You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
New: Added support for the Kurzweil PC3 series, Forte, Forte SE, PC4 and K2700 (PC3, P3K, K2P, P3A, PLE, FOR, FSE, PC4, P4S, K27): the programs of these object files are read with their keymaps, user samples, envelopes, velocity tracking and filters, in the sample layout of the PC3K as well as in the one of the Forte generation; files of the PC3K (P3K), Forte and Forte SE (FOR, FSE), PC4 and PC4 SE (PC4, P4S) and K2700 (K27) are written. Written files are not yet verified on hardware.
New: Added support for the Groove Synthesis 3rd Wave (PGDATA, BIN): a unified program file is read as one instrument with the level, tuning, panning and keyboard range of the parts which play its samples (e.g. stereo, split and layered programs), a multi-sample file as its sample slot; unified program files, whose parts layer and split the samples, and multi-sample files are written. Written files are not yet verified on hardware.
New: Added support for LFO modulation with modulation wheel (CC 1): Akai AKP, Akai S900/S950, CWITEC TX16Wx, discoDSP Bliss, Decent Sampler, DLS, Kontakt, Kurzweil K2000/K2500/K2600, Kurzweil PC3/Forte, NI Maschine, Polyend Tracker, Reason NN-XT, Roland S-10 Series, Roland S-7xx, SFZ, Soundfont 2, Spectrasonics Omnisphere, Synthstrom Deluge, TAL Sampler, Waldorf QPAT, Yamaha YSFC (Montage, MODX)
New: A note reports when the samples of a multi-sample sound away from the root keys they are mapped to - e.g. an octave, because the instrument was transposed when it was recorded - and the Transpose processing option set to 'Auto' (-Zpa on the command line) moves the root notes of each multi-sample onto its measured pitch. Only the root notes are moved, the audio and the key ranges stay as they are. A multi-sample of only one or two zones - typically one recording stretched across the keyboard - has no other zones to agree with and is only trusted with a whole number of octaves: a root key set in another octave than the recording is the common case, while a chord or an inharmonic sound can measure as any interval. A chord of two or three notes - e.g. a fifth pad or a dyad of two synthesizer notes - is measured at its root note instead of at the common fundamental of its notes, which lies an octave or more below them and does not sound.
New: The Process dialog lists the loop options in the order in which they are applied: 'Snap loops which click' first, then 'Cross-fade loops' (formerly 'Set fixed loop-crossfade') with 'Only loops which still click' (formerly 'Only loops which click'). Together they snap the loops which click and cross-fade only the loops which snapping could not fix, which the dialog did not show. 'Only loops which still click' is available when a cross-fade above 0% is selected, since it does nothing without one. The log counts the loops which snapping 'moved' instead of 'adjusted' - a moved loop can still click - and the note about loops which still click after snapping names 'Only loops which still click' instead of the loop cross-fade, which cross-fades all loops. When loops still click although they were cross-faded, the note says that they start at the beginning of their sample, where there is no audio for a cross-fade.
New: A loop only counts as clicking - for the note about clicking loops and for 'Only loops which still click' - if its loop point also breaks the curve of the waveform, which is how 'Snap loops which click' decides whether a loop clicks. The jump at the loop point alone also counted loops which wrap seamlessly through a steep part of the waveform, e.g. at the edge of a square or pulse wave, where the waveform moves that far from one frame to the next anyway. The snapping left such loops alone as clean, and 'Only loops which still click' then cross-faded them.
Improved: The note about a preset which plays its samples as recorded far away from the middle of the keyboard names the setting of the Transpose processing option which plays them as recorded closest to the middle, in whole octaves, which keep the keys in tune. The note said that the option can move the preset, which suggested 'Auto' as well, but 'Auto' only moves root keys onto the pitch measured in their samples, not towards the middle of the keyboard. A multi-sample which the processing transposed no longer gets the note, since it no longer plays as the source maps it.
Improved: The pitch of the samples - for the Transpose processing option set to 'Auto' and the note about it - is measured in the loop of a sample also when the loop ends on the last frame of the sample, which was skipped, and in the loudest part after the attack when the loop has no clear pitch. Noise in a sample no longer decides between a pitch and the octave below it: the octave above a candidate is the pitch whenever the odd harmonics, which it does not explain, hold less than 5% of the power on its harmonics.
Fixed: Loops played the wrong audio after the loop point when the samples were resampled (e.g. -Zf 22050) and several zones referenced the same sample, e.g. E-mu Emulator III presets using the chorus option, TX16Wx regions sharing a wave, Ensoniq Mirage presets with both oscillators active.
