Setting up audio for livestreaming and recording
- Windows settings
- The physical amp
- Other common audio mistakes
- Virtual audio
- Configuring audio devices
- Keeping track of all audio settings
- Troubleshooting software audio
- Troubleshooting hardware audio
- What next?
This guide was written with a single-PC setup as the main focus. There are some specific tips for a dual-PC setup, however.
Setting up your audio for advanced streaming is a total pain. Why even bother doing more with your audio you ask?
I touched on this a little in the tools guide. It’s perhaps best explained with this screenshot of OBS Studio’s mixer interface:
This gives you an idea of how you can now manage, adjust, and mute different audio sources such as your music; you can hide it from your viewers altogether if you want.
It’s probably important to know that my (physical) playback device is a pair of headphones, Sennheiser 600, connected to my amp, Scarlett Solo by Focusrite. So replace that with whatever you use instead.
On Windows 10, right-click the volume icon and pick Sounds.
Here are my settings:
- Format: 2 channel, 24 bit
- Sample rate: 48 kHz
- Exclusive Mode: disabled
- Format: 1 channel, 24 bit
- Sample rate: 48 kHz
- Default: Amp (Line In)
- Levels: 100
- Exclusive Mode: disabled
(Same settings for the virtual audio cables, just in case.)
If you’re a moron like me, you might make the mistake of setting your recording channels to 2 when it should be 1.
With my audio interface and microphone, this results in recording audio on only the left speaker. The right one will get no audio.
Furthermore, a lot of dopes will tell you to fix your microphone outputting to only one side by changing your OBS output to Mono.
This is wrong for two reasons, the first of which is that you’re applying a local, per-app, fix to a global issue. The problem will still persist outside OBS, which is why you should change this in your Windows Sound settings.
The second reason this is a terrible idea is because OBS’s Mono setting averages the volume between your left and right channel.
What’s the average between one active speaker and a silent one? If your answer is half the volume, your passed the maths test.
How’s that for downmixing.
Your devices may support different kHz formats; common ones are 44.1, 48, and 96 kHz.
I recommend going with 48 kHz over 44.1 if your hardware supports it, as it’s more or less the default nowadays.
Whatever you set your recording sample rate to, make sure it’s the same across all software and devices. If you start getting really weird and choppy audio, you probably set your sample rate to something different in multiple places. More on that when using VoiceMeeter in this video:
Also remember that ASIO is better than WDM, which is not covered in the video.
For 16 bit vs 24 bit, here is Focusrite’s own knowledge article “Is my Scarlett interface in 16 bit, or 24 bit on Windows?”.
When using the WDM driver, only 16 bit modes are seen as an option in the Advanced section of the interface’s Properties. However, this setting has absolutely no effect when recording using the ASIO driver, which users should always use in DAWs, due it’s its low latency and increased peformance over WDM (Windows’ own audio driver).
The audio stream captured from our USB 2.0 hardware is always 24-bit samples. These samples are not altered or processed in any way by the driver when recording via the ASIO driver. Windows audio control panel may say the device is in 16 bit mode, but the ASIO driver bypasses the Windows audio system, and is therefore not affected by this setting.
Therefore, using 24 bit depth in your recording software, while using the ASIO driver, is still 24 bit when recording/creating, where 24 bit actually matters.
Be sure to check out Podcastage’s video “16 bit vs. 24 bit Audio, What Should You Record At?”:
If you use anything that goes through the USB port for your microphone and/or headphones, your settings and driver might reset if you move them to a different port than you used to, so make a habit out of checking them every now and then.
I’ve had the above settings reset for me a bunch of times, perhaps because I switched my amp to a different USB port, so make a habit out of checking your settings every now and then.
You may also find yourself missing sample rates in Sound. Because of this amp, I know I’m supposed to have 44.1, 48, 96 kHz at 24 bit. Unfortunately, I’ve had cases where I messed up the drivers and had to fix it. I learned that I have to install the driver a certain way before I would see all my sample rates:
- Uninstall the amp driver software
- Called “Scarlett USB 2.0 Audio Control Panel”
- Shut down computer
- Unplug amp
- Turn on computer
- Install software
- Shut down computer
- Plug in amp
- In the USB port I’m planning on using for perpetuity
- Turn on computer
- All the sample rates should be there now
- Go to the Scarlet control panel and set it to Recording
Finally, open your Volume Mixer (by right-clicking your volume icon) and make sure that the levels for all your apps are par with your audio device on the left. Sometimes, they’ll be off by just a few ticks.
