Hacker News Reader

Comment by phailhaus | Hacker News Reader

We are still missing basic primitives in video conferencing interfaces, and this failure of abstraction surfaces as "virtual meetings feel weird".

* We still don't have a button to signal that we intend to speak; we don't need a fancy "pressure sensitive chair" to accomplish this.

* Video conferencing apps still refuse to show everyone the same layout of faces. Because we don't share the same virtual space, of course our experiences feel disjointed.

* We can't silently "react" to someone else speaking, making it difficult to intuit the consensus opinion. These reactions need to appear on _their_ tile, not on your own (looking at you, Zoom).

These are not hard to implement, but conferencing solutions are so sticky that they just don't seem to have the incentive to improve their stagnant UI's.

View on HN

I think a big one is actually latency and framerate. If you have a low latency videosetup (e.g. SDI Cameras and two monitors) you can try out what real low latency feels like. This works soo much better already because you can react to one another's mimic expression etc.

Unfortunately, that requires the end user to update their hardware. We have to live with the reality that people are not going to have low-latency setups, and design with that in mind. Everybody wins!

>We have to live with the reality that people are not going to have low-latency setups, and design with that in mind.

Or we can live with the reality that some people will shell out for low-latency setups. There is room for both.

If your video conferencing platform _requires_ all participants to have low-latency setups, you are going for a very niche market. I'm talking about fundamental design changes we can bake into the general-purpose video conferencing apps that exist today, such as Teams, Hangouts/Meet, and Zoom.

latency in audio is definitely a big deal. I've definitely had the experience many times of starting to talk a second or two after someone else finishes, only to find someone else had started talking - and all this takes is maybe 500ms of delay to become super noticable.

I've also found it extremely noticable when someone is screen sharing, and their screen share lags behind their audio several seconds often. If they are going through an interface fast, it can be hard to tell what they are referencing on their screen.

I frequently dial into zoom calls via phone, which seems to improve the latency issues a bit. It also helps with stuttering audio in general. It sometimes results in AV sync issues, but I’d rather have smooth sound than jagged sound connected to jagged video.

I dial in all the time, whenever possible. A client's corporate Webex will call me when the meeting starts - that's been great. Another client's zoom will send an invite with a clickable phone number but it doesn't have the appended , and passcode on it, so I have to look that up and type (before the system thinks I'm not there) - super annoying.

It gives me the freedom to pace/walk/move without being tethered to the laptop/desk/screen. Now... I often connect in there as well, so I can see any shared screen, but if I need to walk away for a bit (coffee/etc) no one knows, but I can keep listening. Muting myself on phone is easy too.

And yes, I've found it helps with some audio stuff too - often I'll hear other people making comments about losing audio, or delays, but I rarely experience those when dialed in by phone.

Indeed, even in VR this is a problem as demonstrated by this informal talk with John Carmack:

https://youtu.be/P7mFEGah27c?t=1359

It's not just primitives, but more generally power features are lacking in all packages I've used. For instance, the most basic tool used in audio production is a vu-meter. On my Linux workstation, I keep the Gnome sound settings window open to check the gain.

I understand you want the experience to be as automated as possible, but in large meetings it would be extremely useful for at least one admin type participant to have access to tools to diagnose sound issues. Same for layout, I don't understand why there's no way to force a particular layout. In Webex, you can't seem to be able to view chat and video at the same time, which is idiotic. Just sharing an URL is cumbersome.

> * We still don't have a button to signal that we intend to speak; we don't need a fancy "pressure sensitive chair" to accomplish this.

Jitsi Meet has this: https://meet.jit.si/

You can also use the hand you are born with. It will be even more visible. It didn't feel that awkward in school, so we can use it now too.

You can't in Google meet when there are more than a couple people, because meet refuses to show all participants.

It now supports up to 49 people: https://gsuiteupdates.googleblog.com/2020/09/49-people-tile-...

This depends heavily on what layout everyone is using. We usually have a screenshare taking 90% of the space and if we have cameras on then most video conference software seems to show random 3 people in a small tile somewhere while the rest are just hidden.

Agreed, but the thing that makes it "weird" for me is that anytime people stop talking in a zoom meeting it becomes dead air that people feel compelled to fill up.

Natural breaks in the conversation don't work virtually right now.

Dead air is a cultural issue. Some cultures (Americans) can't handle silence and perceive it as awkward.

In my mind that means the agenda has been covered and it is time to wrap the meeting up.

> We still don't have a button to signal that we intend to speak

MS Teams has a "raise hand" feature, is that the type of thing you're thinking about?

https://support.microsoft.com/en-us/office/raise-your-hand-i...

Discord lights up the avatar's outline when they speak, which is fantastic by itself, but combined with push to talk it's perfect as you can convey the exact intention of taking over the conversation and solve it much more quickly when there's overlap than any 'raise hand' mechanism I've used. Those tend to just not work as raised hands get more ignored than just talking.

Google Meet also has this too.

Zoom has it too - we use it regularly.

Chime does too. And it feels very clumsy.

