G'bye 3rd Party ASP/AVC Filters in WMP/MCE on Win7?

[CruNcher]

Are you absolutely sure lexor ? i mean you have to be ultra sure about this claim (better triple check) as if this would be true and Microsoft has now began to even wider the spectrum of this and that even after being criticized for their "We have to protect our own Apps from evilness and dumb users" move it would be outrageous, though this is still RC not final and maybe Microsoft want's to make people and 3rd partys shaking in their shoes a little (to show them who the boss is) and then on RTM day they gonna say "hey was just a joke"
Though seeing that Win 7 RC is out for a while now and no one else on the .Net saying Microsoft tightened it even more in terms of 3rd party media playback behavior makes me not wanna believe in this story somehow, common sense forbids it i would rather believe they changed something about the Graph building that had a bad effect and now it broke entirely with some filters

[Shakey_Jake33]

I can't say I've had a single problem getting ffdshow (or any other 3rd party codec) to play nice with MPC-HC in Windows 7 RC tbh.

[saint-francis]

Same here. I use FFDShow with MPC HC every day on Windows 7 RC. I haven't tried any other players but I can say for certain that when ever I render any file in graphedit it always uses MS filters.

[benwaggoner]

Originally Posted by lexor

That is demonstrably false:

First, benwaggoner said that MPC had to be fixed.

If it has its own graph builder, it shouldn't need to be fixed.

I'm getting unclear info here on what is or is not working. If there's an app developer who thinks they should be getting access to 3rd party filters but can't make it work, that's a bug and they should file that on connect or PM me.

Any app developer should be able to use any DirectShow or MFC that they want to in their media pipeline.

Some may require an update to do so, of course.

[Snowknight26]

Why the horrible image quality for Xvid files? It's almost as if the Microsoft decoders are skipping deblocking.

MPC-HC's internal decoder:


WMP12 using Microsoft decoder:

[Sagekilla]

It looks more like they're going overkill on deblocking and killing fine detail. There's barely any from the WMP decoded one.

[LoRd_MuldeR]

Originally Posted by Snowknight26

Why the horrible image quality for Xvid files? It's almost as if the Microsoft decoders are skipping deblocking.

Xvid uses MPEG-4 ASP. And MPEG-4 ASP does not have any kind of deblocking!

Yes, many MPEG-4 ASP decoders have a deblocking filter. But that is optional post-processing. It's not an integral part of the decoder

Micro$oft's decoder apparently doesn't offer that option. This is yet another reason why allowing the user choose his/her preferred decoder is so important!

As if there weren't enough already

[clsid]

MPC-HC's internal MPEG-4 ASP decoder uses no deblocking at all. As Sagekilla said, the MS decoder uses too much post-processing, making things look worse.

[LoRd_MuldeR]

Originally Posted by clsid

MPC-HC's internal MPEG-4 ASP decoder uses no deblocking at all. As Sagekilla said, the MS decoder uses too much post-processing, making things look worse.

No matter whether too much post-processing or missing post-processing is the problem, fact is that MPEG-4 ASP doesn't have any pre-defined deblocking. That's why different MPEG-4 ASP decoders offer different type of post-processing (or no post-processing at all). And we must allow the user to decide what he/she prefers! If the M$ decoder uses a type of post-processing that I don't like, but I cannot easily adjust the decoder's post-processing or choose a different decoder, then that makes the entire player useless for me. Who would use a player that screws up the video?

[Manao]

And we must allow the user to decide what he/she prefers!

Not at the cost of complexity. Most of windows users don't even know what deblocking is, so from microsoft's point of view, there is no need to add complexity.

If the M$ decoder uses a type of post-processing that I don't like, but I cannot easily adjust the decoder's post-processing or choose a different decoder, then that makes the entire player useless for me.

Indeed, but you have alternatives. Microsoft has alienated you - but you (and me, and most of Doom9 users) are not mainstream users. I do think the lack of configurability was done on purpose, and is the good choice, from MS's point of view.

Who would use a player that screws up the video?

Somebody that doesn't know the video is screwed up. Roughly 95% of people. Hell, my parents watch TV with an incorrect aspect ratio, as most people with flat screens do. They have the option not to do so, and they don't use it because it's too complicated.