Fixed: 'Snap loops which click' could move a loop which does not click to a position which does. A loop whose last frame repeats its first frame bends a little at the loop point, which the snapping counted as a click, and the nearest zero-crossings of a pulse wave can be half of its level apart: a loop of the 'Prophet-5 Fat PWM' presets of the Hollowsun TAL-Sampler library, which repeats seamlessly, was moved by 4 frames to a position which jumps by 45% of the level at every repeat. A position which jumps at the loop point is not taken anymore for a loop which does not jump there.
Fixed: Reducing the sample frequency with the processing options clipped samples which peak close to full scale. Such a sample is now lowered to fit and its zone plays it louder by the same amount.
Fixed: The note about loops which click reported loops with a cross-fade of less than 1% of the loop, while 'Snap loops which click' leaves every loop with a cross-fade alone, since the cross-fade already smooths the wrap. The note then named loops which the option had skipped - '0 adjusted' - and recommended the option, and 'Only loops which click' replaced the cross-fade of these loops with its own. A loop with a cross-fade is now never reported, whatever its length, and when the loops were snapped, the note says that the loops it names still click instead of recommending the option. The converted audio and loops do not change, and 'Only loops which click' keeps the cross-fades of the source.
Fixed: The note about loops which click reported many loops of 24 and 32 bit samples which do not click, and the loop cross-fade for the loops which click was set on them. A jump at the loop point counts as a click when it is many times the normal movement of the waveform around it, but that movement was computed with integers, which overflow for loud 24 and 32 bit audio, and turned negative: every jump above 2% of the level counted.
Fixed: 8 bit samples were converted to full-scale noise for destinations which need samples with more bits, e.g. 16 bit for Waldorf Quantum/Iridium (QPAT) and the Teenage Engineering OP-XY - samples from 8 bit WAV and AIFF files as well as e.g. the voices of the Fairlight CMI IIx. WAV stores 8 bit samples unsigned and all others signed, but the conversion kept the unsigned format of the source for the 16 bit samples, which are then read as signed ones: every sample is moved by half of the full scale and the waveform wraps around at each zero crossing. The 6 CMI IIx test voices wrapped 255 to 4,864 times per sample when converted to QPAT, and the 48 converted 8 bit samples of the 'RobotA/B 9 octaves' presets of the Iridium MK2 factory library, written for the OP-XY, were shifted by up to 30% of the full scale and no longer followed their source; now they match it within one 8 bit step. Destinations which keep the 8 bits, e.g. SFZ, were not affected.
Fixed: A WAV file which ends before its audio does, e.g. because a copy was interrupted, stopped the conversion of its multi-sample when processing was enabled. Without processing, the file was copied as it is, with a header which states audio that is not there, and the zone ended behind the audio. The audio which is present is now used, faded out over its last 10 ms so that it does not end with a click, and written with a header which matches it; a warning names the file and how many of its sample frames are present. The play range of a zone ends with the audio, a zone which only plays the missing part is left out, and a loop which reaches into the faded out or missing part is removed, since it would jump from silence back to the loop start at every repeat. Reading such a file also dropped up to 16 KB of the audio in front of its end.
Fixed: A WAV file which the check of a loaded sample rejects - one with more than two channels, without a format chunk or without audio data - is left out, as the logged error says. The result of the check was ignored, so the file was loaded anyway, e.g. a file with 4 channels was written into a converted QPAT preset, although the instruments play only mono and stereo samples.
Fixed: WAV files with a MIDI unity note in the wrong byte order (e.g. written by the Akai MPC500/1000/2500) created huge root keys (e.g. pitch_keycenter=1006632960).
1010music bento / blackbox
New: Read/write the LFO and its routing to Level, Pitch and Filter Cutoff incl. the modulation wheel depth. Tempo synchronized rates are converted assuming 120 BPM.
New: Read/write the polyphony (Poly Mode: Mono, Poly 2, 4, 6, 8, X). When writing, the smallest Poly Mode which can play the polyphony of the source is selected, e.g. Poly 4 for 3 voices; Poly X is used if the source has no polyphony set.
Fixed: bento - The loop crossfade (loopfadeamt) was not read.
Ableton
New: Read/write the 3 LFOs and their routing to Volume, Pitch and Filter Cutoff incl. the modulation wheel. Tempo synchronized rates are converted assuming 120 BPM.
New: Samples of the Core Library which comes with Live are found for presets, Live Sets and Live Packs, also for presets which were copied out of the installation: in the Core Library which contains the source, otherwise in the one of Live installed in its standard location. They were searched like the samples of a project, relative to the preset.