It's...something. But it still doesn't feel quite right. A formal "raise hand" button feels too clumsy of a replacement for the natural body language we use to imply that we want to jump in to say something. I don't want a button that "requests the floor", we need something more fluid than that.

Basically the situation we need to avoid is the example raised in the article, wherein two people stumble around each other trying to figure out who should speak and who should back off. The reason that happens is that both participants have no clue that the other person is about to speak until they already have.

Discord solves this beautifully with push-to-talk. Since I have to press the push to talk button _before_ I speak, the UI can hook into that to naturally indicate that I intend to speak just by lighting myself up. I can just hold down the button and wait for the current speaker to cede, or jump in myself at a suitable pause. This latter case is now possible since everyone in the chat can see that I am about to do that, rather than be caught by surprise. But more importantly, _I can back off myself_ just by letting go. In Teams, you'd have to "rescind" the raised hand explicitly (not even sure if this is possible).

Now combine this with a shared virtual space: everyone sees the same order of faces, and when someone pushes-to-talk, they move up. Everyone sees this at the same time, so it's much more natural for that person to start speaking!

As a counterpoint, I feel that a lot of in-person meetings would be improved by having people discretely request the floor rather than relying on having your body language acknowledged. Certainly it would help for those of us to whom aggressive/imposing body language doesn't come as easily.

That's just reality: some people are not attentive and will steamroll others. It's up to us to jump in and help people get heard. If we start running into the same problems as real life, then we're going in the right direction.

I think a "request the floor" button still has its place in much larger group events (imagine a seminar), but for small to medium meetings it's pretty overkill. The goal should be to lower the bar for fluid conversation, rather than to formalize it into a turn-taking process.

> * We still don't have a button to signal that we intend to speak

There is a raise hand button in teams. But as all our meetings are small enough, we never used it, so I don't know how useful it is

Speaking as a masters student who's had a lot of seminars through teams it is a fairly good solution. However, it requires effective moderation - the situation is still lacking in social cues.

> There is a raise hand button in teams. But as all our meetings are small enough, we never used it, so I don't know how useful it is

Not all that useful really. The icon is small and gets easily missed. Also, when you're presenting, you can't see it at all which is mind-blowingly stupid to me. If there was ever a time I want to know if somebody has a question, it's when I'm in lecture mode and can't see people's faces.

Just raise your hand on the web cam. Move it closer to make it visible.

Unfortunately that doesn't scale, and requires everyone to have their video camera on.

> that doesn't scale

Why not?

Can't fit arbitrarily large numbers of people on a laptop screen, it's way harder to pick up this kind of motion from a mosaic of small 2d tiles, and everyone needs their camera on which isn't always feasible (connection issues, privacy of others in a home office setting, etc.)

Good luck noticing that in the grid of 16 people on your tiny laptop screen.

Did your in-person meetings ever had this many participants? If this many people needs to be involved, then I'll argue a meeting is a nonsense anyway. It would be extremely hard even in-person.

You are trying to replace in-person meetings without using cameras? It scales pretty well to a reasonable amount of people. How many people do you need in a meeting? If you have meetings with more than 12 people, then you are probably holding a conference and your meetings suck anyway. A written announcement would work better.

You are trying to bend reality to your will, it doesn't work like that. The fact of the matter is that sometimes you have large meetings, and often times participants will have their cameras turned off. Instead of fighting it and saying that people are Doing It Wrong, we need to accept this reality and work with it.

A push-to-talk system with shared layouts solves all these problems, is scalable, and doesn't require cameras.

For your first point, how is that different from the "raise hand" button?

As I outlined elsewhere in the thread, a "raise hand" button is way too bulky since it's essentially a "request the floor" button. When you're talking naturally with a group of people, you don't submit formal requests for the floor; you simply lean in and everyone subconsciously acknowledges the fact that you intend to say something. The speaker can naturally cede, or you can jump in at a suitable pause. Or you can back off entirely and that's fine too!

The simplest way to accomplish this is through push-to-talk. Since you have to push a button before you can speak, the UI can highlight your speaker tile and allow others to acknowledge your intent. It's much more fluid and natural than formally "raising your hand" and then having to rescind it if you decide to back off. The point is to provide low-level primitives that can become almost reflexive parts of video chats.

But zoom has push to talk and indications that you are unmuted, so I still don't quite get what you are talking about. That's ok.

Push-to-talk is necessary but not sufficient: you need to actually light up the square and move that speaker up so everyone notices that they intend to speak. Not sure if Zoom does this. Hiding a little "mute" icon in the corner is not good enough, we need to bake this functionality deep into the UI so that it becomes a natural and reflexive part of meetings. Zoom doesn't share layouts across all participants, which is a big problem too.

"Raise hand" is an attempt to solve the right problem, but it doesn't solve it in a realistic way. Virtual meetings feel weird because we're all more Trumpy than we realize - in actuality nearly every speaker's turn ends with an impatient interruption by the next speaker. This just doesn't feel "natural" with the latency of a phone call or video conference. I know, it's cynical, but it's one reason we love to hate meetings - we never get to finish tal

You can "raise your hand" in MS Teams as a signal to everyone else

Reply on news.ycombinator.com