Tips for immersive video calls
benkuhn.net
benkuhn.net
* Is the call 1:1 or extremely small? If so, it's down to the preferences of the people in the call. Otherwise, for larger calls:
* Are you in a quiet environment? * That was a trick question. You are not in a quiet environment. You think you are, because your human brain is good at filtering out background noise. Your microphone is not. You are not being forced to actually listen to what your microphone hears. The rest of us are. Mute your fucking microphone!
Sounds like a horrible policy though.
I can vacuum my house while on my bluetooth headset and people won't hear it on the other end.
I leave my headset unmuted, and even if my kids are shouting in the next room my coworkers can't hear.
Also, cardiod mics and good software filter out background noise.
Imagine yourself talking to someone that is always looking to the side (even tough, at their point of view, they are looking to you).
Jabra has a similar USB dongle (Jabra Link), but I haven't used it yet.
Keep in mind: Bluetooth audio _sucks_ for calls, it's way lower bitrate than normal music-mode bluetooth audio.
Bluetooth headsets are impressively awful still. Even when they're really good they still fail in surreal and unexpected ways (ignoring all the problems I have pairing them to my laptop). Listening to an audiobook on the beach would regularly have problems talking from the phone in my pocket to the headset if I was laying down on my back or leaning on a wall. I've also recently moved near a railway line and whilst I don't get disturbed by the trains, 8 seconds before they go past I get a blip in my headphones which is screamingly annoying. There are still a stunning amount of upsides for me personally; I'm not sure I'd go back to cheap in-ear wired headphones readily at all.
Ex: http://www.rode.com/wireless/wirelessgo
Also old school wireless phones don't have the latency & low bit rate issues of bluetooth either and they allowed you to plug in a headset too.
It really comes down to the design of the system. I don't know how practical it would be although in a dense city or apartment building.
In an ideal world, voice would be over (real) PTSN which has tremendously low latency compared to anything modern. Cost and convenience make that unlikely though.
Not everyone has a stable desk they can run wired ethernet to, but if you can it makes things better.
> (~$100) Buy open-back headphones, which let you hear your own voice normally and are extremely comfortable.
I think I might stick with what I have.
After a lot of research (I would recommend this blind test as a start: https://www.youtube.com/watch?v=bBRkKAelaKQ) I decided for the Calisto P7200. Its microphone quality seems to be comparable to a headset, and it will automatically turn off when there are no human voices, so nobody will hear you typing.
> Four microphone directional array technology provides superior echo cancellation and full duplex audio
One of the reviewers on Amazon says this as well [1].
0: https://www.poly.com/content/dam/plantronics/documents-and-g...
1: https://www.amazon.com/gp/customer-reviews/R1NFJRGUUYWF4F/re...
I haven't tried this specific one but from my point of view the odds of a microphone picking up your voice clearly from 2ft away, while happily filtering out a very loud sound being emitted 1cm away at the same time, are quite low. Much lower than the odds of a marketing drone getting away with lying on a product brochure.
I'm not an audio expert, but I doubt the SNR is there for the noise cancelling to be effective when you have to substract a 10db echo to recover a 1db voice.
On the other hand, Katie from the Amazon review seems knowledgeable and she has past reviews of other speakerphones so it might be worth a shot :) Still, I'll believe it when I hear it!
- I configured the 7200 as main speaker and played a youtube video in which someone was contineously talking.
- I measured the distance to the 7200, which is next to my keyboard, and it was almost 2ft away from my mouth (standing desk).
- Then I opened Quicktime to make an audio recording using the 7200 as input device, talking while the youtube video was running.
The result: I could only hear myself on the recording. Even while I was talking, or making other sounds, I didn't hear the video at all.
So yes, I think it's full duplex. And no echos.
You would think so, but subtracting the (speaker-generated, i.e., known-signal!) sound from 1cm away is trivial. The hard part is subtracting the echoes off your walls, laptop, desk decorations, etc. etc. etc.
That's not 10db, but it's still on par with your voice!
You are not wrong at a high level -- speakerphones are terrible -- but it's not because the speaker is right next to the mic!
The only exception I can think of is if your environment is exceptionally quiet -- It's probably not -- or if your mic/computer have better audio filtering then most.