New: Live Packs (ALP) are read: the Sampler and Simpler devices of all presets and Live Sets in a pack are converted with the samples from the pack, without unpacking it in Live first. The samples, which a pack stores compressed with FLAC, are decoded when they are read. Packs whose samples are encrypted (*.eflac) cannot be converted.
New: Live Sets (ALS) are read: every Sampler and Simpler on the tracks of a Live Set, also in racks, becomes a multi-sample named after the Live Set and its track.
Improved: AIFF samples which are encrypted (AIFC compression type 'able'), e.g. some of the samples which come with Live, are reported as such instead of as compressed.
Fixed: Samples were not found when the presets were not written to the root of the User Library. Samples are now referenced relative to the preset (#498).
Fixed: Presets which are stored as plain XML instead of compressed XML could not be read ('Not in GZIP format'), e.g. all 23 drum kits of the '80s hits' pack of Subsocial Studios.
Fixed: The samples of presets and Live Sets which were saved with Live 10 or earlier were not found, since these versions store the path of a sample as a list of folders with the name of the file separately.
Akai AKP/AKM (S5000/S6000/Z4/Z8/MPC4000)
New: The transpose and fine tune of AKM parts are applied.
New: AKM files written by a Z4/Z8 use the more exact Z4/Z8 part settings (level in 0.1dB, tune in cents).
New: Performances can be written as AKM multis (up to 32 parts) together with their programs and samples.
New: AKP programs are written (destination 'Akai AKP', on the command line -d s5000), in one of the 3 file format generations of the S5000/S6000 operating systems (generation 3 is also used by the Z4/Z8). Each program is stored in a folder together with its samples. Key ranges, velocity zones, tuning, panning, level, playback mode, the envelopes, the filter, the pitch envelope and the LFOs are converted.
New: The keyboard tracking of the zones, the mute group of the key-groups and the filter offset of the zones are converted.
New: The velocity and keyboard scaling of the envelope times (Velo->Attack, Keyscale, Vel->Rate 1, Keyboard->R2/R4) are converted.
New: The filter modes BI-NOTCH (as a notch) and PEAK 1-3, WIDE PEAK and BI-PEAK (as a band-pass) are converted.
New: The two LFOs are converted as modulators of the pitch (vibrato), the volume (tremolo) and the filter cutoff, depending on the modulation sources of the program.
New: The amplitude, filter and auxiliary envelopes are converted as pitch and filter cutoff modulators, depending on the modulation sources of the program. Before, only the auxiliary envelope was converted as a pitch modulator.
Improved: Zones routed to a mono output are no longer panned.
Improved: A fully opened low-pass filter without any effect is no longer added to the converted zones.
Fixed: The panning of AKM parts was read from the wrong position.
Fixed: Samples of programs created by PC software (e.g. the Roland XV library) which contain the playback value 6 (AS SAMPLE) in version 1 zones were not looped.
Fixed: Files written by a S5000/S6000 were reported as "Z4/Z8" when the RIFF size was set. Z4/Z8 files are now detected by their 'z48 ' chunk.
Fixed: The pitch envelope modulation via 'Pitch Mod 2' was not detected.
Fixed: Programs which use the key-group or zone tuning to transpose samples by full semi-tones (e.g. +12) now move the root key instead of keeping a large pitch offset.
Fixed: The auxiliary (pitch) envelope stages were shifted by one stage.
Fixed: Playback mode LOOP UNTIL REL is now converted as a loop which ends on note-off; unknown playback values no longer create a loop.
Fixed: Envelope times were much too long, they now use the curve of the S1000/S3000.
Fixed: Samples written by an Akai sampler with a sample tuning were mapped 2 octaves too high since the root key was read from the generic part of the WAV sample chunk instead of the Akai specific data.
Akai MPC60
Fixed: Pads without a decay were silent, the decay now ends at the end of the sample.
Akai MPC500/1000/2500
New: Read the LFO of the pads (MPC1000 OS 2 and MPC2500): waveform, rate and delay (converted at 120 BPM) and the amounts for pitch, level and filter cutoff.
New: Read the filter envelope (time and amount) of filter 1.
Fixed: Programs saved by the MPC2500 could not be read ("MPC1000 PGM 1.01").
Fixed: Velocity to level was applied as the depth of the amplitude envelope, pads with a value of 0 had no envelope.
Fixed: Empty pads could cause 'Referenced sample missing' errors (e.g. in the MPC1000 factory programs).
Fixed: Pads with decay mode 'Start' were silent
Akai MPC2000/2000XL/3000
New: Voice overlap 'Note Off' is converted (stop on note-off), all other pads are one-shot.
New: Read the filter envelope (attack, decay, amount).
Fixed: The start and end of SND sounds were ignored.
