Loading...
Controls RiTA through its WebSocket API (ws://<ip>:26101/api/v1/).
Requires RiTA 2.8.0 or later. With an older RiTA the module still connects and most of it works, but whatever that version's API does not have (mute groups, the average, the alignment APFs, the clear actions, change events) is simply not there, and those actions are rejected. Where this text says from RiTA 2.8.1, it means the 2.8.1 that was released: the betas of that version have some of those things and not the others, and the module says in the log when RiTA refuses one of them.
Every action waits for RiTA's answer, so a button whose actions run one after another does them in that order: the signal is set before the capture that uses it starts. A capture is the one thing that does not hold the button, because RiTA answers it at once with how long it will take and the module reports the result when that time is up. Measurement: export all, AVG: export impulse and Measurement: find delay with Live TF can take a while, and stopping the button's actions in Companion stops their wait too.
The On / Off / Toggle field of an action is a choice, not something to compute, so it has no expression mode: use two buttons, or On and Off in the same one. The Engine, Channel, Mute group and Alignment APF dropdowns do take an expression, as long as what it gives is the number written as text, 1 and not 1.0; anything else is not in the list and Companion skips the action without saying why.
When RiTA would have shown a message box, it answers anyway and puts the text in the reply; the module writes those to the log as "RiTA: …". Warnings and errors reported by a capture or an export show up there too.
Capture measures on an engine with the current signal, and turns the engine on. While a continuous measurement (Spectrum) is running, a one-shot capture is refused: stop it first with Generator: Spectrum on / off, or with Generator: set signal, since from RiTA 2.8.1 choosing another signal stops the Spectrum that was running (it takes RiTA about a second and a half, and $(rita:generator_running) goes to OFF on its own when it does). On an earlier RiTA the signal changes and Spectrum keeps running out of sight.
While a single measurement is running — Sweep, Multi Sweep, Pink or External, fired from here or from RiTA itself — RiTA 2.8.1 and later refuse Generator: set signal, Generator: set gain and Generator: set duration, and changes nothing of what the button asked: those three make the signal again or let go of the sound card the measurement is using, so it would be lost (in RiTA the same controls are greyed out while it measures). The module writes it to the log, and the button works again as soon as the measurement ends. On an earlier RiTA they went through and the measurement could come out wrong. Spectrum is not blocked: writing the signal stops it. A second Capture, a Measurement: find delay and a Measurement: export all are refused the same way while the first one is running, and the log says so.
External needs Dual Channel. It measures what comes in against the reference input, and Single Channel (RiTA's Preferences > Transfer) has no reference input, so from RiTA 2.8.2 External is not in RiTA's signal list while that mode is on: Generator: set signal with External is refused, nothing is changed, and the module says in the log that External only exists with Dual Channel. The button is still right, it just has to wait for the mode, and the module can tell you when: that same RiTA answers the Transfer mode in $(rita:settings_transfer_mode) and in the Transfer mode is feedback, which RiTA only lets a client read. The ready-made Signal: External button uses it and goes grey while Single Channel is on, so you can see at a glance that it is not available; on RiTA 2.8.1 and earlier the variable stays empty and the button simply never greys. Two more things that mode does on RiTA's side, and the module simply follows them through the change events: switching to Single Channel while External is selected moves the signal to Sweep, and going back to Dual Channel leaves it on Sweep — External returns to the list but is not chosen again by itself; and a configuration or a project saved with both opens on Sweep. With an earlier RiTA, External is always there.
Spectrum on/off selects Spectrum if needed and measures it on the chosen engine; off stops it. It can run on several engines: deactivate one with Measurement: activate engine.
From RiTA 2.8.1 the Spectrum signal has no 64 FPPO FFT size, nor 128 FPPO or 65536: choosing Spectrum moves the size to 32768 on its own, writing one of those while it is selected is refused, and leaving Spectrum brings the size back. The module reads the settings again after every change of signal, so what the buttons show is what RiTA has.