- When Windows detects communications activity: Do nothing
- It screws up your Volume Mixer something fierce
I haven’t figured out whether you need to set the Volume/Level to something specific. When you set VoiceMeeter to your default audio device, you should expect the volume to be very loud, as a precaution.
I’m using my regular Line Out recording device which gets the job done and works fine with VoiceMeeter. However, when I use VoiceMeeter as my recording device, my mic stops working in Overwatch, even after I follow Blizzard’s audio guidelines:
Remove the ability for applications to take exclusive control over an audio device.
- Right click on Window’s volume icon in the system tray
- Open Playback devices
- Select the device you want to use in game
- Open the Properties for this device
- Click on the Advanced tab
- Remove the check box for “Allow applications to take exclusive control of this device”
- Repeat the process for Recording devices (step #2 above).
Set everything to 100 or max, unless it’s an application that goes beyond 100% like VLC player.
You’ll want to manage volume in as few places as possible. If you have to manage audio settings in multiple applications, you should still reset to 100 and start from the top and work your way down more granular options.
The physical amp
I generally try to make sure my mic is as loud as possible without any boosts. On a basic amp, you’ll want to turn up your gain as much as possible—unless your mic is always on and you don’t want people to hear stuff in the background.
Some amps and mixers have a gain LED. Mine flashes red when your input signal is clipping during a combination of high gain and high voice. The higher the gain, the lower your voice has to be. And vice versa.
What you want is to stay just comfortably below clipping and triggering the red light using your usual vocal range when recording1. So talk into your microphone with the vocal range and closest positioning you’d normally use in a recording and turn up the gain as much as possible without triggering the light.
Some amps and mixers also have gain LEDS with a yellow colour to settle you into the optimal level while alerting you if your settings might be too low.
Read the the OBS Project’s “Understanding the Mixer” for a more information. This image from the guide summarizes the ideal levels:
Amp software settings
You should download whatever software comes with your amp. If no software is available or you don’t have an amp, check out ASIO4ALL to get ASIO support for use in VoicMmeeter Banana below.
They are listing in descending order of how much CPU power they require to reduce latency. Since this is 2017, my CPU does just fine, so I’m going with “Recording”.
If you get a chance to increase the buffer size of you audio, I’d consider just cranking it to the max of 1024 from a default of 256. This gives your computer a lot more wiggle room to handle your audio, but your CPU has to be up for the task. If it ain’t, you’ll notice some audio distortion or stuttering.
If you use VoiceMeeter, be sure to watch the video on lowering audio latency from above, and remember to use ASIO over WDM for audio whenver possible:
A warning about boosting your headphone audio
From the OBS guide:
Try listening to a song or podcast while recording in OBS. Now compare the original volume with the one in your recording.
There are many reasons why the audio might be completely different—check out the audio guide for that.
One might be that you have an audio interface (amp) like a Scarlett Solo that boosts up your headphone audio after receiving it from the computer. Viewers will only hear the original audio.
The simplest thing you can do is to just try turning down the amp’s headphone dial as much as possible while offsetting it by turning up your Windows volume. If you’re already at 70% volume in Windows, you’ll have to do something like boost OBS’s desktop audio in software with something like OBS’s mixer.
This covers the bases well:
Other common audio mistakes
This is the software you’ll need:
|VB-CABLE2||Create virtual playback devices to split audio to|
|VoiceMeeter Banana||Route and control audio in one place|
|VoiceMeeter Potato||Alternative to Banana with more audio channels|
Audio channels (audio in):
- Comms (Discord)
- Music (iTunes/Spotify)
Or, with suggestions for more audio channels:
- Comms (Discord)
- Copyright (Music, other channels, YouTube videos, local video)
- Game, to isolate alert sounds
- Or OBS, another way to mute or isolate alert sounds
So, we need to change the output device of my Comms (1) and Music (2) programs to each of our virtual audio cables, which lets us split them up into separate audio channels.
Because we installed the free VB-Cable and threw them some money to get two more virtual audio cables, we the following VACs:
- CABLE (freeware)
- CABLE-A (donationware)
- CABLE-B (donationware)
We won’t be using the third one.