Fixed: Pads with decay mode 'Start' were silent, decay mode 'End' faded out directly after the attack.
Fixed: Velocity to level was applied as the depth of the amplitude envelope.
Akai MPC/Force
New: Read and write the key-group LFO (Sine, Triangle, S&H, Saw, Saw Down, Square, Noise) with its rate, reset and the amounts for pitch, volume and filter cutoff. The 'Wheel > LFO' setting is converted to/from modulation wheel control of the LFO amounts.
Akai S1000/S3000
New: Read LFO 1 (vibrato with depth, mod wheel, rate, delay; S3000: waveform) and the assignable S3000 LFO 1/2 modulations of pitch, loudness (tremolo) and filter cutoff.
Fixed: The program parameters after the key temperament were read 1 byte off.
Audiomodern Soundbox
New: Read/write LFO modulation (tremolo, vibrato, filter cutoff) from/to the 4 LFO engines of the Modulation tab (shape, rate, sync, retrigger). Assignments to the layer volume, fine tune/transpose/octave and the filter frequency of the filter effect are converted.
Fixed: The panning of the samples was not converted, all samples played in the center. E.g. split stereo samples - two mono samples of one recording, hard panned left and right - played layered in the center instead of in stereo. The panning is now written to the value which the sample editor of the plug-in sets, for every sample, and read from it.
discoDSP Bliss
New: Added support for reading/writing version 3.8 - 3.20 features (round-robin, one-shot). Writing now always uses version 3.20.
Fixed: Read version is now logged in decimal not hexadecimal.
Fixed: Samples with a sample rate other than 44.1 kHz played out of tune, since the sample rate of a zone was not written and Bliss assumes 44.1 kHz.
Fixed: The amount of a filter cutoff envelope routed in the first envelope destination slot was read from the second slot.
Fixed: The pitch bend range is now written (the larger one of up and down, since Bliss has only one).
CWITEC TX16Wx
New: Read/write the 2 LFOs routed by the modulation table to the pitch, the volume and the filter frequency incl. the modulation wheel as the via source.
DecentSampler
New: The stored values of the controls of the user interface are applied, as DecentSampler does when it loads a preset.
New: One-shots, exclusive groups (silencedByTags), disabled tags, tag volumes and gain effects are read. A message lists the features of a preset which cannot be converted.
New: One-shots and exclusive groups are written.
New: The user interface template (ui.xml) was updated. The template's filter controls no longer overwrite the cutoff, resonance and filter envelope of the converted preset. Its mod wheel vibrato is now removed if the source has its own mod wheel vibrato. If you use your own copy of the template, recreate it with Create templates or add these changes to your copy:
Filter controls: set the values of the Cutoff, Resonance, Filter Attack, Decay, Sustain, Release and Env. Amount knobs to %FILTER_CUTOFF_VALUE%, %FILTER_RESONANCE_VALUE%, %FILTER_ENV_ATTACK_VALUE%, %FILTER_ENV_DECAY_VALUE%, %FILTER_ENV_SUSTAIN_VALUE%, %FILTER_ENV_RELEASE_VALUE% and %FILTER_ENV_AMOUNT_VALUE%.
Filter envelope bindings: replace modulatorIndex="0" with modulatorTags="filterEnvelope".
Mod wheel vibrato: add tags="modWheelVibrato" to its LFO and replace position="1" in its CC 1 binding with modulatorTags="modWheelVibrato".
Fixed: The settings of a group or of all groups were not used for their samples (e.g. velocity ranges), the pan of a group was added to the pan of its samples, and the volume of all groups was ignored.
Fixed: Envelope curves which are not set were linear instead of the default curves of DecentSampler.
Fixed: A linear envelope curve was not written, so DecentSampler played its default curve.
Fixed: An envelope which is the same in all groups was written to each sample, which left the envelope knobs of the user interface without effect.
Fixed: With 'Create folder structure', the presets of a dsbundle were written into an output folder named like the bundle, which the Finder shows as a bundle file again. A dsbundle is a folder which DecentSampler treats as one file, like a dslibrary, whose presets already went into the folder which contains the library file. The presets of a bundle now go into the folder which contains the bundle as well, also when the bundle is the selected source folder or has sub-folders. The name of the bundle is still used for the detection of the metadata like any folder name, and the Contents dialog still shows the bundle as the folder it is.
Fixed: A preset which sets the same attribute twice on one element could not be read: the whole preset was skipped with 'Could not parse XML ... Attribute "textColor" was already specified for element "labeled-knob"', while DecentSampler loads it. A repeated attribute is now ignored and its first value used. The log names the attribute and the element when the two values differ and the attribute is one which the conversion uses.
