14 ms·

Did Google Cripple Edge’s YouTube Performance?

70 points by janderit 8y ago

mscasts 8y ago
Well Youtube performance used to suck under Firefox as well to the point that I felt that running youtube without the addon "Youtube Classic" in firefox was unbearable. I still run Youtube Classic.
- Bayart 8y ago
  Playback or rendering ? I know my browser has performance issues because it doesn't have hardware-accelerated playback for VP9, I don't know if FF is in the same basket (sounds like it might be). Having an addon that only enables h264 solves that.
  - PlutoIsAPlanet 8y ago
    Afaik YouTube uses user agent sniffing to use slower functions for Firefox.
    - krzyk 8y ago
      YouTube started using Shadow DOM V0 (which is deprecated) some time ago, which is implemented only in Chrome/Chromium, as a result all other browsers render it slower.
      - vezycash 8y ago
        Google also blocked Windows Phone from Google maps. https://arstechnica.com/gadgets/2013/01/google-maps-windows-phone-and-an-avoidable-mess/ https://arstechnica.com/gadgets/2013/01/google-maps-windows-...
        zmonkeyz 8y ago
        Not to mention pulling the Windows Phone Youtube app that Microsoft wrote because it wasnt made with html5. Nevermind that the android or iphone app didnt support it either.
        mscasts 8y ago
        They also are still blocking pretty much all browsers except Chrome from Google Hangouts.
  - mscasts 8y ago
    Both but mainly playback.
t0astbread 8y ago
Wow, that's a rude standpoint to take. Of course a monopoly can implement technically better solutions more quickly but the author is completely disregarding the benefits of distributed power: That you as a consumer can decide what you need and what you will use and that you're not a big company's slave. Ignoring this is just short-sighted and an asshole move towards others, IMO. Oh and also, EdgeHTML might not have been the best engine but saying it "sucks shit" is a bit extreme.
- jnoring 8y ago
  Author here. Diplomacy isn't my strong suit, and I don't care if you think me rude. I'm not saying consumers shouldn't have choice. That isn't at all what I'm saying. What I am saying is: Chrome didn't arrive at its market share by some corporate scheme to enforce usage by vertical tying. Chrome arrived at that market share by making a great piece of software that most consumers like and are happy with. And I take particular offense to a browser that is bundled with an operating system by default--as opposed to one a user has to go out of their way to download--as being some panacea to Google. Edge and Safari likely have meaningful market share for no reason beyond: they ship with Windows and macOS respectively. Chrome certainly ships as the default on Chromebooks and Android, but you'd be remiss to ignore that Chrome is the most used browser on Windows, and second most common on macOS. Those are consumers making a choice. Yet so many people in programming and tech circles don't stop to ask themselves why consumers are opting for Chrome. So maybe the final line in my post seems harsh, but here's the reality: Edge is only a thing because it ships with Windows. Beyond that I see no technical reason to favor it over Chrome or even Firefox. On Windows, consumers have already spoken: they like Chrome. Lastly, Edge is a pretty questionable argument in favor of "consumer choice", given it represents a choice no consumer actually made beyond "default browser in Windows."
  - ethelward 8y ago
    IMHO, Chrome arrived at such a market share mostly through (i) ubiquitously ad campaigns (subways, buses, on the web, ...); (ii) an strong push towards the users (it's quite easy when your editor is in charge of what is basically the homepage of the internet for most people); (iii) an aggressive bundling strategy with dozens of other programs (JVM, freewares, ...).
sjansen 8y ago
If you look past the clearly frustrated tone, what he says makes sense. It's much more likely that Google was trying to improve their user experience by working around an IE bug, than deliberately harming their user experience for a PR article. https://en.wikipedia.org/wiki/Hanlon%27s_razor https://en.wikipedia.org/wiki/Hanlon%27s_razor
amatwl 8y ago
I have to say the author's attitude does not come off particularly well in this post, at least in my opinion. Just seems rude.
- jnoring 8y ago
  Author here. That's fair if you feel that way. What I would say, however is "rude" is attributing malice when there is a perfectly reasonable alternate explanation to a story. That's really what I take umbrage with; our standards for truth should be higher than the insinuations of a MSFT intern with no clear evidence, and a total lack of the net community as a whole even to consider alternate possibilities.
fatnoah 8y ago
In all fairness to Google, they've been steadily degrading Chrome's performance on all websites for quite some time now.
bitpush 8y ago
