How Does FaceTime Work?
matduggan.com
matduggan.com
We found a free app Sonobus that works really well for low latency audio for what we were doing in the end. FaceTime is great for just calling the family on the iPad though.
We were so young and naive then.
People have guess that patents prevented it from actually being an open standard. Coincidentally, some of those patents expired in the last year or two and the last should expire this year. But I kind of guess the window has passed for it becoming open.
It’s moving there. They’re adding a web interface this year.
I don't see a web client being a move in that direction.
It enables FaceTime users to stay in the ecosystem when their friends don't use the platform they do (vs going to Zoom, which is available for everyone). And it enables non-users to get a taste of FaceTime without before they buy an iPhone.
It's a brilliant move, but I wouldn't call it open. Maybe "open enough" :D
Not guessed. Apple was actually sued and had to change how FaceTime works.
As a blind person with a lot of friends in different corners of the world, audio quality is very important to me. Surprisingly enough, I've seen side projects that took a weekend to develop that had much better quality than what mainstream services offer. Part of it is probably because of bandwidth costs, but I guess effects (like cancelling echos from participants who use speakers instead of headphones, or reducing noise from crappy mics) also play a role.
Facetime's quality is good, but nowhere near what your devices are actually capable of. The only mainstream solution that is actually good is Zoom, when you enable original sound, stereo audio, high fidelity mode and disable a few annoying filters. To do this, the app needs much more fine-grained control of your microphone than you can get from a web browser, so the native client is essential.
Discord with Nitro is pretty decent too, but really niche, obscure, non-mainstream solutions work best. TeamTalk[1] is one great example.
Oddly, I also landed on zoom has above the rest in terms of quality. Everyone on my team uses headsets, this avoids need for echo cancellation and the weird muting effectings that generates - so I HIGHLY recommend this - two people can talk at once comfortably so you pick up the little cues of interjections etc.
Facetime audio only is good in quality in my experience, but not so great in latency (may be for reasons outside apple's control obviously).
I wonder if there is a market for a meeting app that has the best quality of them all, by not trying to poorly work around bad environments, but simply booting people out of meetings until they fix their setup.
Yes, it would suck if you or someone you really need something from get booted, but after a few weeks everyone's setup would likely be fixed so this would no longer be a frequent issue, and in exchange, you wouldn't have 20+ people get driven mad by someone giving a hour-long presentation over Bluetooth audio.
I feel like shitty audio quality is responsible for at least 80% of the 'zoom brain' effect.
There are so many situational things involved when making a call that I would estimate the size of this market to be approximately $0/yr.
P.S. I'm sure this comment will come back to bite me 10 years down the line just like that Dropbox comment.
However, I also have an air conditioner. You can hear it. I try to keep it turned off during meetings, but sometimes it's just too hot. and so Original Sound goes off and the air conditioner goes on.
What else am I supposed to do? Leave the meeting?
Have all the background noise you need to have, as long as I only ever have to hear it when you're actively trying to speak. If I still can't hear you over your noise, then you have a bigger problem.
If I have to listen to your noise even when someone else is speaking, then we all have a problem, which PTT nicely solves.
Anyone who spend the 90s / 00s competitive gaming likely already has their preferred PTT key, that they can push while continuing to still be productive in whatever activity they're already engaging in.
Some tweaking may be required, given the different input methods used depending on the job vs the game, but that was true of different games back then too.
Your solution is to setup a streaming rig in everyone’s house, and force them into a home cubicle? Sounds absolutely painful to me.
I suspect the lousy audio quality in modern voice conferencing software is to compensate for people with lousy internet connections or underpowered devices: someone using an older phone in the car on 3G cell service with cell tower handoffs, or a crusty laptop using lousy coffee shop WiFi, etc.
It's a tough problem to solve. Everyone on a call can be affected by one person with a lousy connection if the software isn't very sophisticated.
People are generally willing to tolerate lower bitrate audio if it means less dropouts, and lowering the bitrate for everyone is easy to do (if a bit lazy), and solves many problems.
It’s similar story with CPUs. Modern CPUs are insanely fast these days, smart phones included. What once required purposedly-designed ASICs (e.g. Sony Minidisk had one for ATRAC3) is now borderline free in terms of compute power. A cheap Android phone sold 10 years ago gonna have at least 500MHz CPU, and at least 256MB RAM.
96 Kbps AAC audio works on 10-15MHz CPUs, and 64 Kbps HE-AAC works on 50-60 MHz CPUs, both from 2005: https://www.helixcommunity.org/projects/datatype/2005/aacfix...