DLS
New: The filter (DLS Level 2) is read: a 2-pole low-pass with cutoff, resonance (Q), filter envelope (EG2), velocity to cutoff, key tracking and the LFO incl. its modulation wheel depth. A filter which is disabled (e.g. most presets of the Roland GM2 set) or fully open is left out.
Fixed: The articulations of DLS Level 2 files (art2 chunks in lar2 lists) were not read, so their envelopes, LFO, velocity, tuning and filters were missing; only DLS Level 1 articulations were used.
Fixed: A pitch bend range was read 65536 times too large (e.g. 13,107,200 instead of 200 cents), since the fixed point value of the connection was not converted to cents.
E-mu Emulator II
New: The amplitude envelope (attack, decay, sustain, release, velocity to attack), the filter (cutoff, Q, keyboard tracking, velocity to cutoff) and the filter envelope (amount, direction, attack, decay, sustain, release, velocity to attack) of a voice are written; before, every voice got these settings of the template voice, an unfiltered piano with a long decay. The envelope times are rounded to the 31 times the sampler uses (0.01 to 15.3 s); the sustain of the filter envelope is an approximation.
New: The LFO of a voice is written (rate, delay, vibrato, tremolo and filter depth); before, every voice got the settings of the template voice without an LFO.
New: The modulation wheel assignment of a preset to LFO to pitch, filter or level is read as a modulation wheel controlled LFO depth on the voices without an LFO amount of their own, and written for zones whose LFO depth is only controlled by the modulation wheel. Written presets no longer always assign the right wheel to the vibrato.
E-mu Emax / Emax II
New: The modulation wheel assignment of a preset to LFO to pitch, filter or level is read as a modulation wheel controlled LFO depth, which overrides the LFO amount of the voices, and written for zones with a modulation wheel controlled LFO depth.
E-mu Emulator III / IIIX / ESI
New: The modulation wheel assignments of a preset to LFO->Pitch, LFO->Cutoff and LFO->VCA are read as a modulation wheel controlled LFO depth; it overrides the LFO amount of the zone, as the factory banks rely on.
Fixed: Written presets routed the modulation wheel to the vibrato of all zones, which overrode a converted vibrato so that it only sounded with the wheel up; the wheel is now only assigned if the source has a modulation wheel depth, and zones with a fixed vibrato exempt themselves from it. The last 3 bytes of the real-time controls are written like in the factory banks; they were shifted by one.
E-mu Emulator IV
New: Read/write the pitch bend range (the pitch wheel to pitch cord, in whole semitones) and the pitch envelope (the auxiliary envelope with its cord to the pitch). Before, every zone got the default bend range of +/-2 semitones and no pitch envelope when reading, and the bend range and pitch envelope of the source were lost when writing.
New: The centered key source (key ~) to the filter cutoff is read as key tracking as well, e.g. by 82 voices of the ORBIT bank; only the key + source was read before.
New: Read/write the LFOs (rate, shape, delay, key sync) and their modulation cords to the amplifier volume, the pitch and the filter frequency (tremolo, vibrato, cutoff LFO) incl. the modulation wheel control of the LFO depth.
E-mu Emulator X
New: Read/write the LFOs (rate, shape, delay, key sync) and their modulation cords to the amplifier volume, the pitch and the filter frequency (tremolo, vibrato, cutoff LFO) incl. the modulation wheel control of the LFO depth.
New: Read/write the filter resonance (MIDI controller B and its initial value, which is how the banks set it), the filter key tracking (the key to cutoff cords) and the pitch envelope (the auxiliary envelope with its cord to the pitch). The resonance was always 0 before.
New: Banks which store all of their samples in a single sample pool file - 'SamplePool/.ebl' instead of one file per sample - are read.
Fixed: The banks of the ZERO-G library could not be read, each one reported 'No presets with sample zones found'. Their voice lists claim 92 bytes more than the preset holds, and such a list was dropped together with all of its voices. A chunk which runs past the end of its parent is now cut there, which reads all 378 presets of the 68 banks; all other banks are read exactly as before.
Fixed: Samples which store their audio as 32 bit floating-point values were read as 16 bit audio and sounded like loud noise, twice as long as the sample. The Emulator X stores every sample with a resolution above 16 bit in this format, e.g. the ones imported from 24 or 32 bit WAV files. Such samples are now read as floating-point audio with their loops in place; they stay 32 bit floating-point where the destination allows it, e.g. when converting to WAV files, and are converted to the resolution of the destination otherwise. 16 bit samples are read exactly as before.
