Record Stems
Stems records RESERVOIR’s own sound engine to one WAV file per part, plus the drum kit, the full mix and the MIDI that played. The files line up sample for sample, so a DAW session built from them plays back the take as you heard it. Stems capture the audio RESERVOIR renders itself. If you only use RESERVOIR to send MIDI to other instruments, record those instruments in your DAW instead.Making a take
- Click v beside Stems and choose a format (below). The choice is kept for the session.
- Click Stems. Recording starts at once. The strip over the visualizer reads REC STEMS with the elapsed time and a level meter of what is being captured.
- Play.
- Click Stems again to stop. The strip reads STEMS TAIL while reverb and delay ring out: recording continues until the mix falls below -60 dBFS, or 10 seconds at most, so no tail is cut off.
- The strip then reads Saved N stems (C ch) to the take folder, with Reveal and Copy path buttons.
Formats
With a stereo room bound, the room formats give the same files as Per part stereo.
Per part stereo only holds outputs 1 and 2. On a room with more speakers, a note placed on speaker 3 or higher is missing from that file. Choose Room format or Fold room to stereo whenever the room has more than two outputs.
Include
- FX stem (when shared): the shared reverb and delay return. RESERVOIR only writes it when it has to save CPU during a heavy take. From that point the part files are dry, and the reverb and delay are in this file.
- Kit: the drum kit, with its own reverb and delay.
- Mix: the final output.
- Performance.mid: the notes of the take.
Destination
By default each take goes to a new folder in~/Music/RESERVOIR Stems, named with the date, time and preset. Choose folder… sets a different place. Ask every time shows a save dialog when you start.
Keep silent layers is off by default, so parts that played nothing leave no file.
Files in a take
L01 <instrument>.wavtoL64 <instrument>.wav: one file per part that played, numbered by part.Kit.wav,FX.wav,Mix.wavwhen included.Performance.midwhen included.info.txt: preset, room, sample rate, format and the peak level of each file.
Warnings while recording
- Room has 16 outputs, device has 2: you will hear outputs 1-2 only. The bound room uses more outputs than your audio device has. The files still hold every channel of the room, but you only hear part of it.
- Capture is very quiet: Mix peak -45 dBFS after 2 s. Little sound is reaching the stems. Check the levels and that the room you bound is the one you mean to use.
- N host blocks larger than the prepared size not captured. The host sent a larger audio buffer than it announced. Those blocks are missing from the files. Set the host buffer size before recording.