If we're talking about downloading files (and loading websites), you're mostly correct. However, temporary slowdowns, disruptions in wireless connectivity and wireless-specific packet loss affect audio quality a lot, while not having much impact on your download speed. Even when streaming audio or video, you don't feel them as much, as your device usually keeps at least 10 seconds of buffer that it can rely on. When in a call, though, 10 seconds of buffer would mean 10 seconds of latency, which is simply unacceptable.
You can have a 1.2 mb/s connection that sometimes goes down to 64kb/s for two to three seconds. Even if your download speed is still 1 mb/s, you can't really receive more than 64 kb/s when in a call, or you will get dropouts.
I don't understand - are landline phones not in use anymore? I only use a landline phone to communicate with the family, because I simply cannot stand the audio quality of cellphones for calls which last longer than a few minutes.
Edit: In Germany, there are about 40 M active landlines. That is 1 landline phone for every 2 inhabitants, and this number seems to be have been fairly constant over the last 20 years [0].
[0] https://de.statista.com/statistik/daten/studie/274339/umfrag...
I think different countries have different norms around things like landlines - or even texting. In the US, people ignored texting for a very long time and preferred to call people even if it was just a short note.
In terms of audio quality, VoLTE has improved that quite a bit when both sides will inter-operate with HD voice. Older stuff often used 8-12kbps and primitive codecs to encode your voice which simply a lot of compression. I know Speex has had some great results at low bitrates, but that's a far newer codec. However, I believe VoLTE's AMR-WB codec still chops everything about 7,000 Hz. T-Mobile US and Verizon have both implemented EVS which goes up to 14,000 Hz.
https://www.pcmag.com/how-to/how-to-make-your-cell-phone-cal...
Of course I'd find a Sascha Segan article when looking to confirm some information. It has some great recordings to compare and you can see huge differences depending on carrier implementations. The Sprint-Sprint test shows very poor quality. The T-Mobile 2G test also shows how old networks just used poor codecs with limited bitrates. You can even hear distinct differences between the T-Mobile-T-Mobile call and the Verizon-Verizon call where the Verizon call just sounds significantly less rich.
Back in the day, calls used a huge amount of network capacity. Even today, because calls need a more guaranteed bitrate, networks need to leave some extra room for them. 128kbps of streaming music isn't the same as 128kbps of real-time voice because the network can deliver 15 seconds of music and then it can deliver the 16th second of music any time in the next 15 seconds. Real-time two-way communications don't have that luxury. Still, bumping up the bitrate on communications makes a big difference and networks today should be more than capable of handling it.
I don't know what German operators have implemented, but cell phones have come a decent way in terms of audio quality, though carriers being lazy and interoperability problems can mean that consumers don't always get the benefits in the real world.
The other problem with land lines is the amount of robocalls you get on them. The phone companies offer no way to truly block them. Every blocking service they offer relies on Caller ID, and with the robocallers spoofing Caller ID and randomizing numbers, the blocking is totally ineffective. I literally get 5-8 calls a day. They used to offer a call blocking service, where a machine would answer, and tell you to hang up, or press 1 if you wanted to connect. This service no longer exists, and the other call (ineffective) blocking services (they offer at least 3 or 4) each cost about $6/mo each. My belief is that they are making tons of money allowing unauthenticated calls into their network, that they have no desire to actually block any calls.
Also the relatively small number of landlines left leaves a rather small audience of people to complain. So most people don't even know this is a problem.
I've been waiting for that to happen on cell and landline numbers. Authentication for the party calling. At this point, there's no other option than legislation. Just mandate how common voice services like landline and cellular operate their business.
It needs to happen otherwise the only solution is to give them up and only do our own outbound calling, never to actually receive any calls at all.
Other options like blocking all calls from non-contacts is risky because a call from a delivery driver or babysitter in an emergency will be missed. It's about the only option though for enduser today.
It was more intimate, you could hear breathing for example.
They switched to digital because it reduced their costs, instead of multiple trunk lines you could get away with fewer fibre lines.
You mean they had a cartel for decades, kept prices artificially high, didn’t compete and the government did nothing?
Yeah sounds like the cablecompanies…
I remember that happening in the 90s due to deregulation and breakup of at&t. 10-10-321 had a lot of commercials in the 90s[1]. I'm sure they're on YouTube. Most people I knew would go to a convenience store to buy a pre-paid card for long distance. 1-800-collect was a similar service that had a bunch of ads for calling collect.