Elektron Tonverk Preset
Fixed: Presets with loops could not be loaded by the Tonverk when a loop ended on the last frame of its sample or its cross-fade was longer than the audio in front of the loop start, which the Tonverk cross-fades the end of the loop with. The loops of the factory samples always end at least one frame before the end of their sample, and their cross-fades are always shorter than the audio in front of the loop start, by at least 198 frames, and at most 71.5% of their loop. A sample whose loop ends on its last frame now gets the first frame of the loop appended, the frame which follows the loop end when it plays, so the loop and the sound stay the same. A cross-fade which does not fit moves the loop start forward until 256 frames of the sample are left in front of the cross-fade, so the cross-fade blends the end of the loop into the audio which the loop skips, like a sampler which cross-fades into the start of the loop; a cross-fade which covers most of the loop is limited to half of the remaining loop.
Fixed: Converted presets did not show up in the preset menu of the Tonverk. They carried the category and the keywords of their source as their category and tags, e.g. 'Bell' or 'Chromatic Percussion' and 'analog', while the Tonverk knows only its two main categories DRUMS and KEYS, by which the LOAD PRESET menu filters, and its own set of upper-case tags; a preset without keywords had no 'tags' entry at all, which every preset that the device saves has. A preset now gets the category DRUMS when it is written as a drum kit or its source is drums and KEYS otherwise, the category and the keywords of the source which match a tag of the Tonverk become its tags (e.g. 'Bass' as BASS, 'FX' as SOUND FX, 'dark' as DARK), the others are left out, and an empty 'tags' entry is written when none matches.
Ensoniq EPS/EPS16+/ASR-10
New: Read the LFO (wave, speed, delay, reset) and its modulation of the pitch (vibrato), the filter 1 cutoff and the volume (tremolo) incl. the modulation wheel control of the LFO depth.
Fixed: Velocity modulation of the volume and the filter 1 cutoff was read from the pressure source instead of the velocity sources.
Fixed: The pitch envelope depth was far too large (the amount is in 1/8 semi-tones, max. 15.7).
Ensoniq Mirage
New: Read the pitch LFO (vibrato) incl. the modulation wheel control of the depth.
FL Studio DirectWave
New: The tuning of a zone (the Semi and Fine columns of the zone list) is read and the cents of the tuning are written; before, only the semitones were applied to the root key when writing and the tuning was not read at all.
New: Read and write an LFO which modulates the filter cutoff.
New: Read and write the modulation wheel depth of the pitch, volume and filter cutoff LFOs.
New: Read and write the pitch envelope (Zone Env 1 or 2 routed to Voice Pitch in the modulation matrix). Note that DirectWave raises the pitch by at most one octave.
New: Read and write the key tracking of a zone (the K-Trk column of the zone list).
Logic EXS24
New: Read/write the LFOs (rate incl. tempo synced values, waveform, fade-in) and their routing in the modulation matrix to the pitch, the volume and the filter 1 cutoff (vibrato, tremolo, cutoff LFO) incl. the modulation wheel as Via source.
Improved: The conversion of the pitch modulation amount is more accurate for small values (e.g. 343 is now 50 cents instead of 59).
Fixed: AIFF files with certain number of spaces in the comment could fail to load.
Fixed: Zones without a sample played the sample of another zone - the one whose number matched the index of the zone - or logged 'Sample index out of bounds'. Logic stores a zone without a sample, which plays nothing, with the sample number -1. Zones without a sample are now left out and counted in one message. Only if not a single zone of a file references a sample, the zones still play the samples in their order.
Fixed: Sample names which originally contained double quotes in their filenames but got automatically replaced with dashes on Windows are now found.
NI Kontakt
Kontakt 1
New: Read release trigger groups and round-robin.
Fixed: Envelopes were ignored.
Fixed: Bypassed filters were applied.
Fixed: Writing: Samples with non-ASCII characters in their names could not be found. These characters are now replaced.
Kontakt 1/2
New: Added write support for Kontakt 2 (NKI + NKM).
Fixed: Volume conversion used 20 dB instead of 6 dB per doubling.
Fixed: The exclusive group was assigned to the wrong voice group.
Fixed: Sample files were not found if a zone name contained characters which are not allowed
in file names.
Kontakt 1-4.1
New: Decay-break-decay envelopes are approximated.
New: Read LFOs (volume, pitch, filter cutoff), the velocity, key and modulation wheel modulation of the filter cutoff and the polyphony.
Fixed: The sample start was ignored if no sample end was set and negative sample ends were not supported (e.g. loop slices).
Fixed: Removed "Unknown filter type" errors for effects like equalizers.
Fixed: Round-robin was only detected in the first group start criteria.
