
[control]+[e]
Erode contains a number of sub-processes that will allow you to distort and corrode the sound in some interesting ways. The first dial does just that; erodes parts of the sound away by a user-definable amount, adding an amount of noise as a by-product of the process. If however, you just want to add white noise, then there's a special dial just for that (the dial for this controls the volume of noise added)!.
Erode will also allow you to add record crackle and low-frequency hum (like a dodgy pre-amp), to make drum loops and other noises sound as if they've been sampled from vinyl! The frequency of the hum and its level can be controlled, as can the degree of crackling, so it's possible to emulate a new 12" record, or one that's been sitting around in a dusty room for years. Combined with noise, and a low-pass filter, some truly authentic grungy loops can be created from otherwise perfectly sourced sample data.
Each erosion process can be applied as many times as necessary. To disable a particular erosion algorithm, set its depth or volume control to 0, and it will be ignored.
Smooth/Expand
[alternate]+[m]
This is another dual rack-mount processor, and has been combined since each process does the opposite of the other. To choose either processor (since both cannot be applied simultaneously), click on the appropriate LED in the half-rack panel (or press [1] for Smooth, and [2] for Expand). Smooth has a similar effect to a low-pass filter, smoothing differences in amplitudes. However, unlike a normal audio filter, the process can be applied to certain amplitude regions (using the lower and upper threshold controls). Also, Smooth can be made to feed-back on itself (much like an unstable filter), producing very 'unsmoothed' surfaces! This is achieved by setting the depth control dial to greater than half-way.
Expand does the opposite. Rather than smoothing the amplitudes between juxtaposed samples, it accentuates them, resulting in a boost in the high frequency content of the sample. Like Smooth, Expand can also become unstable, and is also prone to sharp peaks in the audio (which are automatically compressed to save digital distortion), when the depth setting rises above half-way. Using the threshold controls it is also possible to use Expand as a pseudo-EQ, as only some parts, and thus some frequencies, will be expanded.
Logic filter
[control]+[l]
This collection of filters is so-called since it just deals with samples as if they were huge blocks of numeric data (which really they are). Each sub-process can be selected by clicking on the appropriate button or pressing keys [1] through to [4] on the keyboard. The two simplest modes are De-click, and Flip, where the latter simply produces a phase-reversed version of itself (negative values become positive, and vice versa), and the former attempts to subdue any audio anomalies that it finds (more drastic de-clicking can actually be achieved with blurring). By flipping only certain regions or channels, it's possible to create some very interesting phase-relationship effects in stereo samples.
Shuffle and Shift are more complex in that they both have three sub-modes of their own. These sub-modes determine what happens to the sample data once it exceeds the maximum allowable value.
Shift allows you to move the sample data up and down in the amplitude domain, by a user-definable amount (to a maximum of -32,768. and +32,767). In terms of repairing samples, Shift allows you to re-set the equilibrium, making all peaks evenly distributed in the positive and negative domains, and can also 'fix' unsigned sample data (since Squash it! assumes all sample data to be signed, regardless of any settings in the file header). To achieve this, shift the sample upwards by 32,767, using the 'wrap' sub-mode. Wrap actually allows you to keep shifting waveform data up and down without actually destroying it. Once the upper threshold has been breached, sample data will begin to appear on the bottom, so it is actually possible to turn a sample inside-out! In 'destroy' mode, all sample data that goes beyond the amplitude threshold is rounded off, resulting in hard compression. 'Squash' mode does a similar thing except that instead of rounding off data, it is folded back on itself.
Shuffle does the same thing as Shift, only in the time domain. As with Shift, the amount to 'move' the sample by can be a positive or negative integer, but this time up to a maximum of the sample frame size. Using the wrap sub-mode, shuffling audio data to the left, past the sample end, will result in it appearing at the start of the sample (shuffling to the right has the reverse effect). In destroy mode, all sample data that is 'shuffled' off the end of a sound is lost, which can actually be beneficial when trying to remove start or end portions of a sound. Finally, in 'squash' mode, the sample data is folded back on itself, resulting in a mix of reversed and non-reversed sample.
Comb filter
[control]+[b]
Comb filters are an interesting variety of filter since they can be used to emphasise a sound's fundamental and all its sub-harmonics, by filtering out all the others. Squash it! actually has two varieties of comb filter, FIR and IIR, the latter of which produces a more resonant sounding tone (since it is based on a recursive or feed-back filter loop). The FIR Comb filter can be chosen by selecting the menu option as normal; for the IIR filter, select Comb filter with the [shift] key held down (or press [control]+[shift]+[b]).
Whichever filter is used, the size of the filter bands are user-definable. When the option is selected, the mouse cursor will initially turn to a text marker, indicating that it needs further input. Use the keys [1] through to [9] (not on the keypad), to control the bandwidth size.
Low/High-pass filters
[alternate]+[f]
Both types of filters are combined into one rack-mount, and can be toggled using either keys [1] and [2], or by pressing the appropriate LED on the rack-mount itself. Each filter section contains two dials that set the cut-off frequency (fine and coarse settings), and a bar graph that displays the roll-off slope. The default is 6dB (single-pole filter), but up to 24dB per octave (four-pole) filter cut-off slopes are possible, although the greater they are, the longer it takes to calculate. To increase or decrease the roll-off slope, press and hold the up or down arrows next to the four segment LED bar graph (the number of lit LEDs denote an increase in 6dB per octave).
