Silverlight vs. Widevine

For the more technical/geeky chat subjects!

Moderator: satrow

User avatar
Moonchild
Pale Moon guru
Pale Moon guru
Posts: 21960
Joined: Sun, 28 Aug 2011, 17:27
Location: 58.5°N 15.5°E
Contact:

Silverlight vs. Widevine

Unread postby Moonchild » Mon, 21 May 2018, 13:40

Decided to make a post about this because there's such a clear difference between the two, and not just subjectively.

I've compared both delivery methods (Silverlight/DRM and Widevine/EME) on Netflix, using the same time points in a movie to compare side by side.
I chose Dr. Strange, it being a modern movie with lots of detail.

The attached screenshots speak for themselves, I think. The Widevine frames are clearly closer to SD than HD, while the Silverlight frames are crisp and show proper detail.

The clear winner is Silverlight. It remains to be explained why EME is giving such sub-standard quality.
Attachments
s1-sl.png
s1-wv.png
s2-sl.png
s2-wv.png
s3-wv.png
s3-sl.png
Improving Mozilla code: You know you're on the right track with code changes when you spend the majority of your time deleting code.

"If you want to build a better world for yourself, you have to be willing to build one for everybody." -- Coyote Osborne

User avatar
LimboSlam
Board Warrior
Board Warrior
Posts: 1024
Joined: Mon, 09 Jun 2014, 04:43
Location: USA

Re: Silverlight vs. Widevine

Unread postby LimboSlam » Tue, 22 May 2018, 04:46

Yes, always thought Silverlight looked sharper. And it does!

What a wonderful year to browse the web with plug-in capabilities.


Thanks man! Keep it up!! :) :thumbup: :clap: :clap:
With Pale Moon by my side, surfing the web is quite enjoyable and takes my headaches away! :)
God is not punishing you, He is preparing you. Trust His plan, not your pain.#‎TrentShelton #‎RehabTime

CharmCityCrab
Astronaut
Astronaut
Posts: 522
Joined: Thu, 25 Jun 2015, 00:47
Location: Eastern US

Re: Silverlight vs. Widevine

Unread postby CharmCityCrab » Tue, 22 May 2018, 05:38

It seems like there is potentially a lot of room for the differences to be in the way Netflix is using the technology vs. an inherent different in picture quality using each Silverlight and Wildvine.

As someone who isn't supertechnical, I am not sure if any of these reasons are truly possible reasons, but it seems like there could some reason similar to them, at least hypothetically:

Netflix is known to encode a lot of its catelog at several bitrates per movie or episode, delivering the end user a higher bitrate if they have a faster congestant and Netflix is less congested, and a lower bitrate if they have a slower connection and Netflix is more congested, something that sometimes results in changes to picture quality in either or both directions even as one is watching.

If you've noticed a show or movie starting blocky like a old Nintendo game and gradually increasing in definition, that's what's going on. Its also building a small buffer.

What if one codec takes less bandwidth than the other, alllowing Netflix to give you their better quality version of that codec more often?

What if Netflix arbitrarily maintains its library differently depending on the codec? For example, of one codec offers things at 420p or 1080p, and the other has 440p and 720p available. In some situations, one codec would look better and in others, the other codec, for reasons having nothing to do with the codecs themselves.

Its also possible that Netflix's infrastructure is more optimized towards Silverlight, having been using that technology longer. Its also possible that people's home infrastructure is more optimized for it- for example, using a Windows PC, which is made by the same company as Silverlight, or a certain chip architecture (i.e. Intel vs ARM or whatever the case may be). There are also a bunch of hardware boxes that are used with TVs that may have varying capabilities even within the same line, as well as phones and devices that have different capabilities. The thing is, though- ultimately, in the long run, both Netflix's internal hardware and software, and that of the end user is destined to eventually turn over as hardware breaks or becomes obselete and software gets updates or is changed to something considered more advanced.