Fixed: The filter cutoff frequency was much too low (e.g. 5Hz instead of 143Hz).
Fixed: A sawtooth LFO was read as ascending instead of descending.
Kontakt 2-4.1
New: Read the filter from the group insert effects.
Improved: Groups which can only be triggered by the sequencer are ignored.
Fixed: monolith: Instruments with an impulse response sample (convolution) assigned the wrong sample to all zones.
Fixed: Kontakt 3-4.1: Group and instrument tuning was wrong and the zone tuning was lost.
Kontakt 5+
New: Kontakt 6.8 can now be selected as output format and writes NKI and NKM files.
New: Reads the insert filter effects of groups (Legacy, Pro-53, Ladder, SV, Daft and AR filter types with cutoff and resonance) and more modulators: LFOs (volume, pitch, filter cutoff), DBD and AHD envelopes, voice groups.
Fixed: Instruments saved with Kontakt 8 lost most of their modulation.
Fixed: The attack curve of envelopes had the wrong sign.
Fixed: A '-' in the sound info was read as the author or description.
Fixed: The MIDI channel 'Omni' of instruments in a multi was read as 'off'.
NKM multis
New: (Kontakt 1/2) the volume and panning are read into the instruments of the performance, Omni is detected, Kontakt 2 programs read the MIDI transpose.
New: (Kontakt 4.2+) the master volume and master tune of the multi as well as the MIDI transpose, volume and panning of the instruments are read into the performance.
New: Muted instruments are not converted; if instruments are soloed, only these are converted.
Fixed: The MIDI channel of NKM instruments could be taken from the wrong slot.
Fixed: (Kontakt 1) the volume, panning and tuning of the instruments were applied twice.
NI Maschine
New: Read/write the LFO of Maschine 1 (MSND) and Maschine 2/3 (MXSND) sounds and its routing to Pitch and Filter Cutoff. The 'MW > LFO Depth' setting (Maschine 1.5 and later) is converted as the modulation wheel depth. The rate in Hertz is approximated from the Speed knob, tempo synchronized speeds are read like unsynchronized ones.
New: Written MSND files contain the parameters of Maschine 1.5 and later, which adds the pitch-bend range.
Fixed: Maschine did not find the samples of written MSND files, they had to be located manually.
Fixed: Written MSND files contained a Metaverb insert effect.
Polyend Tracker
New: Read/write the LFOs of the volume, filter cutoff and fine tune automation. The tempo synchronized speed is converted assuming 120 BPM. Since an automation uses either its envelope or its LFO, an LFO is only written if the envelope is not used; the volume LFO only if the amplitude envelope has no attack, release or sustain level below the maximum.
Reason NN-XT
New: Read/write LFO 1 and its modulation of pitch, filter cutoff and level incl. the modulation wheel ('LFO 1 Amt'). LFO 2 is read if it modulates the pitch and LFO 1 does not. The modes 'Group Rate' and 'Tempo Sync' are read like 'Free Run'.
Fixed: The LFO 1 delay and tempo sync rate were written with wrong default values.
Renoise
New: Read and write an LFO which modulates the filter cutoff (incl. modulation wheel depth).*
SFZ
New: Exclusive groups are read and written ('group' and 'off_by').
Roland S-7xx
New: Read the LFO of the partials and its depths for Pitch, TVF (filter cutoff) and TVA (level) incl. the LFO depths of the modulation wheel (Mod section of the patch). The rates and the pitch depth were measured on an XV-5080: rate 100 is about 4.2 Hz, a pitch depth of 63 is ±1.5 octaves. The scaling of the TVF and TVA depths is estimated (63 = full cutoff range resp. down to silence).
Fixed: Samples with undocumented sample frequencies (index >= 6) are now modded with 6 and a warning is logged.
Spectrasonics Omnisphere 3
New: Read/write LFOs which modulate the pitch, the filter cutoff and the amplitude of a layer incl. the modulation wheel controlling the LFO depth. Tempo synchronized rates are read like unsynchronized ones. Written presets use LFO 1 for the pitch, LFO 2 for the filter cutoff and LFO 3 for the amplitude. The amplitude of layer A is written as the default 0.75, since an amplitude modulation is not audible at the maximum.
Fixed: The depth of the pitch envelope was read from and written to the target parameter of the modulation matrix, which is the coarse tuning, instead of the amount. Converted presets were transposed by the envelope depth.
Synthstrom Deluge
New: Read/write the patch cables from the two LFOs to the pitch, the volume and the filter frequency incl. the modulation wheel (source Y) controlling their depth. Rates and depths are converted with the scaling of the firmware, e.g. the vibrato of the factory preset 'Violin' is 4.7 Hz with a depth of 13 cents. Tempo synchronized rates are read at 120 BPM.
