TG Studios

TG Drop Deck, the manual

Everything the app does, in order, with every key and every setting.

This is the reference. If you want the short version, press 1 in the app and you will hear applause; that is the whole learning curve, and the rest of this is for when you want more.

Nothing here needs a mouse. Every feature is reachable from the keyboard and announced out loud, and the app was built and tested with NVDA running.

Conventions. Keys are written like Ctrl+B. Menu paths are written like On air, Go live. Where a thing has both a key and a menu item, both are given, because a key you have not learned yet is no use to you today.

Two ways to read this. Chapters 1 to 32 are the reference: every feature in the order you meet it. Chapters 33 to 37 are the other way round and start from what you are trying to do, whether that is DJing a set, recording a podcast, running a radio station, going live on YouTube, or reaching more than one place at once. If you know what you want to make, start there and follow the links back.

On a Mac? The Mac copy has a manual of its own, with the Mac keys, at /drop-deck-guide-mac/.

Back to TG Drop Deck for the download.

Contents

Getting started

The soundboard

Running a show

Your voice, and other audio

The picture

Getting it out

Putting it to work

Reference

1. Installing and first run

Drop Deck is free, needs no sign-up and no licence key, and comes two ways.

The installer puts it in your own user folder, so it never asks for an administrator password. The zip installs nothing: unzip it wherever you like, keep the demo folder beside TG Drop Deck.exe, and run it from there. A portable copy can live on a memory stick and updates itself in place.

Windows may warn the first time, because the app is not code signed yet. Choose More info, then Run anyway.

Requirements. Windows 10 or later. Nothing else: no Python, no runtime, no sound card beyond the one you already have. Capturing one program's audio needs Windows 10 build 20348 or later; everything else works on any Windows 10.

The first time you open it you already have forty sounds. Press 1 and you hear applause. Your settings and your board live in AppData\Roaming\TG Studios\TG Drop Deck, not in the program folder, so they are shared between an installed copy and a portable one and survive an update.

2. How the app is laid out

There are two views and one status bar.

The soundboard is eighty buttons in four banks of twenty. Ctrl+Tab moves between banks; the tab you are on says which bank it is.

The running order is a list of songs and drops with columns for title, artist, whether it is a song or a drop, length, when it starts, and its own crossfade if you have given it one.

Ctrl+Shift+P goes to the running order, Ctrl+Shift+S back to the soundboard, and Ctrl+Alt+Tab swaps between the two.

The status bar carries the three volumes, whether ducking is on, whether the microphone is open, whether you are on air, and whether you are recording. It is one line and it is always true; nothing there needs a key press to refresh.

3. How the app talks to you

Your screen reader reads the controls. The app adds speech of its own on top, and Preferences, Speech has three settings for how much of it you get:

There is a separate switch for announcing what is playing as each sound or track starts.

Turning the app quieter never silences the essential channel. Being told whether you are live is not a pleasantry.

4. The four banks

The keyboard map came from the soundboard this app replaces and has not changed. It is muscle memory, and it is not up for redesign.

Bank What it is for Keys
1 Sound effects 1 to 0, and Shift+ those for 11 to 20
2 Dialog drops Ctrl+digit, and Ctrl+Shift+digit
3 Music beds Alt+Ctrl+digit, and Alt+Ctrl+Shift+digit
4 Miscellaneous No fixed keys: give each slot its own

Twenty slots a bank, eighty in all. Ctrl+Tab moves between banks. Ctrl+F2 renames the bank you are on, and Banks, Reset this bank's name puts the original back.

Bank four is the one you design yourself. Its slots have no keys until you give them one, through Sounds, Assign a hotkey or Alt+Enter.

5. Playing and stopping sounds

Press the key and the sound plays. That is all there is to the good case.

Sound effects and drops overlap. Press the same key twice and you hear it twice, and a laugh can land on top of a sting: banks 1, 2 and 4 overlap and never cut each other off. That is what a soundboard is for, and it is the default.

Beds toggle. Bank three is still the looping bank: the same key starts a bed and stops it, and only one bed plays at a time: starting another takes the first down with its own fade, because two pieces of music at once is a mistake rather than a texture.

Three ways to stop.

Stopping can cut instead of fading. A quarter second fade is the default so a stop does not click. Turn it off in the same place if you are riding a mixer or a fader in a DAW and want the app out of the way instantly.

Ctrl+L says what is playing right now.

6. Putting your own sounds in

Land on any button and press Space, or use Sounds, Assign a sound file.

It takes WAV, MP3, M4A, FLAC, OGG, WMA, AAC, Opus and more, twenty formats in all, so an iTunes or Apple Music library goes straight in.

A folder works too. Sounds, Assign a folder puts a whole folder on one key, and each press plays a different one at random without repeating itself until it has been through them. Drop another jingle into that folder later and the app picks it up; nothing needs re-scanning by hand.

Previewing while you choose. In the file browser, turn on Play each sound as I reach it, or press Alt+P, and each file plays as you arrow onto it. You pick by ear instead of by file name. A file that will not play says so rather than going quiet.

Sounds are never copied into the program. The board remembers where they live. That keeps boards small and means one file can be on several boards.

7. Changing a sound

Alt+Enter on a slot opens its properties, which holds everything on one window. The same things are on the Sounds menu one at a time.

8. Removing and restoring slots

Delete empties a slot but keeps its keys and its name.

Shift+Delete takes the slot off the board entirely. Its key then does nothing, which is the point: an empty pad you keep hitting by accident is worse than no pad. Nothing is renumbered, ever, so taking slot 5 away leaves 6 on the 6 key.

Sounds, Put a removed slot back lists what you have taken off and puts one back where it was, with everything it had. Put this bank's slots back restores the lot.

9. Finding a sound

Ctrl+F searches every slot on the board by name, across all four banks. Ctrl+E does the same, because it was the key for two releases and taking a key back off somebody who learned it is not a fix.

Type part of a name, arrow through the matches, and Alt+P plays one without closing the window so you can check you have the right one. Enter jumps to it on the board.

10. Firing sounds from another program

This is the point of a soundboard on a live show: you are in your DAW, your browser or on a call, and you still need the sting.

Right-click a slot, or press Alt+Enter, and give it a global hotkey. Any slot in any bank can have one, and it fires while another program has focus.

It needs at least one modifier, such as Ctrl or Alt. Alt on its own counts, so Alt+A is fine. A bare key would be taken away from every other program on the machine, so the app refuses it and says why.

Ctrl+G arms and disarms the whole set at once, and disarming hands the keys straight back to Windows. The Sounds menu shows whether they are on, because "does this key belong to me right now" has to be answerable without pressing it.

11. The running order

The cue sheet, what is coming up

Ctrl+Shift+C opens a separate window listing everything ticked in the running order, top to bottom, with the one on air at the top. A track leaves the list ten seconds after it starts, so the list is always what is still to come.

Two things it shows that the running order does not. A ticked track whose file has gone is listed and marked File missing, rather than silently left out. And the last row always says the running order ends, so you know it is coming before it happens.

The list drains itself while you are reading it, and the row you are standing on is the one thing that never moves. That is deliberate: removing the row under the cursor makes a screen reader stop mid sentence and read out a track you never chose. It is held until you arrow off it.

Ctrl+Shift+P takes you to it.

Getting songs in. Copy them in File Explorer and press Ctrl+V, as many at once as you like, and a whole folder works. Add songs to the end does the same through a dialog, and you can drag files onto the list.

What it shows. Title, artist, whether it is a song or a drop, length, the time it starts, and its own crossfade if it has one. Title and artist come out of the file's tags and fall back to the file name.

Playing. Enter plays from the row you are on. Ctrl+Shift+Enter does the same from anywhere in the app. Once it is going it works down the list on its own.

Ticking. Space ticks and unticks. An unticked track stays in the list and keeps its place, and is skipped. Shift+A ticks every track and Shift+U unticks them all. Typing a letter still jumps to a track by name.

Moving. Alt+Up and Alt+Down move the track you are on by one place. Alt+Home sends it to the top and Alt+End to the bottom, which is what you want on a long show. Delete takes it out.