Settings: Auto FFT size on / off is the Normal / Auto FFT Size switch of Preferences, off by default. With it on, each engine keeps the next power of two above the length of the sweep, from 65536 to 262144 samples, so a longer sweep gives a longer engine: $(rita:settings_engine_samples) is that size in samples and $(rita:settings_time_constant) the same thing in seconds, which depends on $(rita:settings_sample_rate). The FFT size list then also offers those sizes and 128 FPPO, and the delay of a DSP channel reaches as far as the engine lasts, that same number of seconds: 680 ms for every 32768 samples at 48 kHz (5440 with 262144), half of that at 96 kHz. Off, each engine goes back to 32768 and the measurements are cut to it, with RiTA saying so in the log.
Because the engine follows the sweep, Generator: set duration changes the engine size too, and the module reads the settings again after it. $(rita:settings_fft_size_auto) says whether the switch is on.
The FFT size dropdown offers every size RiTA can ever have, including the big ones. Asking for one RiTA is not offering at that moment is refused, and the module says in the log which of the reasons it is: the switch being off, the size being above the engine of the moment, the Spectrum signal, or a find in place.
Live TF is not offered by RiTA 2.8.0: it is not in the signal list and the generator has no pink noise switch.
Measurement: set inputs sets the inputs of that engine. With Single Channel in Preferences > Transfer and a Sweep, Multi Sweep or Pink signal, RiTA greys out its own reference dropdowns, but the reference input is still writable from here: RiTA stores it and it counts again as soon as the mode goes back to Dual Channel. In 1 Ref. Channel mode the reference input goes to all eight engines, and in 1 Meas. Channel mode so does the measurement input: RiTA says which in its reply and the module writes it to the log.
Reference mode and measurement mode are the two channel modes of RiTA's Preferences. With Several, each engine has its own input. With 1 Ref. Channel or 1 Meas. Channel the eight engines share that input, so writing it on one engine writes it on all of them. RiTA puts itself in 1 Meas. Channel when a two-input sound card is chosen. In that mode turning one engine on turns the others off, in RiTA and from here alike, and Measurement: activate engine writes to the log which ones were turned off; with the Spectrum signal several can still be on. The module reads the eight engines again after a change of mode. Both modes are in RiTA 2.8.0: on an earlier RiTA (a 2.8.0 beta or before) the two actions are refused and the variables stay empty.
EQ filters start disabled in RiTA: a filter that is not enabled is stored but does not sound. Gain applies to Parametric, the shelving types and the U-Shaping bands, order to APF, FIR RevPhase and the U-Shaping bands.
With DSP type RiTA or Kingray there are five filter types: Parametric, Low Shelf, High Shelf, APF and FIR RevPhase. A Marani adds its own, up to 21 in the dropdown, and which of them exist depends on the model and its mode, so one that model does not have is refused: Bell Adaptive Q, Bell Classic Q, the shelvings 1st, 2nd and Slope, Low and High Pass 1st, 2nd and VariQ, Band Pass and Notch. The ones that already existed keep their name, which is RiTA's own and not always the one on the Marani screen: the Marani bell is Parametric, its Hi-/Lo-Shelving VariQ are High Shelf and Low Shelf, and its All Pass is APF. The dropdown also takes a typed name.
Each type uses only some of the fields, and the module sends only those: the gain goes to the bells, the shelvings and the U-Shaping bands, the order only to APF, FIR RevPhase and the U-Shaping bands, and the Q to the bells, the VariQ shelvings and cuts, Band Pass and Notch. On the two Slope shelvings the Q field is not a Q but the slope in dB/oct, 6 to 12. The 1st and 2nd shelvings and cuts use neither Q nor order. With a Marani the Q reaches that model's table, up to 128 (the Notch, 4 to 104), which is why the field goes beyond the 10 of RiTA's own filters.