Discord is first. Change the output device to “CABLE Output (VB-Audio Virtual Cable)” in the Voice settings.
Configuring audio devices
Next you have to change the output device for the music players. Most apps don’t have manual settings for audio devices, which means you have to resort to another way of directing the audio channel.
Update Mar 28, 2018: Come April 2018, Windows 10 will have native support for per-app audio device settings:
On Windows 10, you’ll want to use this instead of Audio Router mentioned below.
Unlike Discord, apps such as iTunes
and Spotify do not have a setting for input and output device.
Update: As of December 21, 2017, Spotify has support for changing the audio output device—albeit in a roundabout way.
Windows 8.1 and older
For this, download Audio Router.
Click the button underneath your music players, choose “Route”, and pick “CABLE-A Input (VB-Audio Cable A)”.
Be sure to check the levels while you’re at it; I set all mine to 100% for the hell of it, but you go be you.
Unfortunately, Audio Router does not save the settings when closed. But hey, it’s free.
Audio Router is unfortunately no longer in development, but it was made as a free alternative to CheVolume which costs money. As with most tools, start out with the free ones and graduate to the paid when you’ve mastered the skill and feel that it’s for you. CheVolume does have a seven-day trial, though.
Unfortunately, the same thread as I just linked suggests that a November 2015 update broke CheVolume, and it’s kind of hard to communicate with the creator. I tried installing it, and it’s utterly broken and useless. Just a painful reminder that great, important tools don’t invent and support themselves.
Best of luck out there.
Update Feb 02, 2018: This YouTube guide also seems to do a decent job of explaining the set-up of VoiceMeeter—can’t necessarily vouch for the settings in the guide, though:
(The only mute in the image above that serves a purpose is the Mic one, but the other mutes make it easier to remember how I’ve set up VoiceMeeter.)
Check out the VoiceMeeter Banana manual from page 8 to see the basic setup guide.
Hardware out (top right):
- A1: ASIO (Amp)
- Choose your regular audio device; preferably ASIO, WDM second3
This is where you receive all your audio streams. In my case, my headphones. All the audio channels combine (mix) into this playback device.
If you want to use both a headset and, say, speakers, you can set your speakers to A2. Maybe you want to hear your comms in your headphones and everything else from your speakers.
- WDM: Line In (Scarlett Solo)
- WDM: CABLE Output (VB-Audio Virtual Cable)
- WDM: CABLE-A Output (VB-Audio Virtual Cable)
- System tray (Run at startup)
- Hook volume keys (For level output A1)
Otherwise your volume buttons won’t work
Therefore you need to map it to a specific virtual audio out.
Keep in mind this will also override your regular volume controls.
This feature is also a little … hit or miss. You can tell it’s working if you don’t see the Windows volume slider. If it does, VoiceMeeter slipped up, and you’ll just have to wait for it to rejigger or something. You also can’t use Volume Up/Down to exit Mute, as I now have to use my Mute-toggle shortcut instead of one of the volume shortcuts.
- System settings
- Set Preferred Main SampleRate to your, well, preferred main sample rate; mine is 48,000 Hz.
- Engine Mode can be switched from Normal to Swift to reduce latency, but it might be experimental.
Notice how the microphone is muted in the screenshot? If you unmute it and keep sending its output to A1, noted by the highlight, you’ll be able to listen to your own microphone. This is an extremely useful feature you’re probably going to be using in your default workflow.
Last, but not least: always try to manage your volume controls in as few places as possible.
Virtual audio over Ethernet
While your at it, you can also cut out a lot of audio hassle with mixers and whatnot by sending your audio to your streaming PC over Ethernet using VBAN. Here’s a quick how-to video:
ASIO with OBS
To add integrated ASIO support in OBS, check out obs-asio. The settings in it don’t automatically update, so make sure it’s the same across different software.
You set it up by adding an ASIO source to your scene after configuring your settings.
I haven’t had a chance to play around with it yet; by the time I get aroun to it, we might have native ASIO support in OBS.
Audio monitoring with OBS
v18.0 of OBS Studio enables audio monitoring. Audio monitoring allows you to listen to an audio source, most important of which is your own microphone, which is incredibly useful.