Getting about. Ctrl+Shift+L puts the cursor on whatever is on air, which matters once the show has run past the row you were looking at. The Playlist menu also has Next track, Previous track and Stop the playlist.

The Applications key opens a menu with everything you can do to the row you are on, which is how you find these without learning them first.

12. Crossfades and segues

Each song hands over to the next before it ends. How long they overlap is the crossfade. The box for it sits under the running order, and the same number is in Preferences, Playlist: two views of one setting.

Three seconds to start with. Set it to zero and each song plays right out before the next begins. The overlap is the outgoing song riding down under the new one as it comes in, so every start time in the list moves when you change it, and the list shows the new times.

A single track can have its own, from its right-click menu. Useful for an ident you want played dry in the middle of a music show.

Shift+Enter crosses into the track you are on, fading out whatever is on air at the crossfade length. That is how you get out of a song early, and it is the same thing the right-click menu calls Segue.

13. Drops and your drops library

A drop is an ident, a stinger or a bumper: something short that goes between songs rather than being one.

The drops library is Playlist, Drops library. Put your regular idents in it once and they are one key away for ever.

A drop in the running order is marked as a drop, so you can see the shape of the show at a glance.

14. The warning before a track ends

Preferences, Playlist. Turn it on and set how many seconds, ten by default. It is the countdown clock a sighted presenter watches.

You hear it wherever you hear yourself, which is the monitor output set on the Microphone tab, so with headphones set up there it never reaches the show or the recording.

Six sounds to choose from: a pip, a double pip, a chime, a bell, three ticks and a rising sweep, with a volume of its own. Each plays as you arrow onto it so you can find one you hear over your own music. They are different shapes rather than different notes, which is what lets you tell them apart over a song, and all six sit at the same loudness, so changing your mind never changes how loud your warning is.

A track shorter than the warning does not get one, so a nine second ident with a ten second warning does not beep the moment it starts.

15. Saving a running order

Playlist, Save the running order writes an M3U file. Open a running order loads one back in place of what is there.

M3U because every player opens one, so a saved show can be checked in VLC or handed to somebody else. Drops, ticks and per-track crossfades are kept in comments this app reads back and other players ignore.

Dragging an M3U onto the running order adds it instead of replacing, so you can build a show out of several saved parts.

16. Volume, and getting out of the way

Three faders, because a bed and a sound effect should never fight over one.

Keys What they move
F3 and F4 Sound effects and drops
F5 and F6 Music beds
F7 and F8 The running order

Each says its new level as you move it.

Ducking. Fire an effect or a drop and the music underneath drops about nine decibels, then slides back when it finishes. Opening the microphone does the same. It is the thing radio does that makes a show sound produced rather than assembled, and here it happens without you touching anything.

Ctrl+D turns it off and on. Preferences, Sounds and beds sets how far it ducks and how long beds take to fade in and out.

The duck follows the microphone being OPEN, not you talking. A gate that opens on your voice clips the first word of every sentence, so there is not one.

While you are on air, F7 and F8 become a monitor fader. Pull the music right down to hear your screen reader and get about the list, and your listeners still get it at full level. F3 to F6 move both what you hear and what goes out. This is the single most useful thing in the app for working a show by ear, and it is why the playlist has its own fader at all.

17. The microphone

Ctrl+M opens and closes it. Alt+Shift+M does the same, for when another program has taken Ctrl+M. Ctrl+Shift+M is where you set it up. Microphone on and Microphone settings are on the On air menu, under Microphone, which is where the Mac has them too. Before 3.9.0 they were on the Playlist menu; the keys never moved.

Hearing yourself is off until you turn it on. On headphones it is how you know you are live. On speakers it is a feedback loop.

The monitor can have an output of its own, so what you hear sits in your headphones and the show goes somewhere else entirely.

Nothing opens your microphone except you pressing Ctrl+M. Not startup, not opening a board. Whether it was open is never saved.

If it will not open, you are told why in plain words: Windows privacy settings, another program holding it for itself, unplugged, or no driver, and where to go to fix it. Before giving up, Drop Deck tries the same microphone through the other ways Windows offers it, and says so if that is how it opened. If the microphone you chose is not plugged in, the system default opens instead and you are told which.

If it opens and nothing arrives, not even background noise, Drop Deck says so a couple of seconds later. That is what a Windows privacy block or a muted input looks like from inside a program.

If Ctrl+M does nothing at all

Another program has almost certainly taken the key for the whole computer. Windows then hands every press of it to that program and never tells Drop Deck. Drop Deck checks for this when it starts and names any of its keys that are taken. Call and meeting programs, microphone mute utilities and AutoHotkey scripts are the usual ones.

Going out is a different question from hearing yourself. On the Streaming tab, Put the microphone on air decides whether listeners get you, and it works whether or not you are monitoring. On speakers you monitor nothing and are still live.

18. Voice processing

The Voice tab in Ctrl+Shift+M is a gate, a high pass filter, a three band equaliser, a compressor and a limiter, on your microphone, in that order. It is what makes a voice sit in a mix rather than sounding like a laptop.

It is one list. Up and down choose a setting, left and right change it, and it says the new value as you go. Hold Shift for a bigger step. Page up, Page down, Home and End move around the list as they do in any list.

Process the microphone at the top switches the whole chain off and on while you talk, which is the quickest way to hear what it is doing.

Stage What it is for
Gate Silence between sentences. Below the threshold it turns you down rather than off, so a breath does not slam shut.
High pass Everything below 80 Hz is desk rumble, a fan, or a lorry going past. None of it is your voice.
Equaliser Low, middle and high, plus or minus twelve decibels. Small moves.
Compressor Evens out loud and quiet, so leaning back from the microphone does not disappear. Threshold, ratio, attack, release, make up.
Limiter The ceiling, and it means it.

The limiter is a real ceiling, not a loudness maximiser. Set it to minus one and nothing leaves louder than minus one, ever. It adds no delay at all, because a few milliseconds between your mouth and your headphones sounds like a barrel.

Everything here is live. The microphone you are listening to changes as you arrow about, which is the only way to set a compressor. Cancel puts all of it back; OK keeps it, and it is saved with your board.

The defaults are set up for a voice already, so a good first move is to turn it on and leave it.

19. VST3 effects

Under the list, Plugin lists the VST3 effects installed on this machine. Choose one and it goes into the chain after the compressor and before the limiter, and its own knobs appear in the same list, in words, with real units, read out as you arrow through them.

The plugin's own window is never opened. Most plugin windows are a picture of a rack, and a picture of a rack is nothing at all to a screen reader. As a list it is the same effect with the same settings, and you can actually use it.

Save preset and Open preset keep a plugin's settings in a file of your own, so a chain you like is one file rather than an afternoon.

Instruments are refused. An instrument makes sound rather than changing it, so one in a voice chain would replace your voice with a synthesiser.

Which plugin you chose, and how it was set, is saved with the board. If the plugin is not on the machine next time, the app says so once and keeps the setting, so it comes back when the plugin does.

20. Audio sources

Holding a source back

A capture card, a console or a games call arrives whenever its own hardware gets round to it, and nothing lines them up for you. Hold this source back, in milliseconds, does what OBS calls a sync offset.

It can only ever make a source later. Nothing can make live audio arrive sooner than it does, so if a card is running behind everything else, hold the other sources back instead. This surprises everybody once.

Alt+Shift+S, or On air, Audio sources. Other things to put on the air besides your own voice: a co-host's microphone, a hardware mixer, or one program's audio.

A sound card or a cable

Anything Windows offers as an input. Pick the device, set a gain, choose which channel to take, and say whether it goes on the air and whether you hear it yourself.

One program

Take audio from, One program, then pick it from the list. Windows hands over exactly what that program is playing and nothing else. There is nothing to set up in the program itself, no driver to install and no cable in the middle. It is what OBS calls Application Audio Capture.

The list is built from three directions, and rebuilt every time you open the picker:

Drop Deck remembers a program by name, so it finds it again next week even though its process number will be different.

Close the program and open it again, and its source comes back by itself, found by the program's name rather than by the window you first chose it from. Drop Deck says when a program source goes, and again when it is back.