New: The option 'Oscillator mode' set to 'Dual (two layers)' (DelugeOscillatorMode=dual on the command line) plays the zones of a synth which sound together on the same note as those of the first oscillator - e.g. a swell on top of a body - with its second oscillator, which was always silent. Only zones which sound together count as layers: key ranges which follow each other, velocity layers, round-robins and the two channels of a stereo recording fit onto one oscillator - hard panned mono pairs are combined into stereo samples - and an overlapping copy of a zone, the same sample with the same root note, tuning and panning, is trimmed instead of being counted. A zone 40 dB or more below the loudest one of the preset, often a layer which the source switched off by its volume, is neither counted nor written; if only one layer is left, only one oscillator is written. The two oscillators together play at the level of one oscillator at full volume and keep the balance of their layers, and an oscillator plays a short silent sample on the keys on which only the other layer sounds, since a Deluge key map has no gaps. A preset on which more than two layers sound together is not written in this mode, and neither is one whose two layers have attack times more than 2 seconds apart, since both oscillators share one amplitude envelope and a layer which fades in over several seconds would start at full level with the note; different decay, sustain or release times and different filters are kept. The log names the preset and the keys, e.g. "Skipped 'Feeling Down': 4 layers sound together on key 38; the Deluge has 2 oscillators."; this is checked before anything is written for the preset.
Fixed: A zone 40 dB or more below the loudest zone of a synth - often a layer which the source switched off by its volume - played at the full level, since the sample ranges of the Deluge have no level of their own, and could even replace the audible zones.
TAL Sampler
New: Read/write the 3 LFOs routed by the modulation matrix to the master tune, the master volume and the filter cutoff incl. the modulation wheel modulating the amount of such an entry. E.g. an LFO with the rate 0.3 runs at 5.4 Hz and the displayed amount 6 semitones of the master tune is a vibrato of +-600 cents.
Fixed: The decay of the envelopes was converted with the time law of the attack and the release, 7.2 x^4 seconds, but TAL-Sampler decays much slower: exponentially towards the sustain level, by 6 dB after 0.35 seconds with the value 0.4 and after 1.35 seconds with 0.6, rendered offline (4.7.2). A decay is now converted to the linear decay which reaches -20 dB at the same moment, 37.1 x^3.75 seconds, and written back with the inverse law.
Fixed: The filter was applied to all layers if the first layer is switched through it and to none otherwise, but TAL-Sampler switches each layer through its filter or not. Only the layers which are switched through the filter get it now, and a written preset switches only the layers through it whose samples have a filter. A destination which knows only one filter for the whole preset writes none when the layers differ, as for any other source whose zones differ.
Teenage Engineering OP-XY
Fixed: The key ranges of presets which were saved on the device were read wrong. The device stores only the highest key of each region - a region starts on the key above the region below it - but every region was read as starting on the lowest key of the keyboard, so that all of them overlapped.
Waldorf Quantum/Iridium
New: A warning is logged when the groups of a preset need more oscillators than the maximum number of layers provides, 3 per layer. The groups which do not fit are merged into the last sample map, where entries which overlap alternate on successive notes instead of sounding together; the warning names the preset, the samples which are merged and the number of layers which keeps them apart.
New: An oscillator whose samples have no filter while others of the layer have one plays around the filter: it is written with the destination 'VCA', which sends it directly to the amplifier, and read without the filter. A filtered and an unfiltered group are not merged into one oscillator anymore.
Fixed: A group whose gain is above 0 dB lost that part of it. The gain of a group is written as the volume of its oscillator, which ends at 0 dB, and it was taken off the entries of the sample map in full. The part above 0 dB now stays in the entries of the sample map. A DecentSampler preset whose two layers are raised by 1.7 and 2.3 dB arrived that much quieter, with the balance between its layers changed.
Yamaha YSFC
New: Read/write the Element LFO of Montage and MODX performances (pitch, amplitude and filter modulation depths, wave form, speed incl. the extended speed). The scaling was measured on a MODX+, e.g. speed 40 is 5.7 Hz and a pitch modulation depth of 127 is +-570 cents.
New: Read/write the Part LFO with the destinations Element Level, Pitch and Filter Cutoff, including control boxes which connect the modulation wheel to 'P.LFO Depth 1-3'. It is written instead of the Element LFO since only its depth can be controlled by the wheel. The scaling was measured on a MODX+: speed 41 is 1.8 Hz and the depth curve is very steep, a pitch depth of 64 is +-113 cents, 127 is +-2400 cents.