Recording

The Video Recorder captures a synchronised video + audio recording of your show directly inside CuePoints, from a webcam, capture card, or network (NDI®) source. Each recording becomes a new track in your project, with the cues you fire during the show landing on it at exactly the right moments, and is saved as a standard MP4 file you can play anywhere. Audio and video are locked together by the recorder itself, so a recording that starts in sync stays in sync, even across multi-hour shows.

There are two halves to it. The Record button in the main transport starts and stops recording, and is all you need once the recorder is set up. The Video Recorder palette: where you choose your camera, audio inputs and file format, opens from the cog beside that Record button.

You only need the palette when you're changing something. Your last setup is remembered, so on an ordinary show day you can launch CuePoints, press Record, and never open it; the transport button, the recorder's status display on the main window, and remote control over OSC all work with the palette closed. The first press after launching starts the recorder if it isn't already running, and begins recording as soon as it's ready.

Choosing A Video Source

Pick your camera from the Video Device dropdown. Three kinds of source appear in the list:

• Cameras and capture devices connected to your Mac, webcams, Continuity Camera, and capture hardware that shows up as a camera (including Blackmagic UltraStudio devices and virtual cameras such as OBS Virtual Camera).

• NDI® sources on your network, PTZ cameras, NDI encoders, or any app broadcasting NDI. These are discovered automatically and listed by their network name.

• Blackmagic DeckLink inputs (SDI/HDMI) when Desktop Video is installed.

If a device you've just plugged in (or an NDI source you've just powered on) isn't listed, click Refresh Devices.

Once a source is selected, the live preview appears and Detected Rate shows the frame rate the recorder measured from the source itself (e.g. 60 fps). Recordings always use the source's real rate; you don't set it manually. If you press Record before the source has been detected, the recorder shows Waiting for source… and starts automatically the moment the signal is confirmed.

The Exposure slider applies a brightness adjustment (in EV stops) to the incoming picture before it's recorded. Leave it at 0.0 EV unless your camera image is too dark or too bright to read.

Resolution, Format & Quality

• Resolution: the size of the recorded file. If the source is larger or smaller it's scaled to fit. 1920x1080 (1080p) is the right choice for almost everyone.

• Format: MP4 / H.264 is small and plays everywhere; MP4 / H.265 is smaller still on modern machines.

• Quality: the compression level (Low / Medium / High).

The Est. MB/min line updates live so you can check the recording will fit on your drive, as a guide, MP4 / H.264 at Medium 1080p is roughly 45 MB per minute.

Set Save To to the folder where recordings should be written. Files are named with the date and time the recording started, so repeated takes never overwrite each other.

Audio Inputs

The recorder captures two audio channels, Audio Input L and Audio Input R: which become the left and right channels of the recording's stereo track. Each side is chosen independently:

• Pick a device for each side (a microphone, an audio interface, an NDI source's embedded audio, or a Blackmagic input's embedded audio). Both sides can use the same device, the usual setup, or two different ones.

• The second dropdown beside each device selects which channel of that device feeds the side, useful for multi-channel interfaces and embedded SDI audio.

• Set a side to None to record silence on that channel.

Gain L / Gain R adjust each side's level in dB before it's written to the file, and the meters below show the live level of exactly what will be recorded. Aim for meters that move well into the bar on loud moments without sitting at the top.

Two things worth knowing:

• Both sides must run at the same sample rate. If you pick two devices set to different rates (say 44.1 kHz and 48 kHz), CuePoints will refuse the configuration and tell you, set both devices to the same rate in Audio MIDI Setup and try again.

• If a selected audio device stops delivering during a recording (cable pulled, device powered off), the recording keeps going: that channel records silence rather than losing the whole take.

Delays: Getting Perfect Sync

Cameras and network sources deliver their picture later than a microphone delivers its sound; an NDI camera can easily be several hundred milliseconds behind a local mic, because its video is encoded, sent across the network, and decoded before it reaches CuePoints. Left uncorrected, you'd hear things slightly before you see them.

Click Show Delays to expand the palette and reveal three compensation fields. Values are in milliseconds (0–2000), and changes apply from the next recording you start:

• Audio L delay (ms) / Audio R delay (ms): shifts that audio channel later, so early-arriving sound lines up with the delayed picture. The delay is written into the recorded file as real audio timing, so the correction holds in every player and editor.

• OSC marker delay (ms): shifts markers fired into CuePoints over OSC (for example from a lighting console) later by the same logic, so cue markers land on the recording at the moment the picture shows the cue firing, not the moment the console sent the message.

Calibrating with a clap test:

1. Set the audio delays to 0 and make a short recording of a single sharp clap, filmed so the moment of contact is clearly visible.

2. Open the recording in your editor and measure the gap between the frame where the hands meet and the spike in the audio waveform.

3. Enter that measurement (in milliseconds) as the audio delay for both channels.

4. Record a second clap to confirm the waveform now lands on the contact frame.

The right value belongs to the source and network, not the show; the same NDI camera on the same network needs the same correction every time. CuePoints remembers your delay values between sessions; re-check them when you change camera, network, or camera settings (encoder bitrate changes can shift an NDI camera's latency).

A note on how sync feels: a clap aligned perfectly to the frame can still read as very slightly early on playback. Human perception of sync is asymmetric, because light naturally arrives before sound, viewers notice audio that's early about three times more readily than audio that's late: formal testing puts the detection threshold at roughly 45 ms for early audio versus 125 ms for late (ITU-R BT.1359-1), and broadcast practice allows up to 40 ms early but 60 ms late (EBU R37). If a frame-accurate alignment still feels early to you, adding one extra frame of audio delay (17–40 ms depending on frame rate) is comfortably within those tolerances and usually reads as perfect.