Your screen reader

NVDA, JAWS, Narrator, ZoomText, Fusion, MAGic, SuperNova and System Access are all in the list, named as screen readers, and capturing one puts its speech on the air. That is how you demonstrate something, or teach with it, without holding a microphone up to your headphones.

They are listed even when they are silent. A screen reader that is not speaking has no window and no audio open, so it would otherwise vanish from the list at exactly the moment you went looking for it.

Two things worth knowing before you put one on air. Everything it says goes out, including it reading Drop Deck's own buttons while you work, so Alt+Ctrl+Shift+S and mute it between the bits you want. And turn "Hear it yourself" off, because you are already hearing your screen reader through your own speakers.

If the program is not running when Drop Deck opens, the source says so. If you close it mid show, the capture notices within a couple of seconds and tells you, instead of going quiet and looking fine.

This needs Windows 10 build 20348 or later. On anything older the option is still there and says it cannot be done, and a cable still works.

A virtual audio cable

The older way, and it works on any Windows. A cable is a free driver that looks like a speaker to one program and a microphone to another; VB-Cable is the usual one. Point a program at the cable in its own settings, then choose the cable as the device here.

Hearing it yourself

Leave it off when the sound already comes out of your speakers from the program itself, or you will hear it twice, slightly apart. Turn it on when the program is sending its audio only to a cable.

Sources are never ducked and never go through the voice processing. Both belong to your microphone: program audio is not somebody talking, and a compressor set up for a voice has no business on a games call.

21. Source control

Alt+Ctrl+Shift+S. This is the one for during a show, and it is worked entirely from the keyboard.

Your microphone is in the list, because soloing a games call has to take your voice down with everything else or it is not a solo. It can be muted and soloed from there. Rename and Remove are unavailable rather than missing while it is the one you are on, and the line under the list says why: Ctrl+M is its switch.

Every source keeps a number, and the number is its position rather than anything to do with its name, so renaming one does not renumber it. That is what makes it worth saying out loud when you are telling somebody which one you mean.

Mute wins over solo. A source that is both stays silent.

Neither is saved with the board. A mute is something you do during a show, and coming back tomorrow to a source that is quiet for reasons you cannot remember is worse than having to press it again.

Turning your microphone off with Ctrl+M does not take your sources off air. They are separate on purpose: somebody putting a call out has not necessarily got their microphone open.

22. Choosing the picture

Alt+Shift+V, or On air, Video source, changes what the stream is showing, and it works while you are on air. Up and down read the choices, Enter puts one out. The connection is not dropped and the sound does not break. The choices are a card with your station name on it, your own artwork, a camera, your screen, or your screen with the camera small in the corner.

Your screen with the camera in the corner puts the screen across the whole picture and the camera small in one corner, rather than the two side by side. Side by side would leave your screen half the width, and at that size ordinary text is unreadable. This way the screen stays legible and you are still on it.

Which corner is up to you. When that source is chosen, a Camera corner box appears in the same window: bottom right, bottom left, top right or top left. It takes effect at once, including while you are live, so you can try each one. Bottom right is the default. Move it if your screen has something important in that corner, or if the platform you are streaming to puts a chat panel over it.

23. Things on top of the picture

Alt+Shift+T, or On air, Screen text, puts your station name, what is playing, a clock or your own words on top of whatever the picture is.

There are four named places and they cannot overlap:

Place Where it is
Top strip across the top
Corner top right
Lower third bottom left
Clock bottom right

Up and down read the places, Enter chooses what one shows, and Delete empties it. Each place can show nothing, your station name, what is playing, the time, your own words, or a text file, which is re-read a second after it changes, so any other program that writes a text file can drive it.

There is no canvas and there are no coordinates, and that is deliberate. Putting something at an exact position is easy; knowing whether it looks right is what needs eyes. The four places are already sized to be readable and cannot land on top of each other, so the answer to "what is on screen" is always short enough to be read out.

Ctrl+Shift+V says what is on screen, on air or off: the picture, and everything on top of it, in words. Nothing else in broadcasting does this.

It tells you when the picture dies. If the camera unplugs, the screen goes black or the capture freezes on one frame, it says so out loud, once, after a few seconds so a blink does not set it off. It says when the picture comes back too.

Going live checks the text fits. A title too long for the lower third is said before you connect, not discovered by somebody watching.

24. Your colours

Alt+Shift+C, or On air, Colours. Three things make up your brand:

Ten ready-made looks are offered first, and every one has been checked, so none of them can be hard to read.

Every colour says how it will read. Arrow down the list and each one tells you: "gold, easy to read, 8.6 to 1", or "light green, too close together to read, 2.2 to 1". You never have to picture it.

That number matters more on video than anywhere. Video stores colour at half the detail of brightness, so two colours that differ only in hue turn to mush when they are encoded, however different they look side by side. Red on blue is the classic one. The number catches that as well, because it measures brightness rather than hue.

This look

Under the list there is a box called This look that you can tab to and read a line at a time. It says the whole brand at once: which ready-made look you are on, all three colours, and how the words and the accent each read against the background. The line above it changes as you arrow through the rows; this box always says the lot.

What it looks like to somebody who can see it

What does this look like to a sighted viewer? draws a sample of your brand and asks the AI what it makes of it. This is the one question the numbers cannot answer. Contrast tells you a pair can be read. It cannot tell you the look is dated, or that an accent colour feels like hazard tape, and those are real things a viewer notices.

You get an impression in one line, then what works and what does not, what kind of station it would suit and what it would suit badly, and one suggested change. It is told to be honest rather than encouraging, because a compliment you cannot check is worth nothing.

There is a question box here too, so you can ask "would this look dated on a music station" or "is the accent too strong" and keep going.

This needs a key, set up on the AI Provider page. See chapter 25.

25. Check my shot

Alt+Shift+D, or On air, Check my shot. This asks a model that can see to look at the picture going out and tell you what is wrong with it.

It is the one part of the app that asks rather than measures. Everything else here works out an answer for itself: how loud a track is, whether two colours can be read together, whether the camera has frozen. Those are good at anything that can be reduced to arithmetic. None of them can tell you that your shirt is the same colour as the wall, that the lamp behind you is blowing out your face, or that your inbox is on the screen you are sharing.

What you get back. A verdict in one line, then the details worst first, then a line beginning "Try:" with the single most useful thing to change. For example:

Your face is under-lit and the window behind you is the brightest thing in frame. You are sitting slightly right of centre, with a lot of empty space above your head. The lower third text is readable and does not cover your face. Try: move the lamp to your left and lower the camera a little.

On a screen it looks for anything private first. Email, chat, a password manager, a notification, a name or an address. It says where the thing is so you can close it.

Setting it up

You use your own account with one of three services, so you pay them directly and nothing goes through TG Studios.

  1. Get a key from Anthropic, OpenAI or Google.
  2. File, Preferences, or Ctrl+P, then the AI Provider page.
  3. Choose who to ask, paste the key, and press OK.

The key is kept in Windows Credential Manager, not in your board file, so a board you send to somebody else does not carry it. You can see it and remove it yourself in Credential Manager under "TG Drop Deck vision key". There is also a Remove this key button on the same page.

The Model box can stay empty, and the default is the quick one. It is a list you can arrow through, and Get the list asks your own account what it can really see, which is the only list that cannot go out of date. You can also type a name in. If you ever hear "does not know that model name", press Get the list and pick from what comes back.

Asking it questions

Under the answer there is a question box. Type a question, press Enter, and it answers about the same picture it just described: "is the plant behind me distracting", "can you read the lower third", "does my shirt blend into the wall". It remembers the conversation, so a follow-up like "what about the other side" means something.

Three things it will not do

It is never needed to go live, and going live never waits for it. It has its own key and does its work on its own thread. If it is slow, or the service is down, Ctrl+B is unaffected.

You do not have to be on air to use it. That is the point of it: it builds the picture, looks at it, and puts it away again, so you can fix your lighting before anybody is watching rather than after. If you are already live, it looks at exactly what is going out and says so.