Disappointed with people talking about the tone and not the content of the blogpost. Didnt realize it was an intern (who probably lacks experience with the code) who made the false accusation in the first place. Even if that werent true, making baseless arguments do not sound like the new MSFT that HN tries to project.
- ABCLAW 8y ago
  The tone of the article and the posts in this thread are literally the most important piece of content. This is the person who was directly called out for a specific action which broke one element of Edge functionality. His commits and comments have a very strong "well, fuck them" vibe to them. I read the initial post that led to this blog reply. It accused google of not playing well with others. The engineer in question is in this thread saying he doesn't give a shit what people think about his tone and that he isn't diplomatic. He is literally the criticism personified - the hubris made person. No wonder people are talking about it.
  - lern_too_spel 8y ago
    > This is the person who was directly called out for a specific action which broke one element of Edge functionality. No, the intern called out Google, not this guy.
    - ABCLAW 8y ago
      If my organization is called out, and the citation given is my specific work, I would consider that someone called me on my work, even if they hadn't named me. Sorry if I wasn't clear.
      - lern_too_spel 8y ago
        He doesn't work for Google, nor did he implement the YouTube feature the intern called out. He implemented something similar in his own project. This was clear in the article.
wonthegame 8y ago
The performance on Safari is abysmal too.
- bunnycorn 8y ago
  Not to mention they removed 4K because... Google force-feeding webm upon us.
  - lern_too_spel 8y ago
    Let's not forget that Apple tried to force license-encumbered formats into web standards. If the publishers don't play along, any resulting poor experience is Apple's fault. Don't support bad actors.
    - bunnycorn 8y ago
      Which ones then?
Lapsed 8y ago
For me, Edge is the only browser that behaves properly when a Youtube video is playing but is not in focus with hardware acceleration on (the exact scenario is I have a game on my main monitor with a Youtube video playing on my other monitor). Both Chrome and Firefox stop loading the video and show the spinning buffering symbol after a few minutes. All three browsers work fine with hardware acceleration off.
smileypete 8y ago
Seems Youtube has gotten slower for me over the years... Playing an HD video Firefox 55 (w/classic, MP4) uses 30%, Chrome 70 uses 20%, but Videolan uses a mere 4%! It's puzzling why a modern browser uses far more CPU than a standalone app these days.
- dhnsmakala 8y ago
  Did you compare the baseline browser usage to browser + video?
  - smileypete 8y ago
    Just a clean restart then a single tab for youtube, and looking at the total CPU in Process Hacker.
ksec 8y ago
I am most frustrated not with Edge going to Blink, Not the tone of the article, but Web Technology as a whole. I remember when I was doing Web programming in pre IE6 era, it was insanely frustrating. Pulling your hairs out trying to make the simplest thing to work across browsers. I didn't hate IE6 for Active X or its Standard forcing practice ( It is actually both good and bad ), but M$ unwilling to improve IE when there are lots of low hanging fruit, while actively stopping other browser from gaining ground. Nearly 20 years passed, it still looks like a bloody mess. Yes it was a lot better than IE5 / 6 era, CSS still sucks, ( just less ), Cross browser handling of Keypress and Video still a mess? And we have now arrived to a point where Chrome/ Blink becomes a living reference implementation to test out ideas before it is "standardise". Web Pages that only works well in Chrome? ( I am not even talking about Web Apps which somehow 90% of the case we don't need but we still made it ) Was this what we envisioned the web 20 years ago?
kup0 8y ago
Now, I can't prove malice, but this does strike me odd: > "Mozilla's Technical Program Manager has stated that YouTube's Polymer redesign relies heavily on the deprecated Shadow DOM v0 API, which is only available in Chrome." https://news.slashdot.org/story/18/07/25/1128225/google-has-made-youtube-slower-on-edge-and-firefox-mozilla-alleges https://news.slashdot.org/story/18/07/25/1128225/google-has-... Relying on a deprecated API, only available to Chrome, and forcing a slow polyfill on everyone else seems like a sure fire way to try to push people to use Chrome and make things slow for everyone else. Whether or not this latest Edge / div debacle was malicious or not, the company has a history of doing stuff leaning that direction.
InGodsName 8y ago
Biggest problem is that apple doesn't support hardware acceleration for vp9/8 decoding which YouTube uses. So our battery empties faster on appple devices. Neither google is doing anything about it nor apple. Google is pushing their own Chromebook and my friends who are pissed off at their YouTube performance in regard to battery switched to Chromebook.

