madVR - high quality video renderer (GPU assisted)

[mzso]

Hi Madshi,

i noticed a "strange" problem on my new media box, during playback in windowed/fullscreen windowed mode and even during pausing a file the CPU/GPU is still active at around 35%/20%?
I did some tests and even reinstalled windows and found one combination of setting that "fixes" this behavior, resulting in 0% CPU/GPU usage. Also the CPU usage in general is higher in all the "problematic" cases.


I too noticed high cpu/gpu usage when I paused and minimized the player and left it for a while.

[n3w813]

Got a question regarding the levels setting....

I have a HTPC running a Nvidia video card connected to a Sony LCD via HDMI. My media are mostly blu-ray rips. Right now these are my settings on the vid card, madvr, and Sony TV.....

Nvidia - 0-255 desktop output with custom resolution hack
MadVR - levels set to '16-235'
Sony TV - HDMI setting set to 'Limited'

My question is, should I set my TV to 'Full Range' and MadVR to '0-255' instead? What settings do I need to set to avoid any kind of levels compression or expansion in the whole video chain?

Thanks! :D

[markanini]

Question regarding the 'color & gamma' tab. Assuming my monitor more or less corresponds to sRGB, does 'changing pure power curve' to 'BT.709/601' make output more accurate/like intended?

Trying to answer my own question:
After reading the .ti3-Parser thread it appears my display is likely 'pure power curve' to start with so I'll keep the default setting only changing the 2.2 value to 2.35 or so when lights are dimmed. :)

Still makes me wonder what kind of displays are 'BT.709/601' or who would calibrate their display to that standard to warrant such a setting in MadVR. A forum search provided no answers on this one :/

[Xaurus]

Got a question regarding the levels setting....

I have a HTPC running a Nvidia video card connected to a Sony LCD via HDMI. My media are mostly blu-ray rips. Right now these are my settings on the vid card, madvr, and Sony TV.....

Nvidia - 0-255 desktop output with custom resolution hack
MadVR - levels set to '16-235'
Sony TV - HDMI setting set to 'Limited'

My question is, should I set my TV to 'Full Range' and MadVR to '0-255' instead? What settings do I need to set to avoid any kind of levels compression or expansion in the whole video chain?

Thanks! :D
I was in your exact shoes not long ago. I did some research and this is what I found:

With HDMI from the Nvidia card it's sending out 16-235 even if you set it to 0-255 in the desktop. Do this:

For those of you that didn't know, for whoever has a HDTV connected with nvidia drivers,
and the entire desktop is being output in 16-235 instead of full range, the best workaround is to
add this to the nv_disp.inf before installing the drivers. Makes the drivers assume your using a fullrange TV
instead of limited range. I need to use this each time I do a clean install of nvidias drivers.

Add

HKR,,SetDefaultFullRGBRangeOnHDMI,%REG_DWORD%,1

to each section of your inf that begins with

[nv_miscBase_addreg_

eg

[nv_miscBase_addreg__01]
HKR,,SetDefaultFullRGBRangeOnHDMI,%REG_DWORD%,1

Then you can set your TV to full range, as well as madvr.

[nand chan]

Question regarding the 'color & gamma' tab. Assuming my monitor more or less corresponds to sRGB, does 'changing pure power curve' to 'BT.709/601' make output more accurate/like intended?

Pure power curve should be much better for sRGB displays than a BT.709 curve. I don't really recommend using the BT.709 curve for display purposes at all - it's a virtual, encoding curve - no monitors actually use this, all of them are either something like CRT (~2.35-2.55 PPC-ish) or sRGB, which is close to 2.2 PPC.

I recommend using a PPC. If you're in a dark viewing environment, I'd actually recommend a PPC of about 2.5 - this corresponds to what the average CRT has.

In effect, you simulate the appearance of a CRT display (~2.5) on your sRGB display (~2.2) - images are “under-encoded” using (1/2.2) instead of (1/2.5) on purpose, because the specifications themselves assume a dark viewing environment - by using an encoding coefficient /below/ the viewing coefficient, the brightness is reduced, which compensates for the lack of contrast in low-light environments.