It never sends a picture of your screen without asking, and it asks every single time. Not once, not remembered. A yes about one screen is not a yes about the next one, and you cannot look at the frame to check what is in it before it goes. The question tells you exactly what is being sent and to whom, and No sends nothing.

It never fails loudly. A wrong key, no internet, or a service having a bad day all come back as a sentence telling you what to change, and the show carries on regardless.

What it costs

One picture, a little under 100 kB, per check. On the default model that is a fraction of a penny and takes about a second. If you put a larger model in the Model box it will be slower and cost more, and it may be worth it before a big show.

26. Streaming audio to your station

Alt+Shift+B goes live on your radio station and comes off air again. Ctrl+B does the same for your video platform. One key each, so the summary you are shown is always for the place you meant.

Ctrl+Shift+B says what the stream is doing: on air or not, for how long, at what bitrate, to where, and whether anything has been lost. Off air it says where the show would go.

Ctrl+B says what it is about to send, and waits for you. Where the show is going, what it is sending, what will be on the screen, and whether your microphone is on the air. Press Enter to go live. It also tells you about the things that would spoil a broadcast without stopping it: a microphone that is not on the air, a picture file you have moved, no camera chosen, or a bitrate the platform will not accept.

If you would rather it did not ask, tick Do not ask again, just go live in that window. Ctrl+Shift+B still answers whenever you want it. To get the question back, tick Say what is going out before Ctrl+B goes live on the Audio streaming page.

Set it up first in On air, Set up streaming, or Ctrl+Shift+P and the Audio streaming page. Everything on it is something the person who runs your server tells you.

Setting What it means
Server Icecast, which also covers Liquidsoap, or SHOUTcast. Video platforms are on their own page, chapter 27
Address The name of the server, with no http in front
Port What listeners use. For SHOUTcast this is still the listening port; the app adds the one a source needs
Mount point The part after the address, such as /live. SHOUTcast does not use one
User name Almost always source. SHOUTcast ignores it
Password The source password, not your listener password
Format MP3, AAC or Ogg Opus
Bitrate 128 is plenty for speech and music together
Station name What listeners see

Test the connection proves it works before the show rather than during it. It connects, says what happened in a box you can read back at your own pace, and disconnects. It never broadcasts anything. The video page has its own button, Check my stream key, which asks YouTube or Facebook whether your key is right.

Which format. MP3 plays everywhere. AAC sounds better for the same bandwidth and is what a lot of stations use. Ogg Opus is better again if your server and your listeners take it.

What goes out. Sounds, beds, the running order, your microphone if it is set to, and any audio source that is on air. Not a preview, and not the warning before a track ends, because those are yours.

Track titles go to the server as they change, so listeners see what is playing. Turn it off on the same tab if your server does its own.

If the connection starts falling behind it tells you, once, rather than quietly losing audio at the far end. If listeners have actually missed something it says that separately, and Ctrl+Shift+B reports both.

27. Streaming video to YouTube, Facebook and Restream

Drop Deck can send your show to YouTube, Facebook, Restream or any other server that speaks RTMP, with a picture, and without a second program in the chain.

All of this is also in the app, on Help, Setting up streaming, and on the How do I set this up? button on the Video streaming page. It is the same text, so it cannot go out of date against what the app does.

Video streaming is a page of its own, separate from the audio streaming page that feeds a radio station. They are separate jobs with none of the same settings: no mount point, no port, no format, a stream key instead of a password, and a picture. Ctrl+B goes to the video platform and Alt+Shift+B goes to the radio station.

Channel name on that page is optional and is not your station's name. Your radio station's name belongs to your radio station; fill this in only if you want your channel called something on the way to air. Leave it empty and Drop Deck just says the platform.

Something has to be on the screen. YouTube will not take sound on its own, so a radio show still sends a picture. A card with your station name and whatever is playing is the default and costs almost nothing.

YouTube Live

Before you start

Setting it up

  1. Open Preferences with Control Shift P and go to the Video streaming page.
  2. Choose your platform in the Platform box. YouTube and Facebook fill in their own address and lock it, because there is only one each.
  3. Press Get my stream key. It opens your YouTube live dashboard.
  4. Copy the stream key from the Stream tab there. It looks like four letter groups separated by hyphens and it does not change, so this is a one time job.
  5. Paste it into Stream key in Drop Deck.
  6. Choose what to show in the Show box. A card with your station name on it is the default and is what most radio shows want. YouTube will not take sound on its own, so something has to be on the screen.
  7. Tick Go live here when I press Control B, so Control B sends the show to this platform rather than to your radio station.
  8. Press Check my stream key. It asks the platform whether your key is right and tells you what the platform said. It sends no sound and no picture, so nothing is broadcast and nothing is posted.
  9. Press OK, then Control B to go live. Control Shift B tells you what the stream is doing at any time.

What happens when you go live

Trying it safely

Picture quality

Facebook Live

Before you start

Setting it up

  1. Open Preferences with Control Shift P and go to the Video streaming page.
  2. Choose your platform in the Platform box. YouTube and Facebook fill in their own address and lock it, because there is only one each.
  3. Press Get my stream key. It opens Facebook's live producer.
  4. Choose where the video is going, then Go live, then Streaming software.
  5. Turn on the persistent stream key in Advanced Settings if you want to reuse it. Without that, Facebook gives you a new key every time and you have to paste a new one before every show.
  6. Copy the stream key and paste it into Stream key in Drop Deck.
  7. Choose what to show in the Show box. A card with your station name on it is the default and is what most radio shows want. YouTube will not take sound on its own, so something has to be on the screen.
  8. Tick Go live here when I press Control B, so Control B sends the show to this platform rather than to your radio station.
  9. Press Check my stream key. It asks the platform whether your key is right and tells you what the platform said. It sends no sound and no picture, so nothing is broadcast and nothing is posted.
  10. Press OK, then Control B to go live. Control Shift B tells you what the stream is doing at any time.

What happens when you go live

Trying it safely

Picture quality

Restream

Before you start

Setting it up

  1. Open Preferences with Control Shift P and go to the Video streaming page.
  2. Choose your platform in the Platform box. YouTube and Facebook fill in their own address and lock it, because there is only one each.
  3. Create the RTMP stream on the Restream website FIRST. Drop Deck cannot do this part for you and the key alone is not enough.
  4. Copy BOTH the address and the key Restream gives you.
  5. Paste the address over the one in the Address box. Restream's address can differ by account and by region, so the one filled in is only the usual one.
  6. Paste the key into Stream key.
  7. Choose what to show in the Show box. A card with your station name on it is the default and is what most radio shows want. YouTube will not take sound on its own, so something has to be on the screen.
  8. Tick Go live here when I press Control B, so Control B sends the show to this platform rather than to your radio station.
  9. Press Check my stream key. It asks the platform whether your key is right and tells you what the platform said. It sends no sound and no picture, so nothing is broadcast and nothing is posted.
  10. Press OK, then Control B to go live. Control Shift B tells you what the stream is doing at any time.

What happens when you go live

Trying it safely

Picture quality

Custom RTMP server

Before you start

Setting it up

  1. Open Preferences with Control Shift P and go to the Video streaming page.
  2. Choose your platform in the Platform box. YouTube and Facebook fill in their own address and lock it, because there is only one each.
  3. Get the address and the stream key from whoever runs the server. There is no page for Drop Deck to open, because it is not a service Drop Deck knows.
  4. Type or paste the address into Address. It must start with rtmp or rtmps.
  5. Prefer an rtmps address if you are offered one. A plain rtmp address sends your stream key across the internet unencrypted.
  6. Paste the key into Stream key.
  7. Choose what to show in the Show box. A card with your station name on it is the default and is what most radio shows want. YouTube will not take sound on its own, so something has to be on the screen.
  8. Tick Go live here when I press Control B, so Control B sends the show to this platform rather than to your radio station.
  9. Press Check my stream key. It asks the platform whether your key is right and tells you what the platform said. It sends no sound and no picture, so nothing is broadcast and nothing is posted.
  10. Press OK, then Control B to go live. Control Shift B tells you what the stream is doing at any time.