70 points by janderit 8y ago

Well Youtube performance used to suck under Firefox as well to the point that I felt that running youtube without the addon "Youtube Classic" in firefox was unbearable. I still run Youtube Classic.

Bayart 8y ago
Playback or rendering ? I know my browser has performance issues because it doesn't have hardware-accelerated playback for VP9, I don't know if FF is in the same basket (sounds like it might be). Having an addon that only enables h264 solves that.
- PlutoIsAPlanet 8y ago
  Afaik YouTube uses user agent sniffing to use slower functions for Firefox.
  - krzyk 8y ago
    YouTube started using Shadow DOM V0 (which is deprecated) some time ago, which is implemented only in Chrome/Chromium, as a result all other browsers render it slower.
    - vezycash 8y ago
      Google also blocked Windows Phone from Google maps. https://arstechnica.com/gadgets/2013/01/google-maps-windows-phone-and-an-avoidable-mess/ https://arstechnica.com/gadgets/2013/01/google-maps-windows-...
      - zmonkeyz 8y ago
        Not to mention pulling the Windows Phone Youtube app that Microsoft wrote because it wasnt made with html5. Nevermind that the android or iphone app didnt support it either.
        mscasts 8y ago
        They also are still blocking pretty much all browsers except Chrome from Google Hangouts.
- mscasts 8y ago
  Both but mainly playback.

t0astbread 8y ago

Wow, that's a rude standpoint to take. Of course a monopoly can implement technically better solutions more quickly but the author is completely disregarding the benefits of distributed power: That you as a consumer can decide what you need and what you will use and that you're not a big company's slave. Ignoring this is just short-sighted and an asshole move towards others, IMO. Oh and also, EdgeHTML might not have been the best engine but saying it "sucks shit" is a bit extreme.

jnoring 8y ago
Author here. Diplomacy isn't my strong suit, and I don't care if you think me rude. I'm not saying consumers shouldn't have choice. That isn't at all what I'm saying. What I am saying is: Chrome didn't arrive at its market share by some corporate scheme to enforce usage by vertical tying. Chrome arrived at that market share by making a great piece of software that most consumers like and are happy with. And I take particular offense to a browser that is bundled with an operating system by default--as opposed to one a user has to go out of their way to download--as being some panacea to Google. Edge and Safari likely have meaningful market share for no reason beyond: they ship with Windows and macOS respectively. Chrome certainly ships as the default on Chromebooks and Android, but you'd be remiss to ignore that Chrome is the most used browser on Windows, and second most common on macOS. Those are consumers making a choice. Yet so many people in programming and tech circles don't stop to ask themselves why consumers are opting for Chrome. So maybe the final line in my post seems harsh, but here's the reality: Edge is only a thing because it ships with Windows. Beyond that I see no technical reason to favor it over Chrome or even Firefox. On Windows, consumers have already spoken: they like Chrome. Lastly, Edge is a pretty questionable argument in favor of "consumer choice", given it represents a choice no consumer actually made beyond "default browser in Windows."
- ethelward 8y ago
  IMHO, Chrome arrived at such a market share mostly through (i) ubiquitously ad campaigns (subways, buses, on the web, ...); (ii) an strong push towards the users (it's quite easy when your editor is in charge of what is basically the homepage of the internet for most people); (iii) an aggressive bundling strategy with dozens of other programs (JVM, freewares, ...).

sjansen 8y ago

If you look past the clearly frustrated tone, what he says makes sense. It's much more likely that Google was trying to improve their user experience by working around an IE bug, than deliberately harming their user experience for a PR article. https://en.wikipedia.org/wiki/Hanlon%27s_razor https://en.wikipedia.org/wiki/Hanlon%27s_razor

amatwl 8y ago

I have to say the author's attitude does not come off particularly well in this post, at least in my opinion. Just seems rude.

jnoring 8y ago
Author here. That's fair if you feel that way. What I would say, however is "rude" is attributing malice when there is a perfectly reasonable alternate explanation to a story. That's really what I take umbrage with; our standards for truth should be higher than the insinuations of a MSFT intern with no clear evidence, and a total lack of the net community as a whole even to consider alternate possibilities.

fatnoah 8y ago

In all fairness to Google, they've been steadily degrading Chrome's performance on all websites for quite some time now.

bitpush 8y ago

Disappointed with people talking about the tone and not the content of the blogpost. Didnt realize it was an intern (who probably lacks experience with the code) who made the false accusation in the first place. Even if that werent true, making baseless arguments do not sound like the new MSFT that HN tries to project.