With DSP type Galaxy, from RiTA 2.8.1, there are only three types and the filters are fixed: a channel has 16 of them, 1 to 10 Parametric, 11 to 15 the five U-Shaping bands and 16 the APF, and the number of bands cannot be changed either. Writing on a filter a type other than the one it has is refused, so the type to choose is the one that filter already is. A U-Shaping band is a cut with a slope, and the order is that slope, 6 dB/oct per step: 1 to 8 on bands 1 and 2, 1 to 6 on bands 3 and 4. Its gain is 15 dB either way and its frequency takes less than the field offers: 20 to 2500 Hz on band 1, 40 to 5000 on band 2, 80 to 10000 on band 3 and 160 to 20000 on band 4. A band has no Q, so the field is not sent. Band 5, which is filter 15, only takes the gain: the Galaxy has neither a fifth cut nor a fifth slope, so the module sends neither frequency nor order on it. The cuts do not cross each other: one moved past its neighbour pushes it, and it stays 1 Hz away, so moving one band can move the one next to it.
Each channel also has a number of bands in use, the Bands dropdown of RiTA's Equalizer window, 20 by default and 16 with a Galaxy. From RiTA 2.8.1 a filter above that number cannot be switched on: RiTA refuses it and the filter stays off, while its frequency, type, gain and Q can still be written. The API does not report the band count and RiTA answers the same thing as for a linked channel, so the module writes to the log every reason that fits what was asked.
In the crossovers (high-pass and low-pass) the last number is not a Q: it is the passband ripple in dB on Chebyshev I and Eliptic, the stopband attenuation in dB on Chebyshev II, and Butterworth, Linkwitz-Riley and Bessel ignore it. It takes 0.1 to 120.
From RiTA 2.8.1, with a processor as DSP type (a Marani, the Kingray DAP0408F or the Galaxy) RiTA checks the crossovers against the model and its mode, the same as its own screen: a type the mode does not have, or an order it does not have, is refused and the module says so in the log. It is no longer moved to the nearest one. With DSP type RiTA nothing changes.
From RiTA 2.8.1 the crossovers have one more type, NXF, on the Marani models and modes that have it (MIR FI, A and D, VA DF and MDA kDM; not the MIR E, the kTM or the PDA, and not in the XOVER-IIR + FIR + 4 PEQ modes of the MIR); anywhere else it is refused. On an NXF the order is not dB/oct but an attenuation code from 1 to 8 (−40, −45 … −70 and below −75 dB). With NXF on one cut the other can only be None or NXF: RiTA accepts whatever is written, and when the DSP is sent to the processor the other cut goes as None and says so in the log. Both are numbers of RiTA's own filter model: with another DSP type they become that processor's numbers, the ones its own screen shows.
DSP type is the DSP Type of RiTA's Preferences > DSP: the filter model RiTA draws and computes its DSP with, its own or the one of an external processor (Kingray, Marani, Galaxy). Changing it makes RiTA convert the filters of the eight channels so the curves do not change, which moves the numbers: the Q of the Parametric and shelving filters and the frequency of the Bessel crossovers (a +12 dB bell with Q 2 becomes Q 3.99 in Kingray). What does not fit the new model is left at that processor's limit, so a +18 dB bell becomes +15 dB in Kingray, and RiTA's warning about it shows up in the log as "RiTA: …". The module writes the change to the log and reads the channels again, and it does the same when the type is changed in RiTA itself. Anything read before the change is a number of the old model.
DSP type is in RiTA 2.8.0. On an earlier RiTA (a 2.8.0 beta or before) the action is refused with "This RiTA has no DSP Type" and $(rita:settings_dsp_type) stays empty; everything else works as usual. The dropdown also takes a typed name, for a newer RiTA that knows more processors.