What happens when you go live

Trying it safely

Picture quality

What goes on the screen

Knowing what the camera can see

This is the part of camera streaming worth having. You cannot look at a preview window, so the app tells you instead.

If something goes wrong

What it says What to do
It says the key was refused Check the key is pasted in full and has not expired. On Restream, check you created the RTMP stream on their website first: the key on its own is not enough. On Facebook, a persistent key can be refused if the live producer session has ended, so open it again.
It cannot find the server Check the address. For your own server it must start with rtmp or rtmps. Check you are online.
It says the connection is not keeping up Your upload cannot carry what you are sending. Lower Picture quality on the Video streaming page. A card needs very little; a camera at 720p wants about 2500 kbps.
The picture is frozen Drop Deck notices a camera that has stopped and shows your card instead, and says so. If it keeps happening, another program may be taking the camera: close OBS, Teams or Zoom.
The camera will not open Another program almost certainly has it. Close OBS, Teams or Zoom and try again. If Windows is refusing, turn the camera on for desktop apps in Windows privacy settings.
Nothing appears on the platform Check On air, Streaming location has the dot on My video platform. Without it Control B goes to your radio station instead, and Control Shift B will tell you which one it is set to.

28. Where the show goes

The key says where. Ctrl+B is your video platform and Alt+Shift+B is your radio station. The show goes to one place at a time, and whichever key you press is the one you get, so the summary in front of you before you go live is never for the other destination.

Each key is a toggle for its own place. If you are on air to your radio station and you press Ctrl+B, Drop Deck tells you where you actually are and does nothing else, rather than dropping one connection and opening another under you.

On air, Streaming location still lists both, with a dot beside the one that was used last:

Ctrl+Shift+B answers the same question at any time, on air or off.

The same choice is the Go live here when I press Control B box on the Video streaming page. They are one setting with two ways in, so ticking either moves the dot on the other.

Save as many setups as you run. The Audio streaming page has Save this station and Forget it, and Load a saved setup inside Streaming location lists them. A saved setup carries both pages and where it goes, so loading one can move the show from your radio station to your video platform. It says so when it does.

Changing where the show goes is refused while you are on air. Come off first.

29. Who is listening

Ctrl+Shift+A. Every stream on your server, how many people are on each, the most there have been at once, and what the server thinks is playing. It keeps itself up to date while it is open and reads the total out when it changes, rather than every fifteen seconds.

It works whether or not you are on air, which matters if your station carries on without you.

If you stream into automation, the server you send to is not the server people listen to. That is the normal setup: you connect to a harbor on one port, the automation decides what goes out, and the audience is on the streaming server behind it. Drop Deck looks in the usual places and says which one answered, so most people never set anything. If your listeners are somewhere else entirely, put that address in Where listeners connect on the Audio streaming page.

Nothing here needs a password. Icecast and SHOUTcast both publish their listener counts already. If your server says nothing, it is usually missing the status-json.xsl file Icecast normally serves, and whoever runs it can drop one in.

30. Recording

Recording the picture as well

Ctrl+R records the sound. Ctrl+Shift+R records the picture and the sound together, into one MP4, and neither of them needs you to be on air.

It records the same picture that would go out: your camera, your screen, your card, with whatever you have put on top of it. If you are already live it takes the identical frames the stream is sending, so it costs almost nothing. If you are not, it builds the picture itself.

Both keys now say what is really in the recording when they start, and that includes naming anything you can hear but that is not on the air and therefore not in the file. Ticking "Hear it yourself" on a source is monitoring; a recording takes the on air mix. Those are two different things and the app used to leave you to find that out afterwards.

On sync, since you cannot check it by watching: the sound is the master clock and the picture is fitted to it, so a camera that runs slow has a frame repeated rather than the timeline moving. Measured over three minutes of real recording, decoded back frame by frame: the offset held at about five milliseconds and drifted by 0.10 milliseconds per minute. With the camera unplugged for eight seconds mid recording it stayed in step and the file kept a constant thirty frames a second throughout.

The track list, for Mixcloud and anywhere else

Every recording now gets a .cue track list beside it, with the same file name: Drop Deck Stream 004.mp3 gets Drop Deck Stream 004.cue. It lists every track from your running order that went out and the moment it started, measured from the top of the recording.

Upload the pair to Mixcloud, or open the .cue in anything that reads one, and your tracklist is already written.

It is written as the show goes rather than at the end, so a show that stops unexpectedly still has the track list for everything it got through. Pads and sound effects are deliberately not in it: a track list with forty sound effects in it is not a track list.

If the app crashes, the file still opens. It is written so that a crash costs the last second rather than the whole recording.

A rough size: about 2.8 GB an hour, usually less.

Starting, stopping and pausing

Ctrl+R starts and stops it. The status bar says RECORDING while it runs, and RECORDING PAUSED while it is paused. Drop Deck says Recording started and Recording stopped every time, at every speech level, because a file being written makes no sound.

Alt+Shift+R pauses it. Press it again to carry on into the same file. Nothing that plays while it is paused is in the file, and the track list carries on from where you resume. If you are recording the picture as well, both pause together, so the two files still line up. It is on the On air menu as Pause recording.

It records the same mix that goes on air: every sound card, the running order, and your microphone if that is set to go out. The warning before a track ends and previews are never in it.

It does not need you to be on air. A show taped for later is the ordinary case. Recording and streaming run together and neither takes audio from the other.

Files go to Documents, in a folder called TG Drop Deck, named Drop Deck Stream 001 and counting up, so nothing you have already recorded is written over. On air, Open the recordings folder takes you there.

Preferences, Recording sets the format, the bitrate and the folder, and tells you what the next file will be called. WAV if the recording is going into an editor; MP3 for everything else, at about a tenth of the size. AAC and Ogg Opus are there too.

Closing the app finishes the file first, so a recording always opens.

31. Outputs and routing

Preferences, Output, or Ctrl+P.

The output device is where everything goes by default, including a virtual audio cable.

Each bank can go to a card of its own: beds to one output, effects to another, the running order to a third. That is how you feed a stream or a recorder while you keep listening on your own speakers.

Devices are remembered by name, so plugging in a headset later will not silently move your sound somewhere unexpected. If a device is missing when you open the app it says so rather than playing out of the wrong card.

The monitor output, on the Microphone tab, is where you hear yourself and where the end of track warning plays. Set it to your headphones and neither reaches the show.

Changing any of this stops what is playing and clears the decode caches, because cached audio was prepared for the old device's rate. The app warms them again immediately so the next key press does not pay for it.

32. Sending your show to another program

Sometimes the thing that needs your show is not a stream at all. It is TeamTalk, or Zoom, or Discord, or OBS, sitting on the same computer and wanting to hear you the way it would hear a microphone.

On air, Send this show to another program, or Alt+Shift+O.

What you need

One free virtual audio cable. VB-CABLE is the usual one and it costs nothing. It gives your machine two new devices: CABLE Input, which is a speaker nothing is plugged into, and CABLE Output, which is a microphone that hears whatever went into CABLE Input.

  1. In Drop Deck, press Alt+Shift+O.
  2. Tick Send this show to another program.
  3. Set Send it out of to CABLE Input.
  4. In the other program, set its microphone to CABLE Output.

That is the whole setup. Your speakers keep working, because the send is a separate output and not a re-routing of the one you listen on.

Drop Deck tells you which one to pick in the other program. The two ends of a cable are named from the cable's point of view rather than from yours, so the thing called Input is a speaker and the thing called Output is a microphone, and almost everybody reads that backwards the first time. Once you have chosen your output, the send page says it in as many words: "In the other program, choose CABLE Output as its input device."

Test the cable

The button at the bottom of the send page. Press it and Drop Deck plays a tone down the cable, listens to the other end of it, and tells you what came back. It takes two seconds.

This is worth doing once when you set the thing up, because a cable that is not working looks exactly like a cable that is. Windows will happily report a perfectly healthy connection that is carrying nothing at all.

Three things it can say.

It will not run while you are on air or recording, because it would put a tone into whatever call or stream the cable is carrying.