Presets

The Preset dropdown at the top of the palette stores a complete recorder setup: video source, audio devices and channels, resolution, format, quality, record mode, and track-naming settings. Switching between venues or rigs is one click instead of ten.

• Configure the palette, then choose Save Preset… from the dropdown and give it a name.

• Select a preset to recall it; CuePoints re-opens with your last-used preset automatically.

• With a preset selected, Delete Current Preset… removes it (with a confirmation).

Your settings are also remembered when no preset is selected, (No Preset) simply means "whatever I last set up."

Record Mode

• Manual Record: the transport buttons at the bottom of the palette are in charge: you press Record and Stop yourself.

• Timecode Triggers Record: pressing Recordarms the recorder (Waiting for Timecode). Recording starts automatically the moment incoming timecode starts running, stops when timecode stops, and re-arms itself for the next run, ideal for unattended capture of every timecoded run of the show. While armed or recording, the armed/recording status is also broadcast over OSC.

Mode changes take effect the next time you press Record.

Recording

• Record starts the recording (or arms it, in timecode mode). The timecode counter starts when the first frame is actually written and the status line changes from Previewing to Recording.

• Stop ends the recording and finalises the file, wait for the confirmation before pulling drives or closing the lid.

• Stop and Delete ends the recording and discards the file, for abandoned takes, so your recordings folder only contains keepers.

The Record button in the main transport does the same job without the palette: press it to start, press it again to stop, and it shows the recording state and elapsed time on the main window as you go. The bin icon beside it is Stop and Delete, and appears only while a recording is running. The same commands are available over OSC, so a console or tablet can drive the recorder for you.

Every recording creates a new track in your project containing the recorded video, and any cues you fire while recording are placed on that track at the exact moment they happened. Track Name controls what the new track is called:

• Custom: you type the name yourself (falls back to the date/time if left empty).

• Date/Time: named with the recording's start time.

• Cue Range: named automatically from the cues fired during the recording, e.g. Q12-Q45 (or Q12 if only one cue fired). The CuePoint Type dropdown selects which marker type counts for this, pick your primary cue type (e.g. Lighting) so standby notes or other marker types don't affect the name. If no matching cues fire, the track falls back to the date/time name.

A few practical notes for long recordings:

• Check Est. MB/min against your free disk space before a long show; a 3-hour show at 45 MB/min needs about 8 GB.

• Set your Mac not to sleep for the duration (System Settings > Displays > Advanced… > Prevent automatic sleeping).

Recording A Copy With The Overlay Baked In

The Overlay dropdown decides whether a second, overlay-branded copy is recorded alongside the clean file:

• None: one clean recording, no overlay copy.

• Current: bakes whatever the main overlay is currently showing, and follows it: if you hide the overlay during the show, it disappears from the recording too.

• A saved overlay set: bakes that set regardless of what the main window is displaying, so you can record an overlay the audience never sees.

Anything other than None records two files simultaneously: the clean recording, and a second copy (named …_overlay) with the CuePoints overlay, cue names, timers, and other on-screen data, rendered permanently into the picture. The second Save To field chooses where that copy goes; it's enabled as soon as you pick an overlay set, and defaults to the same folder as the clean recording.

The clean file is always the primary, if the overlay copy ever hits a problem, the clean recording is protected first.

Troubleshooting

• My device isn't in the list: click Refresh Devices. For NDI sources, check the sender is on the same network and not blocked by a firewall; for DeckLink inputs, check Blackmagic Desktop Video is installed.

• The preview is black: the source is connected but not sending a picture. Check the camera's output, the cable, and (for capture cards) that the input connector (SDI vs HDMI) matches where the cable is plugged in.

• "Starting the recorder…" after pressing Record: you pressed Record before the recorder had finished starting up and finding your devices, which is normal on the first press after launching. It begins recording by itself the moment it's ready; you don't need to press again.

• "Waiting for source…" after pressing Record: the recorder hasn't yet confirmed the source's frame rate. It starts by itself as soon as the signal is detected; if it never does, the source isn't delivering video.

• No audio in the meters: check the device and source-channel dropdowns for that side, and that the side isn't set to None. If the device is listed but silent, another app may have exclusive control of it.

• Recording refuses to start with a sample-rate error: your L and R devices are set to different rates. Open Audio MIDI Setup, set both to the same rate (48 kHz is ideal), and reconfigure.

• Audio and video are out by a constant amount: that's the latency of your source. Calibrate the audio delay with a clap test (above). If the offset changes over the length of the recording rather than staying constant, send us the recording and your debug logs.

• The sound feels early even though the waveform lines up: see the perception note in the delay section; add up to one frame of extra delay.

• Markers from my console land slightly early on the recording: that's the same latency story as audio: the console's OSC message arrives instantly while the picture is delayed. Set OSC marker delay to match your measured video latency.

Technical Notes

For those who care about what's under the hood:

• Audio is recorded as 48 kHz stereo AAC.

• Audio and video are timestamped against a single shared clock, whatever their source. Dropped frames or brief audio interruptions are preserved as gaps/silence rather than shifting everything after them, which is why sync cannot drift over time.

• The audio track begins at exactly the same instant as the video track in every file, no hidden container offsets, so waveform-based tools, editors, and players all agree about where sync sits.

• Recordings are finalised on Stop; the file is not playable until finalisation completes.