In any event, all that aside (And not just because I'm probably wrong :) ), it seems like there is one very good reason for Netflix and other streaming providers to switch off Silverlight:

Microsoft has ended development on Silverlight except for patches and big fixes, and has plans to stop even issuing those eventually. Consequentially, Chrome and Firefox have stopped officially supporting it, and Edge has never and will never support it.

Silverlight lost the platform war on two fronts, to the older Flash initially, and now increasingly to the newer HTML5-based DRM solutions. Many times the loser in the platform war is actually the better technology- some people will staunchly maintain that Laser-Discs were superior to DVDs and that HD-DVD was superior to Blue-Ray. But what can you do? They stopped putting out new movies in the losing formats eventually, and new players, so people who who were determined to stick with them were limited to what they already had plus what they could buy used, with replacement players and discs becoming increasingly hard to find and expensive to obtain, and even if one had unlimited time and money to go after the used stuff to replace, repair, and expand what they have, they'd never be able to buy new movies made after a certain date in those formats.

Eventually, there isn't going to be a streaming service that streams using Silverlight, and eventually any browser offering it as a plug-in will be pirating unsupported old versions, after Microsoft finishes phasing it out and allowing for its distribution.

And, in the interest of looking on the brightside- Silverlight as far as I know never ran properly on Linux, whereas EME should mean that Netflix and the like can offer streaming on Linux through the native browser implentations in Firefox and Chromium. This also could make things a bit easier on streaming boxes of the future that have some sort of Linux at their core.

The only real problem I have with the DRM aspect of EME is that it may be prohibitively expensive for smaller browsers, reducing user choice in that field and the ability of smaller browsers to compete.

I expect some sort of DRM to be on what is essentially a rental subscription service- because you're renting, not buying, and they have to at least try to enforce that somehow or you've defacto bought tens of thousands of hours of ebtertainment for $10 as a one-time fee, which isn't a financially workable model. What bothers me more from a principle perspective is when someone sells a video at full price in DVD format or tied to a specific website and platform with DRM- if one buys something specific at full price, they should be able to male archival copies for personal use and use it when they want on any device or OS they can get to play it- so there shouldn't be DRM on items for sale. But a monthly all you can stream rental service? I'm cool with DRM on that.

The potential of smaller browsers being potentially unable to use this and compete with the larger browsers bothers me, but has anyone actually checked with the people who handle the DRM and asked about allowing their small browser to be allowed to try to implement a version of it? I always hear it assumed that it wouldn't be allowed and that the powers that be behind it would demand some ridiculous sum of money that is more money than small browser companies themselves would make in a million years, but I never here any word about actual inquries. What if they would be happy to let small browsers browsers implement it and only charge on a per-user or per-download basis that would scale and make the total cost much less than they charge big browsers with lots of users?

And, of course, with Pale Moon specifically, cost may be a non-issue because last I heard Pale Moon was taking the stance that they didn't want it at any price, or even for free. If that's changed, I would suggest at least reaching out to whomever controls EME DRM and asking what the price would be for a browser with a smaller number of users owned by a small business that can't match Google in terms of a flat fee. Maybe you'd get a surprisingly positive response.

Also, it may be that forks that stay closer to the browsers from wherence they came are just able to carry some of the implementations over with the rest.

User avatar
SpockFan02
Lunatic
Lunatic
Posts: 459
Joined: Sun, 24 Sep 2017, 16:35

Re: Silverlight vs. Widevine

Unread postby SpockFan02 » Tue, 22 May 2018, 06:27

CharmCityCrab wrote:The potential of smaller browsers being potentially unable to use this and compete with the larger browsers bothers me, but has anyone actually checked with the people who handle the DRM and asked about allowing their small browser to be allowed to try to implement a version of it? I always hear it assumed that it wouldn't be allowed and that the powers that be behind it would demand some ridiculous sum of money that is more money than small browser companies themselves would make in a million years, but I never here any word about actual inquries. What if they would be happy to let small browsers browsers implement it and only charge on a per-user or per-download basis that would scale and make the total cost much less than they charge big browsers with lots of users?