Trying to answer my own question:
After reading the .ti3-Parser thread it appears my display is likely 'pure power curve' to start with so I'll keep the default setting only changing the 2.2 value to 2.35 or so when lights are dimmed. :)

Still makes me wonder what kind of displays are 'BT.709/601' or who would calibrate their display to that standard to warrant such a setting in MadVR. A forum search provided no answers on this one :/

Yes, it'll be PPC or similar. I'd personally recommend generally “darker” values, something like 2.3 for bright rooms and 2.5 for dark rooms.

Like mentioned, the BT.709 curve is only used to *encode* video - the point of them is to give you the maximum amount of information per bit - the same reason why sRGB turns into a flat correlation near the black level. It's used for HDTV broadcasting.

The actual EBU Recommendations 709 document hints that displays have a convention of about 2.35 PPC or 2.2 sRGB.

I don't know if *any* displays do actually use this BT.709 curve directly. Usually you decode the source Y'CbCr to rgb', decode that to RGB using the BT.709 curve or whatever, then encode that RGB to rgb' using the sRGB curve (or PPC 2.2).

madVR doesn't differentiate between PPC 2.2 and sRGB - if you enable PPC 2.2 in gamma correction, the image doesn't change at all.

[Andy o]

With HDMI from the Nvidia card it's sending out 16-235 even if you set it to 0-255 in the desktop.

Hmm, I've read this before. It's not happening with my Pioneer display though, and GTX460 and 285 drivers. I get Full RGB out of the box. AMD owners also had problems like this, and us Pioneer users (probably other displays too, but I only remember a couple other persons for sure and they also had a Pio) also were in what I think is a minority that didn't have the problem. Are most people having this problem here with NV?

[6233638]

I recommend using a PPC. If you're in a dark viewing environment, I'd actually recommend a PPC of about 2.5 - this corresponds to what the average CRT has.Reference monitors that content is created on, are 2.40 gamma.

Everything else you say is more-or-less correct. No display should be running the 709 curve, it is for camera encoding, not playback.

[kirakami]

will madvr work on my old graphic card
Nvidia Ge-Force 4 MX440 64MB AGP 8x with MPC-HC?

[fastplayer]

No, it won't.

[deado]

Hi guys - I need some help here, I have been desperately trying to get smooth playback of 23.976/24fps content on my PC (output to 24hz HDTV) - typically 1080p MKVs. Specs are Windows 7, Nvidia GTX 480, i5 2500k overclocked, so PC is more than capable. Using ffdshow software video decoding, madvr, mpc-hc etc so PC hardly breaks a sweat.

I have found that madVR in MPC-HC using exclusive fullscreen mode gives overall the smoothest playback (and trust me, I've tried everything - all combinations of players and renderers). But it still is not perfect. It still visually skips every 20-30 seconds or so - even though madVR with Ctrl + J info shows NO dropped frames! Sometimes it takes longer to happen though - maybe 1-2 minutes... but it's visibly there for sure - watching credits is a good test. Reclock enabled or disabled seems to make little difference. My graphics card has two options for output - 23hz (which madVR reports as ~23.974hz, and 24hz, which madVR reports as 23.999998hz. Which I choose also makes little difference - both have the similar stutter. Best way to describe it is it 'jumps' for a split second.

I seem to be at a dead end as to what is causing it. Any suggestions?

[Andy o]

Have you enabled VSync on ReClock? If so, disable it. It might also be a good idea to "clean-up video clock timings database" on ReClock's advanced settings.

[nand chan]

Reference monitors that content is created on, are 2.40 gamma.

Everything else you say is more-or-less correct. No display should be running the 709 curve, it is for camera encoding, not playback.

Okay good to know. I'll use and recommend 2.40 gamma when viewing in a properly setup reference environment. (D50, 64 lux ambient, 80 cd/m² illuminant)

[deado]

Have you enabled VSync on ReClock? If so, disable it. It might also be a good idea to "clean-up video clock timings database" on ReClock's advanced settings.