It kind of makes calling confusing (and required going to a store and buying something upfront), but I remember younger people (who probably weren't paying for the landline at home) never really considered using their home phone's long distance service. I think most landline services kept their high prices because the few that used it were too lazy or didn't know better. I wouldn't be surprised if it's still expensive (I haven't had a landline in over a decade).
Also, for people getting an internet signal from the phone company (DSL), the landline was usually bundled in a package. They tried to make them viable for as long as possible.
Even the older grandmas who did not go past grade school who do not know a non English non Latin language like to use their WhatsApp to consume media from their cousins and other friends/relatives. And of course, FaceTiming grandkids.
> In the past decade, landline subscriber base has been on a decline, with connections reducing from 36.76 million as of 31 January, 2010 to 20.58 million as of 31 January, 2020. Of the total telecom 1.18 billion connections, only 20.26 million, or less than 2%, were landline as of February. - https://www.livemint.com/industry/telecom/trai-says-declinin...
(And yes, landline phones do offer better voice quality. Even if I am calling from a cellphone, I prefer to call someone on their landline. )
Of course now (for the most part) packet loss concealment and other advanced technologies have equalized a bit of this but there are still days where I miss B-channels on a T1 PRI.
I’ve made some calls, but less than five in the last few years.
This was around 2009 and many users, having not experienced that kind of frequency response/fidelity were kind of "unnerved" by it because (as you say) it can sound "too good". Users were freaked out by the extended frequency response and added fidelity. A common complaint was that previously undetectable breath sounds, etc were audible.
"These days" with FaceTime, WebRTC, etc sampling rates go all the way to 48 kHz, being able to represent the entire range of human hearing (and far beyond the range of human speech[1][2]). Not only have people become more used to the extended audio range, various other parts of the stack (quality microphones/hardware, speech detection, noise cancellation, etc) has caught up and some of those early pesky issues like breath sounds, etc have been largely addressed.
[0] https://en.wikipedia.org/wiki/Nyquist%E2%80%93Shannon_sampli...
[1] https://en.wikipedia.org/wiki/Voice_frequency
[2] https://www.frontiersin.org/articles/10.3389/fpsyg.2014.0058...
It's weird to me how, after a century of obsessive attention to quality and latency all throughout the whole analog era, through the digital isochronous transition, and up into the first-gen analog cellular system, digital cellular was so much cheaper that we just rolled over and accepted all the terrible compromises. It's not like customers who preferred AMPS had the choice to stay on it.
Only recently have I experienced VoIP calls that start to claw back some of that lost quality. Frequency range, bit depth, maybe even latency if I'm lucky. I don't think we'll ever again have it as good as ISDN, but I can dream.
The BEST cell phone call I've ever had was back in 2004 on a sony ericsson in Bangkok of all places calling my parents on their landline in Chicago. Crystal clear, almost zero latency, it felt like I was in their living room talking to them. I've never had that in the states and I would like to.
I too miss the days of a B channel on a T1 PRI!
We're looking for such a service for use in developing countries where iPhones are rare and connectivity is often mediocre at best.
Tests with the like of Telegram, Duo and others have been okay-ish for voice-only calls but pretty bad for video. Mostly because of lack of Apple hardware on the other end we are until now unable to conduct any comparative tests to see whether FaceTime would outperform these apps in terms of quality and stability.
https://techcrunch.com/2021/03/04/whatsapp-rolls-out-voice-a...
Strangely I've never noticed this audio problem during video calls.
So I tend to do voice only calls using WhatsApp and video calls with FaceTime.
For home, you chose fiber if you’re lucky to have fiber. Most people only have access to coaxial cable internet. If you’re unlucky, you do not even have that.
For mobile, you have Verizon, ATT, and T-Mobile so it is not so bad there.
Yes, with 1.5Mbps one should hear crystal clear audio, but you have no idea how much bandwidth is actually being used.
Reminds me of
> The four F's: fighting fleeing feeding and mating
It is a shame that FaceTime isn't easier to use as an alternative to Slack-calls and screen sharing. At least last time I tried, or I'd use it for work as well.
In the 21st century, this is unacceptable. The app is so clunky I loathe using it but it's unfortunately still a good (the best?) solution for dialing landlines and mobile phones abroad for extended periods of time.
[1] https://en.wikipedia.org/wiki/Voice_over_LTE
Edit: mdasen says more about VoLTE below: https://news.ycombinator.com/item?id=28249217
It's also such an equalizer, allowing even the non-computer literate to use it and focus on the actual conversations.