When Windows has the two ends set differently

Both halves of a cable have their own sample rate and bit depth in Windows, in two different tabs of a dialog nobody opens, and nothing in Windows keeps them in step. It is easy to end up with one end at 48000 hertz and the other at 192000 without ever choosing that.

It still works, because the driver converts. It just converts twice for no reason, and at the higher rate it carries four times the data for no benefit at all. Drop Deck now says so, with both numbers, on the send page and on Ctrl+Shift+O.

To fix it: Windows Sound settings, find each end of the cable, Properties, Advanced, and set both to the same thing. 48000 hertz is a good choice.

Drop Deck only ever says this about the device your show is going to. Your headset has two ends as well, and there is no reason its two halves should match.

Drop Deck Audio, on the Mac and coming to Windows

On a Mac, Drop Deck installs a virtual audio cable of its own called Drop Deck Audio: nothing to download, nothing to work out, one name that appears in every program's microphone list.

Windows will get it. It is not here yet, and the reason is worth knowing. On a Mac a virtual audio device is a plug in that sits in a folder. On Windows it is a kernel driver, and a kernel driver has to be signed by Microsoft before Windows will load it at all. That is a longer road than one release, and rushing it is how a free app starts tripping virus scanners.

Until then, VB-CABLE, and everything above is about making that painless.

Do not point the main output at the cable

This is the mistake worth naming, because it looks like the obvious answer and it breaks both halves of your show at once.

Preferences, Output is where your SOUNDS play. It carries your pads, your beds and your running order. Your microphone and every source you are catching are not on it: they live on the on air mix, which is what the stream, the recorder and the send all read.

So if you point that output at a cable, measured on a real machine:

Neither output has the whole show. Use the send, which does.

Where is everything going

Ctrl+Shift+W, any time, on or off air. It reads out the lot in one go: which card your sounds play from, which card you are listening on and whether it carries everything, where the show is being sent and what it is leaving out, and anything that is wrong with the arrangement.

There are three separate questions about where your audio goes and this is the one command that answers all three.

Give the send a card of its own

The card you send from should be doing nothing else. Share it with a bank, or with the card you listen on, and that card carries the show twice: the other program hears everything doubled a fraction of a second apart, which sounds like a strange hollowness rather than like an obvious fault.

Drop Deck refuses to let that pass quietly. It says which card, and which bank, and what to do about it.

What is in the send

Everything the audience of a stream would get: your pads, your music beds, your running order at full level, your microphone with its processing, and every audio source you are catching. It does not include anything you are only monitoring, so hearing yourself in your headphones does not send your voice twice.

You do not have to be on air, and you do not have to be recording. A send is wiring, not a broadcast.

Leave out whatever you are sending to

This is the one setting worth understanding, and it has a name in broadcasting: mix minus.

If you are also capturing the other program as a source, and most people are, then a send that carried everything would hand that program its own audio back. Everybody in the call would hear themselves a moment late, which is the oldest fault in radio.

So set Leave out of the send to the source you are sending to. If you capture TeamTalk as a source called TEAM TALK, and you are sending to TeamTalk, leave out TEAM TALK. Everything else still goes.

If the source is renamed or removed later, nothing is left out any more, and Ctrl+Shift+O says so in as many words rather than letting a call full of echo be a mystery.

Hearing what you are sending

Ctrl+Shift+H puts the send in your own headphones, exactly as it is being handed over. A radio desk calls this a confidence feed and it is the only way to be certain that what you think you are sending is what is going.

It goes to your monitor output only. It can never find its way back into the send.

Is it arriving

Ctrl+Shift+O says where the send is going, what it is leaving out, whether it is arriving without gaps, and whether your ordinary output is keeping up.

That last one matters more than it sounds. A virtual cable that is losing audio does not report an error anywhere: it simply drops small pieces, and what you hear at the far end is a little choppiness and a pitch that seems to wander. Drop Deck measures its own output rather than trusting the driver, which is why it can answer this at all.

Feeding OBS, Twitch or YouTube

The same thing, and there are two ways round it.

Use the send, as above, and add an Audio Input Capture source in OBS pointed at CABLE Output. OBS then gets your whole show, microphone included.

Or route by bank. In Preferences, Output you can send individual banks to their own sound card, so you could put only the drops on a cable and keep the music on your speakers. That is a mixing decision rather than a "give this program my show" one, and chapter 31 covers it.

If the sound is choppy or the pitch wanders

Check that both ends of your cable are set to the same sample rate. In Windows Sound settings, look at CABLE Input under Playback and CABLE Output under Recording, open Properties, Advanced on each, and set both to the same thing. 48000 Hz is a good choice. They are separate settings and Windows does not keep them in step for you.

33. DJing a live set

A DJ set is the running order doing the work and your hands free for the drops.

Set the decks up once.

  1. Put your music in the running order. Ctrl+V pastes a list of files straight from Explorer, or use Playlist, Add songs to the end.
  2. Set a crossfade. Playlist, Crossfade between tracks sets the default for every track; Alt+Enter on one row overrides it for that track alone. Six to ten seconds suits most music, and two or three suits speech.
  3. Put your stabs, horns and shout-outs in bank 1, and your longer beds in bank 3.

Running it.

The crossfade is measured from where the MUSIC stops, not where the file does. Most MP3s carry a second or two of silence on the end. The app measures each track's real ending once, in the background, and cues the next one against that, so a segue does not disappear into digital silence. You do not have to do anything for this; it is why the segues sound tight.

Talking over it. Ctrl+M opens the microphone and everything musical ducks underneath while it is open. It ducks by being open rather than by listening for your voice, so it never clips your first syllable and never un-ducks because you paused for breath.

A tip worth the thirty seconds. Put a bed on a looping slot in bank 3 and it will hold under you for as long as you talk. Bank 3 is the looping bank, whatever you rename it to.

34. Recording a podcast

A podcast is the same machine with the stream turned off and the recorder turned on.

Before you start.

  1. Ctrl+Shift+M for the microphone: pick the device, set the gain, and turn monitoring on if you want to hear yourself in headphones.
  2. Go to Preferences, Voice and set the compressor and the gate. A light compression and a gentle gate is the usual answer, and it is applied to what is recorded as well as what goes out.
  3. Preferences, Recording sets the format, the bitrate and the folder. The page tells you what the next file will be called, so it is not a surprise afterwards.

Recording. Ctrl+R starts and stops. You can record and stream at the same time; they are separate.

What gets recorded is the programme: everything, whatever sound card it was routed to. Your monitoring is deliberately not in it, because you hearing yourself in headphones is not part of the show.

Guests. A guest on Zoom, Teams or TeamTalk goes in through Alt+Shift+S, Audio sources: choose One program and pick the app they are in. That takes only that program's audio, so your own notifications do not go into the recording. See chapter 20 for the other kinds.

Two habits that save re-records.

35. Running a radio station

This is what the app was rebuilt for.

Connecting to your station. Preferences, Audio streaming, or On air, Set up streaming. You need the server type (Icecast, which also covers a Liquidsoap harbour, or SHOUTcast), the host, the port, the mount point, your user name and your password, and a bitrate. Your station's administrator has all of these.

Your password is kept in Windows Credential Manager, not in the board file, so a board you share does not carry it.

Going on air. Ctrl+B. Before it connects it puts a summary in front of you saying where the show is about to go and anything wrong it can see: a missing file, a picture that will not load, a title too long for the screen. Go live is the default button, so Ctrl+B then Enter is the whole gesture. If you would rather not be asked, there is a box in that window to turn it off, and a way back on in Preferences, Audio streaming.

Where the show goes is On air, Streaming location. One list, one dot against the one that is live. If you have a radio station and a video platform both set up, this is how you choose between them, and it is worth looking at before every show.

While you are on air.

Saved setups. If you present on more than one station, save each one under Streaming location, Set these up. Loading one fills in both the audio and the video pages in a single go.

A rehearsal worth doing once. Set up a private mount on your server, go live to it, listen back on a phone, and check your levels there rather than in your own headphones. Your headphones are the one place a level problem does not show up.

36. Going live on YouTube

What you need. A YouTube account with live streaming enabled, which takes 24 hours the first time, and your stream key from YouTube Studio.