That seems to have done the trick, :thanks: :D

It's crazy though that it takes such specific settings just to get smooth playback. It should not be so difficult to achieve!! It basically means anyone who doesn't know exactly what to do is almost guaranteed to get stuttering of some sort. Especially if they are using other more standard renderers and if Aero is enabled - if using Nvidia and newer drivers it seems. Not sure about ATI. Thank god for madVR and the fullscreen exclusive mode. madshi, you rule :cool:

Default settings in players (ie. EVR using WMP12) have also had much worse stuttering ever since nvidia driver 258.96. I made a thread over at the nvidia forums here (http://forums.nvidia.com/index.php?showtopic=188990&st=0). I had stuck with driver 258.96 for a long time but now using this new solution with mpc-hc, madvr etc I can use the newest 285 driver :)

[Andy o]

What "specific settings" are you referring to? The "clean-up video timings" thing you don't normally have to do, but sometimes it just gets corrupted.

[deado]

What "specific settings" are you referring to? The "clean-up video timings" thing you don't normally have to do, but sometimes it just gets corrupted.

I just mean generally - ie. Vsync settings.

And having to use madVR in exclusive mode just to get smooth playback- ie. no other renderer for me gives perfectly smooth playback under any circumstances when doing 24hz output - they all stutter somehow or another. I think this is the combination of crappy Nvidia drivers lately and the complications of Aero and GPU usage.

From what I have tested on my setup, any other renderer (EVR, VMR9, Overlay, Haali, whatever) gives stuttering of some sort - and I have been testing this for months and months trying different settings. Nothing is perfect except this madVR solution in fullscreen exclusive mode with Reclock Vsync disabled like you suggested. From what I've found, it seems to be Aero on = stuttering regardless. So madVR's exclusive mode with Aero off (and no tearing) gives the best results.

It's just a sad state of affairs that to get smooth playback from a PC output is so difficult in the first place! And that 99% of combinations of players/renderers/settings cause stuttering. At least for me anyway - but I assume it is similar for others? At least those with Nvidia cards - since I think that is where a lot of the source of problems come from. Maybe it's just my particular set up, dunno, but I'm glad I've finally got smooth playback :)

[Andy o]

The only generalized problem I've seen with Nvidia here in the madVR thread is probably the glitches thing, but that's fixable with madVR's anti glitch options.

[kalston]

The only generalized problem I've seen with Nvidia here in the madVR thread is probably the glitches thing, but that's fixable with madVR's anti glitch options.

Not anymore for me with the 280.xx drivers. I don't have a DX11 capable GPU though. Even with all the options enabled I still get a lot of glitches.

I'm using the old exclusive mode now and it seems to work better, at least I don't notice anything wrong with it.

[TheProfosist]

Im currently having a issue where when i playback on my second monitor using D3D11 for presentation and i enter exclusive mode MPC-HC crashes with the below error:
Problem signature:
Problem Event Name: BEX
Application Name: mpc-hc.exe
Application Version: 1.5.2.3456
Application Timestamp: 4e29d332
Fault Module Name: madVR.ax
Fault Module Version: 0.74.0.0
Fault Module Timestamp: 4e516bc2
Exception Offset: 0003231a
Exception Code: c000000d
Exception Data: 00000000
OS Version: 6.1.7601.2.1.0.256.1
Locale ID: 1033
Additional Information 1: 0a68
Additional Information 2: 0a6884bbdf5111fa2279f2535a7cd566
Additional Information 3: 749e
Additional Information 4: 749e8f106870e5fcdbc114acbc3c00d2

the other problem that i have is that unless i enter exclusive mode there is a blinking box in the upper right hand corner of the display (which is rather annoying).

I am also still having this problem on my laptop. when i try to play a fullscreen video with madVR and MPC-HC. After a few seconds i it crashes and i get this error from windows:

Problem signature:
Problem Event Name: BEX
Application Name: mpc-hc.exe
Application Version: 1.5.2.3456
Application Timestamp: 4e29d332
Fault Module Name: madVR.ax
Fault Module Version: 0.74.0.0
Fault Module Timestamp: 4e516bc2
Exception Offset: 0003231a
Exception Code: c000000d
Exception Data: 00000000
OS Version: 6.1.7601.2.1.0.256.1
Locale ID: 1033
Additional Information 1: cfd3
Additional Information 2: cfd39af6f2554026609aee2812317e49
Additional Information 3: ba1d
Additional Information 4: ba1d1af66b681e7b15d51afc345f34da

[Andy o]

Not anymore for me with the 280.xx drivers. I don't have a DX11 capable GPU though. Even with all the options enabled I still get a lot of glitches.