The low-pass filter has an 'ideal' range of 0 - 5KHz, and the high-pass from 1KHz to 20KHz, although of course the roll-off slope has a profound influence on the actual frequencies that are affected. For example, with a slower roll-off slope, more frequencies above or below the cut-off will pass through, whereas a 24dB slope will allow only the frequencies above or below the cut-off point.
Digital Delay
[alternate]+[d]
Digital delays have many uses as a studio effects processor; not only can they be used for their intended purpose, but they can also be used to create pitch-related effects (when short delay times are used), and pseudo-reverberation. Delays can be applied to either left or right, or both channels (when used with stereo samples), but cannot be applied to user-defined ranges.
Squash it!'s digital delay has a set of four dials: two are used to set the delay time itself; one for fine tuning (values between 0 and 25ms), and one for coarse settings (0 - 625ms). Used together, it is possible to set any delay time from anywhere in between 0 and 650ms.
The other dials control the quality of the delay, allowing you to control rate and feedback characteristics. Feedback determines how much of the original volume is mixed in with the delayed signal, where the maximum setting will actually cause the delays never to reduce amplitude (a setting of 0 is normal). Rate determines how many delay-taps there are in the effect; larger values obviously produce longer delay times and a more pronounced effect, though low rates can be very useful for generating pitch effects, when a short decay time is set too.
Harmoniser
[alternate]+[h]
Harmonisers add a pitched up or down effect to an input signal, and are frequently used in studios for embellishing vocals to making chorusing effects and so on. Squash it!'s harmoniser works only with mono samples, and can only be applied to entire sounds (ignoring any edit ranges set). However, it does have two modes of operation (toggled by pressing the 'scale to fit' LED), making it very flexible. Also, the maximum range of pitch harmonising is ±1 octave. With 'scale to fit' disabled then the length of time of the effect is variable. This depends on whether it is pitched upwards, in which case, the sample effect is shorter than the original sample, or downwards, making the effect longer (and thus truncating it). 'Scale to fit' ensures that, to the best of Squash it!'s time compression/expansion algorithm, the pitched effect is the same duration as the original sample. For sounds such as drum loops and vocal passages, where the timing is critical, this is essential.
The Harmoniser rack-mount may also be used as a standard pitch-shifter, since there is a control for the dry:effect balance. Setting this dial to the maximum 'effect' bias will mean that none of the original signal is mixed with the pitched effect. Of course, it's primary use is to control the mix of original pitch and affected pitch, so you can create more subtle harmonising effects.
Ping-pong Delay
[alternate]+[p]
Ping-pong delay has some similarities to the standard digital delay except that, like the Rotary Speaker Simulator, it is a special stereo effect. Thus, neither the range markers or the edit channel flags are used, and of course, you will end up with a stereo sample at the end of processing! Ping-pong delay differs from the standard digital delay in that the decay-taps are positioned alternately on the left and right channels. Also, although all the other controls for decay rate and delay times are the same as the Digital Delay rack-mount, Ping-pong delay does not have provisions for feedback control.
Rotary Speaker Simulator
[alternate]+[r]
This effect automatically generates a stereo sample from a mono sound (unless additional memory is unavailable), but will utilise a stereo sample too, if one is already created. Naturally, as it is a stereo effect, the channel edit select buttons are ineffective, although it is possible to perform the actual effect on user-definable ranges.
Rotary Speaker Simulator creates an effect whereby a sound is constantly moved around the stereo image. There are two controls, depth, and rate (in Hertz), where the depth determines the degree of stereo separation (at its greatest setting, sound rotation is panned hard left or right). The rate sets how quickly the sound pans from left to right, with lower values producing much more subtle fades. As well as controlling the attributes of the speaker rotation, it's also possible to change its shape from the more usual sinusoidal waveform, to square or sawtooth, for some interesting effects.
Naturally, Squash it! cannot perform stereo conversions on existing stereo samples. Instead, there's a whole set of tools specifically designed for these, to convert them to mono samples. The advantages of swapping a sound from mono to stereo are two-fold; not only does it save half as much memory (thus allowing you to further process a sample), but you will also be able to use all of Squash it!'s tools (like the Harmoniser, and Granularise).
Mix to Mono will take both left and right channels and sum them into a mono sample. The effect is equivalent to 'mix' in the Overlay rack-mount, and so will produce little or no noticeable effect if both channels are identical. Convert to Mono will actually throw away one of the channels; a dialogue box will appear if this option is selected from the menu bar, and you may press [L] or [R] to choose the channel that you don't want.
The final option for creating a mono sample from a stereo sound is Interleave channels. This mode will create a hybrid sample, twice as long (in terms of time in seconds), and half the pitch. Effectively, left and right channel data is merged in an interleaved fashion, and this can be exceptionally useful when applied to samples that are radically different on each channel.