A USB speakerphone is designed for multiple speakers and a wider area in which they could be positioned.
You also get some advantages from isolating the mic/speaker from the laptop chassis, noise control and such. Same reasoning as using an external DAC/Amp for headphones rather than the onboard interfaces.
People just don't realize how unacceptable speakers are. Tools should just have a prompt that says "hey, you are ruining everyone else's life by using speakers".
This makes your whole system half-duplex (if someone else talks, they can't hear you). Have you tried using Zoom's "original audio" mode?
Here's an excerpt from an article I'm writing:
----
Computers, sadly, are way dumber than you. When the microphone on your computer captures the sound coming out of your computer's own speakers, it can't really tell that it's not you talking. It just hears a sound. The obvious consequence of this confusion is an echo - you can hear yourself. Every single time you've heard your own voice repeated back to you on a conference call, it was caused by someone else on the call using speakers.
The sound engineers at Zoom, Google, Skype, Jitsi are very clever though. They noticed that they could fix that problem by writing software to do echo cancelation - clever algorithms that detect an echo and then remove it from the audio signal. This software is the reason why most times when you hear an echo on a conference call in 2020, it usually goes away on its own after a few seconds (Note: this doesn't prevent people on the call from thinking that whatever they did to try to fix it actually worked). Echo cancellation is one of many forms of what's called Digital Sound Processing, which we'll talk more about in a minute.
----
What specific microphone/speakers are you using?
yes I'm aware that some software does echo cancellation, as you say it doesn't need to be full mute while you speak, it "just" needs to cancel the signal that would cause feedback...
I started using the dedicated mic because I got complaints that my audio was pretty quiet, due to the fact (I thought) that I 1. keep my laptop farther than you usually would, and 2. sometimes I turn to my second monitor to refer to some info; mic seems to have fixed the issue but I hope echo cancelling does the rest of the job without me needing headphones :D
But lack of an echo is not proof that the setup works, because the problem with echo cancellation is that it is too aggressive, not too subtle. There is no echo when you use the shitty ambient microphone that is 2cm away from your laptops builtin speakers.
But in removing that echo, the software is also removing your voice from the call signal, or other people's voices from your speakers' signal. The symptoms are more subtle, but you and others will miss parts of the call when you speak over each other (or in rapid succession). This creates a lot of "huh? can you hear me?" or weird interruption timings.
Try having someone on the other side of a Zoom call that wears a proper headset, then both of you can turn on "original audio" mode on Zoom. You can actually speak over each other like you would be able to do in person, and back and forth is much more natural, especially if there is more than 2 people having a discussion.
PS: Keep in mind I do 10-12 hours of calls a day, and a lot of them are sales and/or management which requires a lot of active careful listening. My standards can be slightly unreasonable for a more normal use case (say, if you are a developer and do two meetings a day where people take turns to talk).
<this is what you sound like when you do X>
* Why can't I quickly and immediately flag my intent to speak? Once you introduce lag into a conversation (and _every_ video chat conversation includes lag), you can't get away with "let's just be unmuted and hope for the best".
* Why can't everyone in a chat see participants in the same order? How can we truly feel like we are speaking in a virtual space together when we do not share the same view? This one is the most frustrating since it has the greatest impact compared to how easy it is to implement. It is hard to overstate the benefit of having a consistent shared experience.
* Why can't I tell at a glance whether I'm muted or not? Why do I have to find the mute button every time? Everyone has a keyboard, why are we avoiding it like the plague?
* Choosing the correct input/output sources is one of the most common source of issues. And yet, we continue to hide it behind confusing settings. I've only noticed Discord get the memo and let you adjust sources straight from the main screen.
* Why do people have to appear in a random order?
* Why can't I tell at a glance who is speaking?
* Why can I only see 5 people if someone is sharing their screen?
* When someone is sharing their screen, why do I only get to see either their face or the screen, not both?
* Why can't I tell whether people are present/engaged when their video is off?
* Why can people be moved into breakout rooms only through a tedious and complex procedure, and why can it only be done by the host?
* Why can't I emoji-react to what someone is saying?
* Why can't I quickly answer a poll or join a team?
How would you like this to work? Sometimes I prefer others to think I'm engaged when I'm not
The service we use shows people in order of appearance, except you yourself (who is always prepended to the list).