Setting it up. Preferences, Video streaming. Choose YouTube Live and paste the stream key. Everything else has a working default: 1280 by 720 at 30 frames a second, a keyframe every 2 seconds, which is what YouTube asks for.

YouTube will not accept a stream with no picture, so the app always sends one. Choose what it is with Alt+Shift+V:

You can change between them while you are live. The connection is not dropped and the sound does not break.

Before you go live, in this order.

  1. Alt+Shift+V and choose the picture.
  2. Alt+Shift+T to put your station name, what is playing, or a clock on top of it.
  3. Alt+Shift+C if you want your own colours.
  4. Ctrl+Shift+F to hear whether you are actually in shot, centred and lit.
  5. Alt+Shift+D to have the whole thing looked at and described back to you.
  6. Ctrl+B.

While you are live, Ctrl+Shift+V reads out what is on the screen, and the app tells you by itself if the picture goes black or freezes, which is the failure you would otherwise never know about.

Facebook and Restream work the same way; choose them on the same page. Restream is worth knowing about because it takes one stream from you and sends it to several platforms at once, which is easier on your upload than running two encoders.

37. Streaming to more than one place

Audio and video are separate destinations, and Ctrl+B goes to one of them. That is what On air, Streaming location is for. It is a deliberate choice rather than both at once, because sending a full video stream when you meant to send audio to your station is an expensive accident.

To reach several video platforms, use Restream or a similar service. Send one stream from Drop Deck and let them fan it out.

To reach a radio station and a video platform at the same time, run the audio stream from Drop Deck and feed the video from a second machine, or use your station's own relay. One machine sending two full streams is where upload bandwidth usually runs out.

To feed OBS instead, see chapter 32. In short: send Drop Deck's output to a virtual audio cable and have OBS listen to that. This is the right answer if you want scenes, alerts and overlays that Drop Deck does not do.

Routing audio to more than one place at once is chapter 31. A bank can have its own sound card, so your beds can go to a desk while everything else goes to the encoder. The encoder always takes the sum of everything, whatever it was routed to, so nothing falls off the stream because you sent it somewhere else.

38. Boards and files

A board is your eighty slots, your running order, your settings and your stations, in one file.

Key What it does
Ctrl+N A new board, eighty empty slots
Ctrl+S Save
Ctrl+F12 Save as, so you can keep one board per show
Ctrl+O Open one
File, Export board with its sounds One folder holding the board and a copy of every sound it plays

The board saves itself a couple of seconds after anything changes, and again when you close the app. If it cannot be saved on the way out it asks, rather than losing your evening quietly.

Drop Deck opens the board you last had open. Open one with Ctrl+O, or keep one somewhere with Ctrl+F12, and that is the board you get next time, with its streaming and everything else. If it has been moved or renamed since, you are told which board opened instead, and Ctrl+O finds it.

A board changed outside Drop Deck is never saved over. Copy a board file into place while Drop Deck is open and come back to the window: with nothing playing, the new board simply loads. If you had changed things on screen as well, those are kept beside it under a new name, so neither is lost.

A board that will not load is kept, and you are told. It is copied aside before anything else happens, with "could not be read" and the date in its name, and you hear what was wrong with it.

Taking a board to another computer

File, Export board with its sounds makes one folder holding the board and a copy of every sound it plays: the pads, a folder pad's sounds, the running order and the drops library. Put that folder in Google Drive, Dropbox or OneDrive, or on a USB stick, and open the board from it on any computer, the Mac included. Everything plays straight away. Choose where it goes and a folder named after the board is made there; it says how many sounds it copied, how big they are, and anything it could not find.

A board opened on another computer looks beside itself first. Before you are asked to relink anything, it looks for missing sounds in its own folder and the folders just under it, and says how many it found. It only takes a file with exactly the same name, and it never writes what it found into the board, so the computer the board came from still finds its own sounds; use Relink to make a new place permanent. A sound kept in the board's folder is also saved relative to the board, so a board and its sounds kept together in one cloud folder open wherever that folder does.

File, Relink missing sounds repairs a whole board against a new location in one pass, and does the running order and your drops library at the same time. Move your sound folder and this is the fix. It reads a board made on the Mac as well as one made here.

File, Load the demo pack brings the included forty sounds back at any time.

File, Import an old soundboard bank reads a bank from the soundboard this app replaces.

Where Drop Deck keeps your settings

Your board, and everything Drop Deck remembers, normally live in %APPDATA%\TG Studios\TG Drop Deck.

The portable copy keeps them beside itself instead, in a folder called userdata next to TG Drop Deck.exe, so the whole program is one folder you can move, copy to a memory stick, or delete. The first time a portable copy starts, it copies whatever was in AppData into that folder and tells you the old one can be deleted. Nothing there is removed for you, and the board you had open is followed into its new home.

An installed copy can do the same: make a folder called userdata beside TG Drop Deck.exe and it is used from the next start.

Stream keys are in neither folder. Windows Credential Manager keeps those, for one person on one computer, so a copy of the folder taken to another machine asks for the stream key again.

39. Every setting

Ctrl+P opens Preferences. Eight tabs, and everything is applied on OK whichever tab you were looking at.

Output. The output device, and a device of its own for each of the four banks.

Sounds and beds. Ducking on or off and how deep. How many presses of Escape stop everything, one to four. Whether stopping fades or cuts. How long music beds take to fade in and out.

Playlist. The crossfade between tracks. Whether to warn before a track ends and how many seconds before. Which of the six warning sounds, and how loud.

Microphone. Which microphone, which channel the voice is on, gain, which output you hear yourself on, and whether you hear yourself.

Voice. The gate, high pass, equaliser, compressor and limiter, and a VST3 plugin with its own settings. See chapter 18.

Audio streaming. Saved stations, server type, address, port, mount, user, password, format, bitrate, station name, where listeners connect, whether the microphone goes on air, whether track titles are sent, whether the stream is listed publicly, whether the playlist fader is monitor only, and a button to test the connection.

Video streaming. Platform, address, stream key, and whether Ctrl+B goes here rather than to your radio station. What to show: a card, a picture of your own, or a camera, with the camera, the size and the picture quality, and whether a clock goes on the card. How much the app tells you about the shot. Buttons to fetch your stream key, test the connection, look through the camera and read the setup instructions. See chapter 27.

Recording. Format, bitrate and folder, and what the next file will be called.

Speech. How much the app says of its own accord, and whether it announces what is playing.

Cancel means cancel on every tab, including Voice, where everything you do changes the microphone you are listening to as you do it.

40. Every key

F1 inside the app is this list, and it is always right for the version you are running.

Sounds

Key What it does
1 to 0, Shift+ Sound effects 1 to 10, 11 to 20
Ctrl+digit, Ctrl+Shift+ Dialog drops
Alt+Ctrl+digit, +Shift Music beds
Ctrl+Tab Next bank
F2 Rename the sound you are on
Ctrl+F2 Rename the bank
Alt+Enter Properties
Delete Clear the slot
Shift+Delete Take the slot off the board
Ctrl+Space Stop the last sound
Escape twice Stop everything
Ctrl+L What is playing
Ctrl+F, Ctrl+E Search
Ctrl+G Global hotkeys on or off

The running order

Key What it does
Ctrl+Shift+P / Ctrl+Shift+S Playlist / soundboard
Ctrl+Alt+Tab Swap between the two
Enter Play from here
Shift+Enter Cross into the track you are on
Space Tick or untick
Shift+A / Shift+U Tick all / untick all
Alt+Up / Alt+Down Move by one
Alt+Home / Alt+End Move to the top or the end
Delete Take it out
Ctrl+V Paste songs in
Alt+D A random drop from the library
Ctrl+Shift+D A drop from a file
Ctrl+Shift+L Go to what is on air
Ctrl+Shift+Enter Play from here, from anywhere

Levels

Key What it does
F3 / F4 Sound effects
F5 / F6 Music beds
F7 / F8 Running order, and the monitor fader while on air
Ctrl+D Ducking on or off

