It probably saves insane amounts of bandwidth. But at what cost :(
I dunno, I’ve been in a few meetings where people with deep pockets make critical infrastructure decisions based on extremely limited information. Trusting “them” to have a valid metric is a rookie mistake.
The older you get you realize more and more that the people making the decisions are totally clueless.
…Until you become one of the decision makers.
Yuuuuuuup.
“How much will option A cost? Dunno.”
“What about option B? Dunno.”
“My gut tells me B is much more expensive than A though.” “Yeah for sure. But I prefer B.”
Wanna waste a hundred grand a year? Go right ahead, who cares. Wanna hire someone? Woah hold your horses there bucko, don’t you know we have budget limitations??
Curious as to why that would be the case. Unless people are starting videos, letting them buffer, then reloading and doing it again.
It should be the same amount of bandwidth, otherwise, right?
It’s just people not finishing videos. Buffered but never played. In aggregate it adds up to a lot.
People opening 8 hour long music videos, then pausing them after half an hour and just keeping it open while they do something else.
Then they come back after multiple hours and just close the browser.
Letting the entire video buffer is the same as downloading the entire video which you can still do. My favourite tool is yt-dlp
You can also setup a script to automatically download a channels latest vid so you don’t need to check the website anymore.
fun fact: according to sponsorblock, youtube is testing ads that are baked serverside into the video. so one day even downloading might not be ad free
They will never be able to block me just using the mouse to skip forward. If its already downloaded theres zero buffer lag.
I will create another step that converts the format to an open one if they somehow block that too.
Its an accessibility thing for me. Ads literally cause me harm. They cannot possibly win me over i’ll just end up doing something productive instead.
Beat me to it (by several hours).
I’m not watching on YouTube. If I want to watch, I’ll download it first. yt-dlp on the desktop, seal (yt-dlp underneath) on android.
Edit: Big finger problems
I remember when we were still on dial-up and I found a youtube video I wanted to show my brother, I’d let it buffer and load and have to keep the pc on the entire day until he got home from work.
Realmedia has entered the chat
Modern ABRs are actually quite sophisticated, and in most cases you’re unlikely to notice the forward buffer limit. Unstable connection scenarios are going to be the exception where it breaks down.
For best user experience it’s of course good practice to offer media offlining alongside on demand, but some platforms consider it a money-making opportunity to gate this behind a subscription fee.
My internet is intermittently like 100mbps and 256kbps. It sees the 100mbps and acts like it’s going to be that way forever, so doesn’t buffer the whole video while it has the fast speed, then drops entirely when it slows down.
An ABR is generally going to make an estimate based on observed bandwidth and select an appropriate bitrate for that. It’s not out of the question that you run out of forward buffer when your bandwidth takes a nosedive, because the high bitrate video is heavy as all hell and the ABR needs to have observed the drop in bandwidth before it reconsiders and selects a lower bitrate track.
I’m not familiar with ABRs affecting the size of the forward buffer, most commonly these are tweaked based on the type of use-case and scaled in seconds of media.
And then there is youtube which just discards the whole buffer content each time an ad plays. Very sophisticated. Although knowing google that behavoir is likely on purpose
If that were true then users wouldn’t hate and complain about it. This post existing is proof that it’s shit because clearly it’s not as seamless as you’re making it out to be.
The thing is that you can’t notice when it’s working on account of how seamless it is. Yes, sometimes it breaks down, but these are the exceptional cases.