This instrument's sound engine works like a granular sampler, however instead of manipulating bits and pieces from a sampled sound, it uses mathematically generated "bytebeat". A very interesting thing on its own, a kind of fractal where complex structures emerge from a very simple formula. T-ape allows you to hear it, browse, zoom, cut and mangle its most interesting pieces and arrange them into sequences, arpeggios or just play directly using eight touch-sensitive keys. There are two light sensors too, adding a kind of Theremin-like functionality.
T-APE does not know frequencies, scales or octaves; however it can be mimicked by carefully tuning each grain. There are 8 touch sensitive keys so a typical scale can be constructed. To each key you can assign a selected slice of the bytebeat with all its parameters tuned using rotary encoders to where it sounds the way you want.
Arpeggiator and sequencer work with octaves to some degree in the sense that they can divide and multiply grain lengths which, at least for short grains, gives result similar to pitch shifting; the illusion of octaves breaks down at 3rd multiplication, which produces major fifth. The 4th multiplication is 2nd octave, but the 5th multiplication shifts the perceived note by a major third. Higher up the intervals progressively get out of tune.
There is tempo adjustment option but no external sync signal. Tempo parameter is stored within a patch, because sequences or arpeggio settings are part of the patch as well (so you do not need to adjust tempo when switching between patches). Light sensors can be calibrated to compensate for ambient light levels, this calibration data is stored among the patch parameters too. The only parameter that is not part of the patch is volume level.
T-APE comes with a few pre-programmed demo patches (most of which you can see in the videos below. There is space for total of 64 patches (8 banks selectable by rotary encoder and indicated by orange LED, then press one of the keys to select patch from the bank). In addition, each patch can contain 8 sequences (each up to 64 steps). You can modify and overwrite the demo patches and there is a "factory reset" command, which restores everything to defaults.
(many thanks to everyone! :)
A few buit-in example patches, sequences and arpeggios for you to play with or edit into something new
Step by step guide how to create a set of sounds from scratch, assign them to keys and arrange them into a sequence
They were recorded directly with Canon EOS 550D running Magic Lantern firmware (for adjusting the input sound levels), with 60 Ohm Koss Porta Pro headphones plugged in parallel for monitoring, using a simple 2-way splitter. No post processing was done, not even levels normalization.
Let us explain a few things about MMXX T-APE which might not be immediately obvious.
To avoid any confusion, as the word "granular" usually evokes "sampler": there is no microphone or line input. The base for these sounds is bytebeat, a sonic fractal, which can generate really wide spectrum of sounds. Some of them may sound similar to fm, wavetable or sample based synthesis - with short enough grains the difference fades, it all becomes almost the same thing. Here the pitch is lifted by merely shortening the grain, not by manipulating the waveform by a pitch-shifting algorithm.
This results in metallic sounding timbres and often ends up sounding quite melodic in faster running arpeggios or sequences. On the other end of the sonic spectrum there is an infinity of droning or, rather harsh sounding noises. By manipulating bit-rate of calculation you can explore the bytebeat at different resolution and reveal new patterns hidden inside.
The sound comes from a pair of 8-bit DAC without analog gain control, and the output signal level cannot be regulated very well. Volume control works by bit-shifting the data, so at the very low volume the least significant bits are lost, and at high volume, depending on how "busy" the sound is, clipping may occur. Ideally the T-APE should be plugged into a mixer, amp, active speaker, sound card or recorder (i.e. devices that have high input impedance).
You can use headphones too, but with certain caveats: those with higher impedance work better (headphones with 38-60 Ohms are ideal). With lower impedance headphones you may need to turn the volume up too high and then there will be more distortion in the signal. There is is a simple low-pass filter at the output but it's there rather to adjust the output impedance and protect the DAC, and it is set high enough to let the most of the spectrum through including the aliasing artifacts.
The rotary encoders do not have movable inner part like sprockets in a cassette tape, so it is not possible to insert a pencil and twist them. Only the orange ring rotates.
There is no battery included because of the shipping restrictions, however the device has a connector for standard Li-Po. It is not advisable to put in a cell with larger capacity than 250mAh, and such cell will fit neatly in the middle between the two holes. It must only be 3.7V Lithium Polymer (Li-Po) with protection circuit (like this one) and JST 2pin connector, wired by the unofficial "IoT standard": on the battery side, with the key facing up, and the holes facing you, the right side is + (red). On the PCB the positive and negative pin is marked, so please make sure that your connector wiring matches this.
Firmware updates are not likely to follow any time soon, but eventually, as we get feedback from you over time, we may release an update. In such case, you'll need to use an USB cable and PC. The process is easiest on Windows or Linux, but more problematic on Mac (requires installing drivers, and running a Python script). If you do not have access to a Windows PC, we will not be able to give you support or advice.
At the secound layer there is an 8‑bit sine, trianle & square wave piano.
You can program eight sequences with up to 64 steps for each patch.
Adjust the bitrate of the bytebeat, this can be done individually for each key.
Adjust echo delay - from short, reverb‑like timing to 1.5 seconds maximum.
Select one of the eight pre‑programmed bytebeat formulas.
Select from standard arpeggiator modes (up, down, up-down, random).
Adjust some of the variables in the bytebeat formula to futher mangle it.
Adjust the output volume - use with care when listening on headphones!
In your package you will get:
Working with bytebeat, arpeggiator, sequencer, layers, wavetable sounds and patches. Open the video on YouTube for clickable timestamps.
A showcase of directly playable sounds, arpeggios or sequences controllable by touch sensitive keys, encoders and light sensors