Your voice and other audio

Key What it does
Ctrl+M Microphone on or off
Alt+Shift+M Microphone on or off, when another program has Ctrl+M
Ctrl+Shift+M Microphone and voice settings
Alt+Shift+S Set up audio sources
Alt+Ctrl+Shift+S Source control: mute, solo, rename, remove
Alt+Shift+O Send this show to another program on this machine
Ctrl+Shift+H Hear what is being sent
Ctrl+Shift+O Is the send keeping up
Ctrl+Shift+W Where is everything going
Ctrl+Shift+R Record the picture as well as the sound
Ctrl+Shift+C The cue sheet: what is coming up

On air

Key What it does
Ctrl+B Go live, and come off air
Ctrl+Shift+B What the stream is doing
Ctrl+Shift+A Who is listening
Ctrl+Shift+F What the camera can see: in shot, centred, lit
Alt+Shift+V Change the video source, on air or off
Alt+Shift+T Screen text: what goes on top of the picture
Alt+Shift+C Colours: your background, words and accent
Alt+Shift+D Check my shot: have the picture described to you
Ctrl+Shift+V What is on screen right now
Ctrl+R Start and stop recording
Alt+Shift+R Pause and resume the recording

Files

Key What it does
Ctrl+N, Ctrl+O, Ctrl+S, Ctrl+F12 New, open, save, save as
Ctrl+P Preferences
F1 This list, inside the app

41. Questions people ask

Does it cost anything? No, and it will not. There is a donate page if it earns a place in your show, and nothing is held back from anyone who does not.

Do I need to install it? No. The zip runs from anywhere, including a memory stick, and updates itself in place.

Will it work with JAWS? It is built and tested with NVDA in mind and uses standard Windows controls throughout, so it should. If you use JAWS and try it, please say how it went.

Can I use the included sounds in my show? Yes, for anything you make. They were generated with AI rather than taken from a commercial library, which is what makes them safe to hand out.

Can I go out on YouTube without being on camera? Yes, and it is what most people do. Drop Deck draws a card with your station name and whatever is playing. YouTube will not take sound on its own, so something has to be on the screen, and a card costs almost nothing. See chapter 27.

Can I stream to my radio station and to YouTube at the same time? Not yet. Ctrl+B goes to one of them and you choose which. Restream is the way to reach several places at once today: send to Restream and it passes the stream on to everywhere you have switched on there.

How do I know I am in shot? Press Ctrl+Shift+F. It says whether a face is in the picture, whether it is centred, whether you are close enough and whether the light is good enough. It answers whether or not you are on air.

Why does pressing a key twice play the sound twice? Because that is what a soundboard is for. If you want one particular sound to stop instead, tick Pressing its key again stops it in its properties. Music beds already work that way.

Why does Escape need two presses? Because one key that silences a live show is one key away from silencing it by accident. Set it to one in Preferences, Sounds and beds if you would rather.

Does turning my microphone off stop my audio sources going out? No. They are separate on purpose. Mute a source in Source control, Alt+Ctrl+Shift+S.

Can I record without being on air? Yes. Ctrl+R works whether or not you are streaming, and both run at once without affecting each other.

Can I put a Zoom or TeamTalk call on the air? Yes, two ways. Add it as a program source, which needs nothing set up in the other program, or point it at a virtual audio cable and add the cable as a device source.

Will my listeners hear my screen reader? Only if your screen reader is going to the same output as the show. Give the show its own output, or your monitoring its own, and they will not. While you are on air, F7 and F8 pull the music down for you alone so you can hear yourself think.

Where are my recordings? Documents, in a folder called TG Drop Deck. On air, Open the recordings folder takes you there.

Where do my settings live? AppData\Roaming\TG Studios\TG Drop Deck. Not in the program folder, so an update never touches them.

42. When something is wrong

A sound will not play, or says it is missing. The file has moved. File, Relink missing sounds repairs the whole board, the running order and the drops library in one pass.

No sound at all. Check Preferences, Output. If a device was unplugged the app says so on startup rather than playing out of the wrong card. Check the three faders in the status bar too.

I can hear a program twice. Turn off Hear it yourself for that source. The sound is already reaching your speakers from the program itself.

YouTube or Facebook will not take my stream key. On Restream, create the RTMP stream on their website first and copy BOTH the address and the key: the key on its own is not enough. On Facebook, a persistent key can be refused once the Live Producer session has ended, so open it again. Everywhere, check the key is pasted in full.

The picture is choppy, or the app says the connection is not keeping up. Your upload cannot carry what you are sending. Lower Picture quality on the Video streaming page. A card needs very little; a camera at 720p wants about 2500 kbps.

The camera will not open. Another program almost certainly has it. Close OBS, Teams or Zoom. If Windows is refusing, turn the camera on for desktop apps in Windows privacy settings.

It says I am on air and nothing appears on my channel. Press Check my stream key on the Video streaming page first. It sends no sound and no picture, so it cannot put you on the air, and it tells you what the platform said.

If the platform accepts the key, your stream is arriving and something at their end is holding it back. On YouTube the usual reason is auto start switched off for that stream, which leaves YouTube waiting for you to press Go live in your own live control room. After that, check the broadcast is not set to private, and that you are looking at the same channel the key came from: one Google account can hold several channels and a key belongs to one of them. The first time you ever turn live streaming on, YouTube can take up to twenty four hours.

On Facebook nothing is posted until you press Go Live Now on Facebook's own page, so a preview and no post is Facebook working normally.

It says I am on air and I am not sure anything is going out. Press Ctrl+Shift+B. It says how long you have been on air and, when nothing has left the machine yet, it says that too. That usually means no sound is reaching the stream: check something is playing and that your microphone is on the air.

The stream will not connect. Use Test the connection on the Audio streaming page. It says exactly what the server said. The usual causes are the source password rather than the listener password, a mount point with no leading slash, and for SHOUTcast, giving the source port instead of the listening port.

Who is listening says nothing. Your server may not publish status-json.xsl. If your listeners are on a different address from the one you stream to, put it in Where listeners connect.

Ctrl+M does nothing at all. Another program has taken the key. Use Alt+Shift+M, and see "If Ctrl+M does nothing at all" in the microphone chapter for how to find out which program.

The microphone will not open, or opens and I am not heard. Drop Deck says which. The usual fix is Windows Settings, Privacy and security, Microphone, with Let desktop apps access your microphone turned on, and then checking the microphone is not muted in Windows Sound settings.

My streaming set up has gone. Drop Deck opens the board you last had open, and says so at startup if that board has moved. Ctrl+O opens it again.

A program source is silent. Check the program is running, and that it is actually playing something. If it was closed and reopened, Drop Deck finds it again by itself within a few seconds.

Program capture is not offered. It needs Windows 10 build 20348 or later. Use a virtual audio cable instead.

The app will not open, or opens and does nothing. Only one copy runs at a time; a second launch brings the first to the front. Check the taskbar.

Something else. Help, Submit feedback sends a report with your settings attached, which usually answers the question on its own. It never includes file names, sound names or paths.

43. Updates, feedback and licence

Updating. Help, Check for updates. If there is a newer version you are told what is in it and asked whether you want it. Say yes and a progress window opens: it draws a bar, says the percentage out loud as it climbs, and has a Stop the download button that really stops. When the download finishes the app closes, the installer runs, and the app opens again by itself.

Your board, your sounds, your hotkeys and your settings are all kept.

Help, Check for updates looks for a new version and tells you what changed before it downloads anything. Updates are signed, and one that does not match its signature is thrown away rather than installed.

An installed copy downloads an installer and replaces itself.

A portable copy downloads a zip and replaces itself in place: it unpacks beside, waits for the old copy to close, replaces the folder and starts the app again from the same path, so shortcuts still work. If the folder cannot be written to, it unpacks alongside and says where.

Help, Submit feedback sends a message with your version, your settings and what the app has been doing. Never a file name, a sound name, a bank name or a path.

Help, User manual opens this page. F1 is the key list inside the app, always right for the version you have.

Free and MIT licensed. Source on GitHub.

Get in touch if this manual does not answer your question, and I will fix the manual as well as the problem.