And, of course, with Pale Moon specifically, cost may be a non-issue because last I heard Pale Moon was taking the stance that they didn't want it at any price, or even for free. If that's changed, I would suggest at least reaching out to whomever controls EME DRM and asking what the price would be for a browser with a smaller number of users owned by a small business that can't match Google in terms of a flat fee. Maybe you'd get a surprisingly positive response.
Maybe, but why should small browser developers need to ask for permission to implement decryption modules? Even if they are allowed, it's not the best situation when that allowance is conditional and in the hands of "the powers that be." That's a big part of what's scary about EME: Someone can't just follow the W3C specs and write a browser that's compatible with the Web if you need the permission of some corporation to be compatible. And why would, say, Google want there to be more alternative browsers? Wouldn't they have a vested interest in retaining as many users as possible?

I don't have a comment on which plug-ins have higher-quality playback or why, but that's something on EME.

User avatar
Moonchild
Pale Moon guru
Pale Moon guru
Posts: 21960
Joined: Sun, 28 Aug 2011, 17:27
Location: 58.5°N 15.5°E
Contact:

Re: Silverlight vs. Widevine

Unread postby Moonchild » Tue, 22 May 2018, 14:49

Long essay aside: I have a HD subscription with them, and made sure to give ample lead-in time to switch bitrates to prevent SD quality while things are buffered. That is NOT what is going on -- the screenshots are final quality screenshots. considering my ISP has its own Netflix servers and I have plenty of bandwidth, there is no technical limitation in the way it can be streamed and the resulting quality of the screenshots is purely a Widevine vs. Silverlight thing on the same subscription using the same PC, same connection and same movie.

As far as costs for DRM goes, that is entirely something between the streaming service and the DRM provider and has nothing to do with the client in use. If I pay for a certain quality, the streaming service has to deliver that quality.
Last edited by Moonchild on Tue, 22 May 2018, 14:52, edited 1 time in total.
Improving Mozilla code: You know you're on the right track with code changes when you spend the majority of your time deleting code.

"If you want to build a better world for yourself, you have to be willing to build one for everybody." -- Coyote Osborne

CharmCityCrab
Astronaut
Astronaut
Posts: 522
Joined: Thu, 25 Jun 2015, 00:47
Location: Eastern US

Re: Silverlight vs. Widevine

Unread postby CharmCityCrab » Tue, 22 May 2018, 17:32

SpockMan02 wrote:Maybe, but why should small browser developers need to ask for permission to implement decryption modules? Even if they are allowed, it's not the best situation when that allowance is conditional and in the hands of "the powers that be." That's a big part of what's scary about EME: Someone can't just follow the W3C specs and write a browser that's compatible with the Web if you need the permission of some corporation to be compatible. And why would, say, Google want there to be more alternative browsers? Wouldn't they have a vested interest in retaining as many users as possible?


Is the DRM implementation used in EME actually completely owned by Google? I honestly assumed it was owned (or controlled) by a consortium that included Google, but that was only an assumption. If it's directly owned by Google, that would be a darker scenario, not because of any intrinsic problem with Google (I mean, they have problems, but don't they all?), but because a standard shouldn't be in the hands of one company with one browser product that competes against other browsers they license to. The reason I assumed consortium is because I figured that there was no way Mozilla, Microsoft, and Apple would implement this if it was *totally* in Google's control.

In theory, Google has said it's goal with Chrome is just to get people onto the net on a browser that is compatible with modern standards and keeps innovating for greater speed and functionality, to drive more visitors more frequently to sites with it's ads, Google cloud services, etc.. Originally, when they introduced Chrome, they said they'd be just as happy if Chrome only had a small market share and drove other browsers to adopt it's features and innovate as if they controlled the market. However, in practice, of course that hasn't always been the case- with Google's websites pushing users of other browsers specifically towards Chrome, Google sites and products being optimized for Chrome, and the spread of Google specific HTML markup on websites.