DSP model and DSP mode, from RiTA 2.8.1, go with it. The model is the Marani of Preferences > DSP (MIR FI, MIR A, MIR D, MIR E, MDA kDM, MDA kTM, PDA and VA DF); setting it puts RiTA in Marani with that model. The mode is the processor channel RiTA models, IN or OUT and, on a Marani, its EQ and crossover mode, with the same labels as the Marani software, for example OUT: XOVER-IIR + 8 PEQ. Kingray and Galaxy have an IN and an OUT mode; with DSP type RiTA there are no modes and both variables are empty.
The mode decides how much of the channel there is: how many EQ filters, which crossovers and slopes, and the FIR. Because the labels depend on the model, there are more than forty of them and the API cannot list them, the mode is typed in by hand: $(rita:settings_dsp_mode) shows the one in use, and a label that is not of that model is refused. Settings: DSP model has a mode field of its own, which sends both in one go, the way RiTA checks them. Changing the model switches off the filters whose type the new model does not have, and RiTA says which in the log.
Alignment APFs are the 2 all-pass filters per channel that RiTA's Auto Align writes (it replaces both on every run), separate from the 20 EQ filters. They can also be set by hand: frequency, order (1 or 2) and Q. With DSP type Galaxy an all-pass is always of order 2, so from RiTA 2.8.1 a 1 is not refused there: it is stored as a 2, and that is what the variable then says. The module follows them, so buttons update after an Auto Align. Older RiTA versions do not have them.
Mute groups are the 4 assignable mute buttons of the external processor (DAP0408F, GALAXY, Marani…). Which channels belong to each group is set in RiTA, not from here, and a group can mix channels of several processors. Toggle does what the button does: it unmutes the group when every channel is muted, and mutes it otherwise. Feedbacks: Mute group is muted (red) and Mute group has channels; the name given in RiTA is in $(rita:mute_group_N_name), so it can be the button text. A group with no channels on a connected processor answers "no channels", and a channel the processor did not take is written to the log.
AVG is RiTA's average of the selected engines (not the active ones). What selects one matters, and RiTA's own window does it by itself: touching a channel's EQ selects its engine, so that engine joins the average. Ticking a filter, the Chingonizer or an alignment APF already did it, and from RiTA 2.8.2 so does designing, ticking or importing an X-Over FIR. No action of the module selects an engine, so an average that changes contents on its own was changed from RiTA; the Measurement engine is selected feedback follows it, so a button can show which engines the average is using. AVG: on / off is the same control as Settings: averaging on/off. From RiTA 2.8.1 AVG and Sum exclude each other, like their buttons in RiTA: switching one on switches the other off, and the module follows both. AVG: export impulse writes the impulse response of the average to the Memory Bank folder, in the format chosen in RiTA; the export runs in the background and its result (or error, such as no export folder or AVG off) is written to the log.
Measurement: rename engine and DSP: rename channel write the same field, because in RiTA an engine and its DSP channel have one name: $(rita:meas_N_name) and $(rita:dsp_N_name) always show the same thing. From RiTA 2.8.1 a name another engine already has is not repeated: RiTA gives it the first free one, Bajo and then Bajo_1, and because RiTA answers with how the field ended up, the buttons show the real name at once and the log says what RiTA kept. An empty name, and Name, can still be repeated.
Memory: store trace from engine is RiTA's Snapshot of one engine: it saves the trace the engine has now as a row of the Shot table, with the engine's name. From RiTA 2.8.1 a name that is already in the table takes the first free one, Main and then Main_1, Main_2, so the memory may not be called like the engine. The module reads the table back after storing and writes to the log which row it is and what it ended up called, because that name is what Memory: show / hide, Memory: rename and Memory: delete take, along with the row number; asked for by name, the first row with that name is the one that answers. Renaming two rows to the same thing by hand is still allowed, in RiTA and from here.
Measurement: sync all is RiTA's Sync All (button and Y shortcut): it aligns with each other the engines that already have a measurement, without touching AVG. The new delays show up on the buttons at once and are written to the log. With no measured engine it answers "no active measurements".