>* Why can't I tell whether people are present/engaged when their video is off?
Why should you be able to tell? Quite often, meetings are a waste of time, and even in the ones which aren't, only a small amount of information will be useful to each person. It's easy to read HN and pay attention for your name or the name of the team you're part of while on a call. Besides that, what counts as 'present/engaged'? I think everyone would be present (in case something be asked of them). Does 'engaged' mean they need to sit there and nod at everything even if it's irrelevant to them?
This reminds me of my two huge pet peeves:
* Why am I forced to see people? If I'm deep into helping someone else out, I want to make their screen as big as possible, I'm not looking at their face at all.
* Zoom spends some effort making my lips and voice sync up. Why don't you make my clicks sync up as well when I'm sharing?
It also has the space bar to temp unmute, basically a tap to speak mode.
I think the UI route for Zoom is wrong though as most users are just lazy and can't be bothered with even minimal effort. So, it should play a loud-ass sound when you speak on mute.
Or just electrocute, to solve humanities problems long term.
Mute is safe, unmute is not; interfaces need to be loud if I'm transmitting audio, rather than the other way around.
Would be nice if they just drew a giant red X when you’re muted.
My gripe is multiple screens while presenting. Give me the choice to show whichever screen my mouse is on.
teams and webex have indicators when someone is speaking
screw muting interfaces though. that shouldn't be something that fades out of the screen.
They tend to optimize for the lowest common denominator. Despite having a nice headworn mic and earbuds, great webcam and good lighting setup, Zoom and other apps tend to optimize for using as little bandwidth as possible so my audio doesn't sound nearly as good as it could.
Ben makes some good points that I didn't mention or consider myself: - I wasn't aware of Zoom's "use original audio" feature so I'll need to give that a try. - Dedicated monitor is great; having a presentation monitor allows you to keep your shared screen clean but still have notes and other stuff available. - Open back headphones to hear your own voice: good idea if you have a dedicated space to yourself, but not if you share the space with anyone else
The biggest problem with this is other people: you won't get any benefits from others if they have a bad setup. So it requires everyone to have a reasonable setup (even just a headset makes a huge difference).
[1] https://jonpurdy.com/2020/03/how-to-improve-your-zoomskype-t...
I have a dual monitor setup but what I find works even better for notes and other shared docs a lot of the time is working on a laptop alongside my desktop computer. Admittedly I'm in a dedicated office so I just need to turn a bit to type on my laptop.
One of the advantages is that I have a very clicky mechanical keyboard on my desktop so typing while on a call is noisy if I'm also speaking so can't be muted.
I don't use a headset mic. My experience is that those rarely sound good and tend to end up with even more breathing and mouth noises and stuff than even a computer's built in mic.
I have a Blue Yeti that work got me. I've asked other people on video chats and they say the mic is a big improvement over my laptop's. I think that particular mic is a little bulky and ostentatious but it seems to get the job done.
You have to do a modicum of adjustment to make sure you are not breathing on the actual microphone, or chewing on stuff while you are not muted. I shift to my ambient microphone if I'm eating on a call.
Unless you are in a studio-grade environment (including not typing on your computer) or doing some serious voice recording, I have found boom microphones to be the better tradeoff for actually talking to others.
People often draw the wrong conclusions from posts like this. I've worked with people who've taken offense if anyone's video conferencing gear wasn't up to their bar, and it felt like a meta-topic that was never worth the time.
The most notable has been a healthcare provider I’ve been working with. For a multitude of reasons not related to COVID we’ve been relying on telehealth. The sessions are predominately about the conversation and it’s downright frustrating that the provider can’t even be bothered to use the earbuds with mic that come with an iPhone. Our audio is constantly stepping over each other. I’m also looking up the provider’s shadowed nose the entire time too.
The answer is somewhere in the middle. I don’t expect the provider to spend nearly a grand as this persons blog post advocates. However, some super basic training about audio, camera positioning, and lighting would make a world of difference.
Never in in any group call, no matter what your microphone configuration is.
But for group calls, no matter what, please mute!
I'm most interested in tips on tools or devices that allow you get higher quality audio calls while taking walks.
https://www.nvidia.com/en-us/geforce/guides/nvidia-rtx-voice...