Here’s how they explain it in the release notes:
Added audio monitoring on Windows/OSX. Audio monitoring allows the ability to listen to the audio of a source, and can be enabled via the advanced audio properties. You can set it to monitor without outputting, or monitor and output. The device used for monitoring can be changed in advanced settings.
I’ve got three ways to do audio monitoring:
- Turn it on in VoiceMeeter (as aforementioned)
- Flip the physical “Direct Monitoring” switch on my amp
- Turn it on in OBS Studio
To turn it on in OBS Studio, go to your mixer settings, and you’ll find the new option:
Not everyone needs to use audio monitoring, but I recommend you turn it on just for a few minutes for testing as it’s a dead-simple way to find a ceiling for your gain controls for background noise.
Since I manage my audio settings in VoiceMeeter, I’ll keep using that for audio monitoring, but this is an extremely useful tool for streamers who now won’t have to go through a lot of the crap outlined in this guide.
v18.0 also adds these new audio processing features you might want to check out:
audio compressor filter
audio filter for Windows/OSX. Allows applying VST plug-ins (only up to version 2) to filter audio
VST plug-ins are basically the go-to plug-ins for audio processing, and they can be very useful for cleaning up your audio, if you’ve got a shoddy microphone setup.
Keeping track of all audio settings
Wouldn’t it be great if you only had to set up all your audio settings in one place? If you think I’m going to tell you a way to do that, I’ve got bad news for you. As the README for obs-asio says:
Important: make sure the settings selected are those of your device as set in the Device Asio Control Panel (from its maker).
The settings set in OBS MUST reflect those or the plugin won’t work.
Instead, here is a checklist of the general software that requires configuration:
- Windows Sound settings
- OBS Studio
- Audio, amp, and ASIO drivers
- Eg ASIO4ALL or Focusrite’s ASIO Control Panel
- Music and audio-editing apps
If you’re still scratching your head, check out the
audio section of my general settings.
Troubleshooting software audio
The easiest way to see if something up is to get LatencyMon. It’s very straightforward and tells you in plain English if something’s up.
- If your audio crackles, you’ve got mismatched sample rates
- If your audio cuts out, your buffer size is too big
Troubleshooting hardware audio
Check out “How to get rid of hum and eliminate other noises from your audio and video systems” by TechHive.
- Do virtual audio cables affect audio quality?
- Dunno. Probably.
- Does this setup affect audio latency?
Yes. Here is what VoiceMeeter’s vburel said:
Voicemeeter adds delay according latency (buffer size) used by audio device.
[T]o minimize this delay we recommend to use ASIO device as output A1 (with a buffer size between 128 and 256 samples).
[I]t is also possible to reduce the latency with WDM/KS device in Voicemeeter by decreasing preferred buffer size, but it needs specific expertise to find optimal value (set back to default values in case of problem).
In optimal case, Voicemeeter could playback with less latency than windows system itself. REM: some devices (like USB or Bluetooth) can add big delay anyway (whatever the buffer size used by the stream).
Even some audio pro board with ASIO driver can have a 10 to 20 times more delay than the one given by buffer size (pending on how the driver has been programmed).
To resume, latency is pending on different factors, and in some cases cannot be reduced.
- What can I do about latency?
- Your audio buffer size might be too large; try lowering it.
- In System settings of VoiceMeeter, try changing Engine Mode in the bottom right from Normal to Swift.
- How do I calibrate audio delays to compensate for latency?
- Put on a counter including miliseconds in the middle of your OBS recording and maybe use another app like Audacity to see where exactly the audio is active relative to the counter’s timestamp.
Congrats on finishing what I consider the most complicated of the guides so far. The Discord guide should be more fun.
“Your usual vocal range when recording” sounds like a weird phrasing, but there are people who think it’s cool to scream and whatever for their YouTube or Twitch.
Conversely, there are people who’ll test their mic by putting their mouth close to the mic and enunciating loudly while whisper-mumbling with side address when they’re actually recording.
Optimize for how you actually record. ↩︎
You’ll need to pay up to get more than the one free virtual audio cable. Also don’t forget to install as administrator. ↩︎
Quoth page 10 of the VoiceMeeter manual:
We recommend to select first ASIO (if exists) or WDM device type to get best Latency (WDM / KS playback device are used in exclusive mode per default, bypassing the windows mixer and possibly its Volume Control).