In fact, my 103 year old grandmother is able to do video calls with me, which is great because I rarely see her in person these days. I don't think there exists an alternative she would be able to use.
What Apple does to simplify things (prioritize simplicity and accessibility, obsessively remove or hide options, make reasonable defaults, broaden the happy-path) can be replicated on more affordable platforms.
That said, I've struggled to teach my elderly mother how to even answer her iPhone -- she finds it overly complicated. If she could get over the technical hurdle, she could stay in better touch with remote family and friends who now prefer texting and video calls to snail mail and begin to alleviate her loneliness.
Even in poor areas, lots of people are walking around with iPhones. $200-500 gadgets are not out of the range of poor people. That works out to less than a dollar a day.
One time amortized expenses aren't that much of a barrier, from what I've seen. The real barriers the poor face are things like high rents, high medical cost, high cost of education, etc.
[1] https://www.worldbank.org/en/news/press-release/2018/10/17/n...
(I say cheaper because you don't need to buy an iDevice to use those)
I remember my 6 very well as I fell on my motorcycle and I pulled this banana shaped phone out of my pocket.
I definitely meant the 6.
Minimum monthly wage (formal jobs, before taxes): R$1100
New iPhone 6: R$1700 ~ R$2400
New iPhone 12: R$7700 ~ R$10000
> Does average monthly salary mean much if the majority is really poor?
For those poor and outside the formal job market, buying a new iPhone (even the older ones) are out of reach or a very significant investment.
Cheapest iPhone: 120000 HUF
Cheapest Android: 9700 HUF
Cheapest reasonable Android: 21000 HUF
Minimum wage after taxes: 111000 HUF
Whatsapp ,Skype, Telegram, Google Duo, etc apps work fine on Android phones which are lots less expensive than Apple Devices that are needed for Facetime
Firstly; she has not heard of WhatsApp, it would have to be taught to her that it is a chat app that has additional functionality to allow phone and video calls.
Blank stare.
She does not chat and has no interest in doing that.
The UI of WhatsApp is confusing and unintuitive if you assume you know nearly nothing except that you want to make a video call.
> Firstly; she has not heard of WhatsApp
I am pretty sure she had not heard of FaceTime as well before you introduced it to her? :)
Disc: Googler.
Not signing up to anything Facebook is the first criteria in any service I want to use.
FaceTime is limited to Apple devices until iOS 15 allows anyone with a browser to join a call, so you’d think they’d be emphasizing the benefits of portability more.
https://www.theverge.com/2021/6/7/22522889/apple-facetime-an...
It was originally announced as an open standard, though. Anyone who remembers Steve Jobs would have caught on. It is easy for us old-timers to forget how long ago that was now, so I extend my apologies to the youngsters.
Older ladies might also have long nails which makes touching icons hard too.
In the end i guess it all depends on the persons ability to manipulate and see things.
I've pushed my parents hard into the Apple ecosystem and it has worked out great.
My mom first got the iPhone 4, before that she had a nokia flip phone, she could use it from day 1. Nowadays I buy the newest iPhone, and after 1-2 years or so I sell it to one of my parents for a 'nice' price. The iPad I made them buy (and they upgraded it once) is very popular.
But the most used Apple product is the AppleTV. It has enabled them so watch the news when they want, and Youtube has given them access to content (travel) they'd never had before.
Sharing photo albums, facetime video calls, imessage and ease of use is just on a level I don't think exists with android - of course, sometimes theres the small hiccup.
But they are not 93 years old so the mileage might vary :)
To her, it's almost like a television, as family members are able to call her and even initiate calls. (Although we often have to call her on the normal phone to get her to realize that we're calling her on her computer, or to plug it in --- this is the one thing she does do with the hardware.)
The key point is the single function: a general purpose computing device has more UI & management than some people want to deal with. This is something the industry really hasn’t done a great job with, especially since it overlaps with the problem of treating things as disposable because the revenue model for single purpose devices is mostly broken and so people are often forced to churn them because the device they’re used to has been discontinued or no longer works with their WiFi.
My grandmother as mentioned uses an iPad which is the only user friendly tablet or "computer" I would ever give her.
She uses it for general browsing and e-mail as well, as well as taking photos and browsing albums in iCloud. No messaging though.
I would also recommend a handle/knob type thing on the back so it is easier to hold it without accidentally touching the screen.
I'd legit pay money for a Consumer Reports around Google Meet, Jitsi, Zoom, and FaceTime. Load up a LAN with a packet loss/jitter simulator and gimme some charts.