Nvidia used to restrict installs to RTX cards but people worked out how to remove the restriction. Recent RTXVoice versions removed this restriction so now you don't need to modify ini files to install it.
I've been using it myself and it improved the sound quality quite a bit. Or rather, it eliminated a lot of the noise that was coming through.
No, please, do mute. I don't want to hear a baby crying in the background while the SO tries to calm them down, nor do I want to hear someone doing the dishes, or the police siren that seems to have stopped in a congested area. We have work to do. Please, let's get it done and stop the meeting.
Also, thank you Google for enabling people to mute other participants who do not realize how loud their environment is.
This article seems to come from a place where people live in big, spacious places (possibly rural North America? I could see Canada). The reality for me is that most of my coworkers work from a 1- or 2-bedroom studios crammed with kids and/or their significant others.
A simple noise canceling microphone takes care of distant, and thus unimportant sounds.
EDIT: Please take note of the "in low participant count videos" caveat above before lambasting me about how this doesn't work in 40 person meetings. I'm talking about meetings with 3-4 people.
Nothing annoys me more than people having background noise and not muting themselves, no one should be pushing crying babies and barking dogs to my headphones without a reason it completely ruins the call for me. If I really need to know if they are distracted, I can just look at their video.
It’s simply rude to broadcast your noise to the entire group of people, let alone dozens of people doing so at the same time. If your child is crying, it’s fine, I understand how things are, and of course I’ll understand you’re distracted. However, I’m not fine with your child distracting all 20 people on the call.
For a 1-on-1, you absolutely want to know if the other person on the call is going to be distracted, or you're wasting your time.
For a 40 person all hands, the meeting could probably never proceed if it waited for everyone to be completely distraction free.
If you have a baby crying or police sirens, then yes, for the sake of everybody's sanity, mute yourself. For sure.
I.e. don't work in the middle of a global pandemic.
A headset mic will pick up your loud typing, sniffling, clearing your throat, and whatever. Even if you are working out of a sound-proofed office, and in a group call, mute your mic when not speaking.
Not necessarily. A cheap one will, since it acts as an omnidirectional microphone. A good one, however, will have a hypercardoid microphone and an externally directed second omnidirectional microphone hooked up to a noise canceling circuit that renders external sounds effectively inaudible.
I can type on my keyboard with my Antlion mic on, and it's not audible to others (even though it's fairly loud to my ears). And yes, I tested this with a local recording first.
That said, I agree that coughing, clearing your throat, and such are definitely mute moments.
Then one day I plugged it in to the port on my laptop's dock, instead of the plug on the laptop itself. In the dock's port it was suddenly so sensitive that it was picking up the slightest sound out of the headphones (and immediately feeding back)! And a conversation going on in the neighboring conference room, that I couldn't even hear on my own! Unless you really focused on something it was just insane loud background noise. But, it was fine once I turned the mic volume down to like 2%. But also, my words could be made out even with the insane sensitivity--I guess that's pretty impressive dynamic range.
So:
1. If you're picking up things on the mic you think you shouldn't, try turning down the mic's volume, it might be that your voice stays loud enough, but everything else drops off.
2. If you're getting background noise now, it might not be your new setting, it might be your computer hardware. If you're using different computing hardware maybe you need to fuss with the volume even though it's the same headset as before.
A poor-quality headset mic and sub-standard audio software might. People dialing in on a phone should always be muted at all times, even when they have something to say, because phone quality with Zoom isn't good. But on my laptop, with my headphones, I can even chew and no sound is transmitted at all. I can see this on my real-time level monitor, and have confirmed it with others. I don't eat during meetings, but I could.
Typing and sniffing sounds like people using a built-in laptop mic as a speakerphone, not a headset.
In a conference with 10 people, 7 will be on Windows PC's with on-board RealTek, where the only mic control in the control panel is a basic gain and a mute checkbox, or else with some cheap USB headsets, 2 will be on the road with their phones, and 1 will be on a Mac.
You will never get a conference with everyone having True Scotsman hardware and drivers, and so the general rule has to be for everyone to mute.
Look, I'm all in favor of people muting by default, especially in larger meetings. But when I chat with my team, 2-5 people usually, the only person who mutes himself usually is the one Linux user.