[LoRd_MuldeR]

We don't need more complexity. Just an "advanced" button with advanced options for advanced users

[Trahald]

Originally Posted by LoRd_MuldeR

We don't need more complexity. Just an "advanced" button with advanced options for advanced users

Ask Jdobbs about advanced settings. he probably sits with his mouse over the 'submit reply' button for a half an hour before he submits a post that reveals a hidden advanced setting in one of his utilities. He knows while you can say only advanced users should touch the setting (and more importantly deal with the consequences), there will be a jump in bug reports due to people messing with the setting that have no business doing so.

As you can tell while I wish microsoft didnt have to , I understand why this was done. Having worked on the phone for years in the past.

[clsid]

Windows is full of settings and hidden tweaks.

An option for post-processing would be one that can do no harm at all.

[Manao]

An option for post-processing would be one that can do no harm at all.

Never doubt the capability of users to create trouble with seemingly harmless options.

And in this case, most probably there's a switch in the registry to reduce/remove deblocking. But if it is, it's hidden and undocumented, thus far less liable to generate bug reports.

[benwaggoner]

Originally Posted by Manao

Never doubt the capability of users to create trouble with seemingly harmless options.

Oh, yeah.

For every option, you have to weigh

• The probability of a user who would benefit using it, proportional to that benefit.
• The probably of a user who shouldn't use it using it anyway, proportional to the cost using it.
• The additional test cost of testing everything each way for each option. Four on/off options just increased requirements for full test coverage by 16x!

In scrubbing the UX for a forthcoming H.264 encoder update, I pushed the team to take options out whenever we could find a good heuristic. For example, bundling speed/quality features into the complexity levels where they make sense so it can just be a slider.

Thus, Adaptive Rounding is off for Complexity 0-1, and on for 2-8, since it's always beneficial, but just a little slower. So, off for when the user wants to go super-fast, on when they're optimizing more for qualtiy.

[clsid]

Testing is so 2008. Formal verification is way cooler.

[benwaggoner]

Originally Posted by clsid

Testing is so 2008. Formal verification is way cooler.

And once we have a formal metric for video quality, we'll be good to go !

[Zanthra]

Edit: I apologize for my idiocy. I got here from a link form elsewhere, and did not check the last post date/time.

[Revgen]

Originally Posted by benwaggoner

Can we at least agree that there's a real reason for this? You're saying excuse like codec pack and multiple decoder issues aren't an actual problem, but there's plenty of evidence here at forums.doom9.org that the contrary is true . And the people who can make it as far as here to ask for help with a problem are going to be the top portion of the audience who have tried.

And every newbie that comes in to complain about codec packs is told to get VLC player or a version of mplayer if they are having issues with codec packs or codec conflicts.

Of course Microsoft isn't going to like that, so perhaps you'd be better served creating WMP lite and WMP advanced. If people decide to use WMP Advanced, then give them a warning like "Warning: this is an ADVANCED VIDEO PLAYER meant for professionals and enthusiasts. We can't ensure that it will work the way you want it to" or something of that nature. Your OS does this every time I install a non-WHQL driver. Why not do the same for WMP?

The excuses I'm seeing from your end don't seem to add up.

[ranpha]

Originally Posted by Revgen

And every newbie that comes in to complain about codec packs is told to get VLC player or a version of mplayer if they are having issues with codec packs or codec conflicts.

Of course Microsoft isn't going to like that, so perhaps you'd be better served creating WMP lite and WMP advanced. If people decide to use WMP Advanced, then give them a warning like "Warning: this is an ADVANCED VIDEO PLAYER meant for professionals and enthusiasts. We can't ensure that it will work the way you want it to" or something of that nature. Your OS does this every time I install a non-WHQL driver. Why not do the same for WMP?

The excuses I'm seeing from your end don't seem to add up.

Uhm... no, if a newbie complains about codec pack problems, he/she will usually be told to just uninstall that codec pack and then just install the Haali+MPC-HC+ffdshow combo or get a better pack such as CCCP.

And what is the difference between your proposed WMP Lite and WMP Advanced anyway? And what make you think that your WMP Advanced (application) is the same as an unsigned device driver?