I understand something like DRM is inherently a problem for open-source or libre software, because it undermines the idea that the complete code of a product is available and forkable by anyone with the knowledge, skills, and I guess in some cases person-power (For larger projects) to do. For example, if someone wants to fork Firefox in the future, they probably won't be able to bring the DRM part of the EME module with them, and even existing users can't see or modify the code within the container Firefox has set up for the DRM (Though you can opt to run Firefox without it). That said, this can sometimes be avoided by the less ideological advocates of open-source software, as a practical matter, if a module like this is licensed freely to any browser project that wants it. It still wouldn't make it open source, but it would mean you could fork whatever, or write your own browser, and keep or add an implementation.

User avatar
SpockFan02
Lunatic
Lunatic
Posts: 459
Joined: Sun, 24 Sep 2017, 16:35

Re: Silverlight vs. Widevine

Unread postby SpockFan02 » Tue, 22 May 2018, 17:53

Widevine is owned by Google (although it's not the only EME).

Mercury
Fanatic
Fanatic
Posts: 202
Joined: Fri, 01 Nov 2013, 23:26
Location: Toronto, Canada

Re: Silverlight vs. Widevine

Unread postby Mercury » Tue, 22 May 2018, 18:23

Curious. What do you get when you watch one of Netflix's test patterns? The actual resolution of the video stream should be displayed at the top right.
https://www.netflix.com/watch/80018588

User avatar
Moonchild
Pale Moon guru
Pale Moon guru
Posts: 21960
Joined: Sun, 28 Aug 2011, 17:27
Location: 58.5°N 15.5°E
Contact:

Re: Silverlight vs. Widevine

Unread postby Moonchild » Tue, 22 May 2018, 21:57

Mercury wrote:Curious. What do you get when you watch one of Netflix's test patterns? The actual resolution of the video stream should be displayed at the top right.
https://www.netflix.com/watch/80018588


Both go up to 1280x720 (720p) after the lead-in and up to 3Mbps streaming. Unfortunately EME doesn't allow html5 "stats for nerds" or anything else on the video element (by design) so I can't verify it with the browser's own tools, but the still image is equal for either Silverlight or Widevine.
The actual moving image fidelity of Widevine sucks in comparison, which is what really matters, of course!
Improving Mozilla code: You know you're on the right track with code changes when you spend the majority of your time deleting code.

"If you want to build a better world for yourself, you have to be willing to build one for everybody." -- Coyote Osborne

User avatar
Tomaso
Keeps coming back
Keeps coming back
Posts: 944
Joined: Thu, 23 Jul 2015, 16:09
Location: Norway

Re: Silverlight vs. Widevine

Unread postby Tomaso » Sat, 26 May 2018, 12:43

To get the best quality possible at Netflix (if you have enough bandwidth for it), you should also make sure that you've set your choice to "HIGH" here:
Account > My profile > Playback settings
Theoretically, the "Auto" setting should give you the best quality that is supported by your connection.
..but from what I've read, this isn't always the case.

User avatar
Moonchild
Pale Moon guru
Pale Moon guru
Posts: 21960
Joined: Sun, 28 Aug 2011, 17:27
Location: 58.5°N 15.5°E
Contact:

Re: Silverlight vs. Widevine

Unread postby Moonchild » Sat, 26 May 2018, 15:14

It's not a matter of settings!
The test patterns showed very clearly that NetFlix has no issue getting to its maximum quality with auto streaming, too.
In addition, my ISP is a cable company that has its own Netflix streaming servers, so it's all local-net traffic. The streaming itself is not the problem. My downstream is 150Mbit, which is more than enough for their 3Mbit streams @ max bitrate.
Once again, it's compared on the same PC, same browser, same account, same time of day, and verified a few times by switching back and forth between the two methods. I've done everything needed to minimize other factors before posting this comparison.
Last edited by Moonchild on Sat, 26 May 2018, 15:15, edited 2 times in total.
Improving Mozilla code: You know you're on the right track with code changes when you spend the majority of your time deleting code.