It won't if you spend 2 minutes setting your audio thresholds.
but if you can't isolate yourself from background noise, then there's no point in getting fancy equipment for the rest of the stuff. just put yourself on mute and use your laptop's built in mic and webcam, because nothing you try will fix the background noise.
Like, I don't have a baby crying in my house because I don't have a baby. Or pets. If you do, stay muted. Where I live my environment is very, very quiet. Which the article says is step 1. If you can't do step 1, then yeah, this article isn't for you.
Which I sort of agree with in a small call where people are having back and forth conversation with each other. If I'm in a 1:1 I don't mute.
But it's still a good idea on larger calls or even a smaller one where you're just going to be listening to a presentation. In general, there's not a lot of cost associated with going on mute.
2. This hits the nail on the head with the duplex audio problem and the only video service that gets it right is.. actually Discord who started as audio only originally.
Sorry, as someone who is completely out of the loop, can you explain what you mean?
(emphasis added)
https://www.latimes.com/business/technology/story/2020-06-10...
1) Zoom has been reported to send video call encryption keys to China.
2) Zoom has sent participant lists of particular video calls to China authorities.
3) Zoom now segregates Chinese from the rest of the world, so that the keys and participant lists can be disclosed to Chinese censors, without impacting the rest of the world.
edit: and not just China. What makes you thing Google wouldn't do this in the US?
There is no reason sufficient to justify such behavior.
edit:
As to Google in the US, there is no legal framework requiring them to obey any administration and in fact significant protection if they would like to make fun of or openly attack it. There is no comparison here that isn't in utter bad faith.
Horrible - to use your word - is horrible, no matter how many people are doing it.
Anyone used a Facebook Portal? How is it? Could it be hacked for this?
By default, it's not always easy to figure this out.
I ended up adjusting the image manually with v4l2-ctl (see for example http://manpages.ubuntu.com/manpages/xenial/man1/v4l2-ctl.1.h...) I also pointed my warm white light desk lamp towards the wall behind my desk and bounce a bit of light back on my face off the wall.
If there is a better thirty-party program out there for macOS, I would love that.
The Logitech C920 is IMO a nice compromise. It's as high quality as you need for most purposes and it doesn't take up the desk real estate of a dedicated camera. It's also easier to have your eyes closer to the slides you're presenting or the video of the person you're speaking with.
Nothing wrong with the advice here, but done perfectly it will take your meetings from a 3/10 to a 6. I was hoping for tips on improving communication.
Figure out the little facial gestures, the verbal indications and such that allow you to get some of the richness back. Obviously, you won't get to as good as offline, but if everyone puts in the effort, you start to understand the limits of what can be sent and received online to each person.
In person is better of course but a low latency, high quality connection goes a LONG way and it's rare enough that most people have never experienced it.
Sound heard in two ears is perceived by your brain as louder than that heard in only one ear. A listener using only one ear must raise the volume, and the harm of that over the long term (and even individual event exposures) is well studied.
Anecdotally I have several people in my family that used to work in radio, spend extended amounts of time using monitors in only one ear, and now have advancing or profound hearing loss in that ear.
Also, where are you pulling your notion of 8Khz capped audio quality from? Skype has historically had some issues with selection of the crappiest audio codec and sample rate that resulted in pushing many calls down to the lowest bar of 8Khz, but info I can find for Teams suggests a sample rate of 16Khz and bit rate of ~50kbps. For Zoom it looks to be 32Kbps ~50kbps. By no means approaching the capabilities of a decent audio output setup, but better than you reference.
I just use one AirPod at a time, so I hear normally through the other ear.
I look forward to the day when high quality doesn’t also mean “private conversation is impossible”.
Hand the photo to a friend along with a sharpie. Ask them to write a short message to you on the photo’s surface.
Scan the result, then write some code to produce an image of the message in black on a white background. Well done, it’s probably very close!
Now repeat the exercise using a blank sheet of white paper instead of a printed photo. This is your call with headphones, vs a call on speakerphone!
(Bonus points if you chose a photo of some sharpie writing on a white surface for the first part of this exercise. This is you on speakerphone in a public space filled with people using speakerphones to also make calls.)
$800 dollars for what? "Immersive video call"?
And this one made me laugh:
> Prefer Zoom to most alternatives