RiTA does not send events for the sync, so the module reads it with the level meters, at the poll interval: the feedback Sync All is set stays on while a sync is in place, Sync All is possible says whether any engine has a measurement, and a Sync All done from RiTA writes "Sync All done in RiTA" to the log.
While a sync is in place, RiTA does nothing on a new Sync All (in RiTA the button is disabled): the module says so in the log. Put the feedback Sync All is set on the button to see it at a glance.
Measurement: clear find removes the sync. Measurement: clear delays sets the DSP delay of the 8 channels to 0, and RiTA refuses it while a sync is in place: clear the find first. Measurement: clear everything is RiTA's Clear with Everything and with no dialog, so it wipes measurements, DSP and names of every channel.
Each engine has two delays: meas_N_delay is the arrival of its measurement, and meas_N_dsp_delay is the delay of its DSP channel, which is what Sync All writes.
Measurement: export all is RiTA's Export All: it exports the selected engines (the engine buttons, selected in the API) to the project folder, in the format chosen in RiTA. The export runs in the background; its result (or error, such as no engine selected or no export folder) is written to the log.
Settings: export position is what goes after the name of every exported file, <engine>_<position> and <name>_<position> for the AVG impulse. From RiTA 2.8.1 RiTA has no Position field on screen and this is the only way to set it: it stays until it is changed, empty means no position, it takes up to 60 characters, and whatever a file name cannot have becomes a dash. The field accepts variables, so a button can stamp the position it belongs to, and $(rita:settings_export_position) shows the one in use. RiTA cleans the text up, and the module writes to the log what it really kept.
The Settings variables follow RiTA and are never put back by the module: if RiTA changes one by itself, that is what the buttons show.
Linked channels (Link DSP): in RiTA a channel can follow another one. The linked parts of the slave channel (gain, delay, polarity, crossovers, EQ filters, alignment APFs and FIRs, each group on its own) cannot be written: those actions do nothing and write "read only, this part of the channel is linked to another one in RiTA" to the log. On an EQ filter RiTA answers the same for being above the bands in use and for the fixed types of a Galaxy, so there the log names every reason that fits what was asked. Rename still works. The API does not say which channel is the master.
DSP: clear channel is the exception: on a linked channel it undoes the link and then resets the channel, exactly as RiTA's own Clear button does, with no confirmation.
DSP: clear channel is the Clear button of the row: it also clears that engine measurement, and on a linked channel it undoes the link first.
Connected, Mute group is muted, Mute group has channels, Sync All is set, Sync All is possible, AVG on, AVG has a curve, generator running, generator signal, engine active, engine selected, DSP type is, Reference mode is, Measurement mode is, Transfer mode is, DSP polarity inverted, DSP alignment APF enabled.
$(rita:generator_running), generator_signal, generator_gain, generator_duration, generator_output1, generator_output2$(rita:settings_fft_size), settings_window, settings_smoothing, settings_spectrum_averages, settings_averaging, settings_sum, settings_coherence_threshold, settings_dsp_type, settings_dsp_model, settings_dsp_mode, settings_fft_size_auto, settings_engine_samples, settings_time_constant, settings_sample_rate, settings_reference_mode, settings_measurement_mode, settings_transfer_mode$(rita:sync_active), sync_count, sync_possible$(rita:mute_group_N_name), mute_group_N_muted, mute_group_N_available for N = 1..4$(rita:average_active), average_has_data, average_count, average_engines, average_mode, average_name$(rita:dsp_N_name), dsp_N_gain, dsp_N_delay, dsp_N_polarity for N = 1..8$(rita:dsp_N_apfK_enabled), dsp_N_apfK_frequency, dsp_N_apfK_order, dsp_N_apfK_q for N = 1..8 and K = 1..2$(rita:meas_N_name), meas_N_active, meas_N_delay, meas_N_dsp_delay, meas_N_level for N = 1..8In Companion 5 the text shown on a button is set in the button's Style tab, Text element, Button text string.