"If you want to build a better world for yourself, you have to be willing to build one for everybody." -- Coyote Osborne

User avatar
Tomaso
Keeps coming back
Keeps coming back
Posts: 944
Joined: Thu, 23 Jul 2015, 16:09
Location: Norway

Re: Silverlight vs. Widevine

Unread postby Tomaso » Sat, 26 May 2018, 16:20

Yes, I understand that you're comparing the two plugins, and I have no doubt that you have this stuff under full control, Moonchild. :)
Just pointing out that people should check their playback settings too.

User avatar
Moonchild
Pale Moon guru
Pale Moon guru
Posts: 21960
Joined: Sun, 28 Aug 2011, 17:27
Location: 58.5°N 15.5°E
Contact:

Re: Silverlight vs. Widevine

Unread postby Moonchild » Sat, 26 May 2018, 16:44

Sorry, but to me it sounded like your were critiquing my comparison by bringing that up. If that wasn't your intent, then forget I said anything :)
Improving Mozilla code: You know you're on the right track with code changes when you spend the majority of your time deleting code.

"If you want to build a better world for yourself, you have to be willing to build one for everybody." -- Coyote Osborne

User avatar
Tomaso
Keeps coming back
Keeps coming back
Posts: 944
Joined: Thu, 23 Jul 2015, 16:09
Location: Norway

Re: Silverlight vs. Widevine

Unread postby Tomaso » Sat, 26 May 2018, 17:17

Oh, I see.
No, that was not my intention. :)

On a side note, this is actually the first time I've ever heard about Widevine.

Brulsturm
Newbie
Newbie
Posts: 3
Joined: Sat, 10 Feb 2018, 10:16

Re: Silverlight vs. Widevine

Unread postby Brulsturm » Fri, 14 Sep 2018, 08:46

Interesting comparison Moonchild.
In wich browser did you test the plugins? And did you notice differences in processorload?
I ask this because our Dutch public broadcaster (NPO) use some sort of nasty DRM wich brings my laptop's processor to a 90%-95% load and a stuttering, unwatchable videopicture (PM 28.0.1 x64, Silverlight x64 5.1.50907.0, Win7 x64, i5-540M).
Its a riduculous webplayer anyway, with absolutly no settings at all (oh yes, subtitles on/off) so the quality isn't adjustable to the users. When I test the same webvideo in FF, it plays smooth, after downloading a DRM-"blob", wich is apperantly WidevineCdm. In that browser the same video(stream) only uses 15%-25% CPUload.
Just wondering if the observed differences in quality could be the reason for the differences in processorload, although I have to admit it looks more like the never ending story, named: hardware/software-decoding.
Last edited by Brulsturm on Fri, 14 Sep 2018, 08:48, edited 2 times in total.

User avatar
Tomaso
Keeps coming back
Keeps coming back
Posts: 944
Joined: Thu, 23 Jul 2015, 16:09
Location: Norway

Re: Silverlight vs. Widevine

Unread postby Tomaso » Fri, 14 Sep 2018, 21:49

Brulsturm wrote:stuttering, unwatchable videopicture (PM 28.0.1 x64, Silverlight x64 5.1.50907.0, Win7 x64, i5-540M)

@ Brulsturm:
Would you mind testing with Pale Moon v27.9.4(x64) + Silverlight v5.1.50907.0?
I ask because I'm experiencing similar issues with Pale Moon v28.0.1(x64) at a Norwegian broadcasting site.
..but since the Silverlight videos play just fine when using v27.9.4(x64), I suspect that there might be some sort of regression present.

You can use the portable version of Pale Moon v27.9.4(x64), so that it doesn't interfere with your installed version:
ftp://palemoon:get@archive.palemoon.org ... .win64.exe


Return to “Technical chat”

Who is online

Users browsing this forum: No registered users and 1 guest