ABCLAW 8y ago
The tone of the article and the posts in this thread are literally the most important piece of content. This is the person who was directly called out for a specific action which broke one element of Edge functionality. His commits and comments have a very strong "well, fuck them" vibe to them. I read the initial post that led to this blog reply. It accused google of not playing well with others. The engineer in question is in this thread saying he doesn't give a shit what people think about his tone and that he isn't diplomatic. He is literally the criticism personified - the hubris made person. No wonder people are talking about it.
- lern_too_spel 8y ago
  > This is the person who was directly called out for a specific action which broke one element of Edge functionality. No, the intern called out Google, not this guy.
  - ABCLAW 8y ago
    If my organization is called out, and the citation given is my specific work, I would consider that someone called me on my work, even if they hadn't named me. Sorry if I wasn't clear.
    - lern_too_spel 8y ago
      He doesn't work for Google, nor did he implement the YouTube feature the intern called out. He implemented something similar in his own project. This was clear in the article.

wonthegame 8y ago

The performance on Safari is abysmal too.

bunnycorn 8y ago
Not to mention they removed 4K because... Google force-feeding webm upon us.
- lern_too_spel 8y ago
  Let's not forget that Apple tried to force license-encumbered formats into web standards. If the publishers don't play along, any resulting poor experience is Apple's fault. Don't support bad actors.
  - bunnycorn 8y ago
    Which ones then?

Lapsed 8y ago

For me, Edge is the only browser that behaves properly when a Youtube video is playing but is not in focus with hardware acceleration on (the exact scenario is I have a game on my main monitor with a Youtube video playing on my other monitor). Both Chrome and Firefox stop loading the video and show the spinning buffering symbol after a few minutes. All three browsers work fine with hardware acceleration off.

smileypete 8y ago

Seems Youtube has gotten slower for me over the years... Playing an HD video Firefox 55 (w/classic, MP4) uses 30%, Chrome 70 uses 20%, but Videolan uses a mere 4%! It's puzzling why a modern browser uses far more CPU than a standalone app these days.

dhnsmakala 8y ago
Did you compare the baseline browser usage to browser + video?
- smileypete 8y ago
  Just a clean restart then a single tab for youtube, and looking at the total CPU in Process Hacker.

ksec 8y ago

I am most frustrated not with Edge going to Blink, Not the tone of the article, but Web Technology as a whole. I remember when I was doing Web programming in pre IE6 era, it was insanely frustrating. Pulling your hairs out trying to make the simplest thing to work across browsers. I didn't hate IE6 for Active X or its Standard forcing practice ( It is actually both good and bad ), but M$ unwilling to improve IE when there are lots of low hanging fruit, while actively stopping other browser from gaining ground. Nearly 20 years passed, it still looks like a bloody mess. Yes it was a lot better than IE5 / 6 era, CSS still sucks, ( just less ), Cross browser handling of Keypress and Video still a mess? And we have now arrived to a point where Chrome/ Blink becomes a living reference implementation to test out ideas before it is "standardise". Web Pages that only works well in Chrome? ( I am not even talking about Web Apps which somehow 90% of the case we don't need but we still made it ) Was this what we envisioned the web 20 years ago?

kup0 8y ago

Now, I can't prove malice, but this does strike me odd: > "Mozilla's Technical Program Manager has stated that YouTube's Polymer redesign relies heavily on the deprecated Shadow DOM v0 API, which is only available in Chrome." https://news.slashdot.org/story/18/07/25/1128225/google-has-made-youtube-slower-on-edge-and-firefox-mozilla-alleges https://news.slashdot.org/story/18/07/25/1128225/google-has-... Relying on a deprecated API, only available to Chrome, and forcing a slow polyfill on everyone else seems like a sure fire way to try to push people to use Chrome and make things slow for everyone else. Whether or not this latest Edge / div debacle was malicious or not, the company has a history of doing stuff leaning that direction.

InGodsName 8y ago

Biggest problem is that apple doesn't support hardware acceleration for vp9/8 decoding which YouTube uses. So our battery empties faster on appple devices. Neither google is doing anything about it nor apple. Google is pushing their own Chromebook and my friends who are pissed off at their YouTube performance in regard to battery switched to Chromebook.