I'm using the old exclusive mode now and it seems to work better, at least I don't notice anything wrong with it.

I'm using the 285 drivers and GTX 460. Glitches are fixed for me when enabling the 2nd and 3rd options. The 1st (separate thread) gets me stuttering after skipping sometimes, but either enabled or disabled, it's the same for glitches.

[kalston]

I'm using the 285 drivers and GTX 460. Glitches are fixed for me when enabling the 2nd and 3rd options. The 1st (separate thread) gets me stuttering after skipping sometimes, but either enabled or disabled, it's the same for glitches.

But did you tick "use a separate device for presentation" and "use D3D11 for presentation"? Madshi said he needed these for smooth playback, as long as the GPU is DX11 capable.

[nevcairiel]

DX11 fixes all glitch issues for me, however the DX11 mode is kinda annoying, because it has some bugs with redrawing the desktop when you exit the full screen mode.
I now run without either DX11 or "use separate device", i only tick the 2nd and 3rd glitch tweak, and it seems ok.

[madhatter300871]

I was in your exact shoes not long ago. I did some research and this is what I found:

With HDMI from the Nvidia card it's sending out 16-235 even if you set it to 0-255 in the desktop. Do this:


I am using madVR, Nvidia 8400GS, latest drivers (as of a few weeks ago). Nvidia outputs 0-255 for the desktop, and allows the media player to control output for video.

In the madVR settings I have selected that my monitor expects 0-255. I have then set my projector to accept 0-255 input. For me I don't have a problem, if I change madVR to output 16-235 I see a noticeable difference as would be expected.

Regarding using a pure power curve as opposed to BT.709/601 curve .... should I be using a pure power curve ? I did 'play' with this setting but in truth was not really sure what I should do for the best. Is my Z4 an sRGB display ? I have set the PC to use Adobe RGB colour profile for my projector.

[nand chan]

Regarding using a pure power curve as opposed to BT.709/601 curve .... should I be using a pure power curve?

Yes, most likely.

I did 'play' with this setting but in truth was not really sure what I should do for the best. Is my Z4 an sRGB display?

The best way to demonstrate is to register the madTestPatternGenerator (located in madVR's folder) and run the smallramp.ytp sample. If you set it to PPC (at which level it's exactly linear depends on the gamma curve of your projector, you'll have to experiment).

Then, observe the values near the black point. For PPC, it should be a relatively smooth drop-off to black.

For the BT.709 curve, you'll notice that it sort of just drops off at a certain point. (It's best to test this at a lower gamma setting like 2.2 (instead of 2.4/2.5), and the effect becomes more noticeable as you go lower, eg. 1.8)

I have set the PC to use Adobe RGB colour profile for my projector.

You need to be very, very careful here - madVR doesn't use or respect your windows color profile settings at all. It's only capable of using a .3dlut for color management.

Fortunately, you can use the built-in option for the yCMS mode to have it generate this automatically.

Just choose “calibrate this display by using yCMS” and enter the following in the top field:

red, Yxy, 1.0, 0.6400, 0.3300
green, Yxy, 1.0, 0.2100, 0.7100
blue, Yxy, 1.0, 0.1500, 0.0600
white, Yxy, 1.0, 0.31271, 0.32902

[kalston]

DX11 fixes all glitch issues for me, however the DX11 mode is kinda annoying, because it has some bugs with redrawing the desktop when you exit the full screen mode.
I now run without either DX11 or "use separate device", i only tick the 2nd and 3rd glitch tweak, and it seems ok.

Sadly it's not enough for me on my GTX275.
If the next nVidia drivers (I don't like using beta drivers) don't fix it I'll probably try going back to 260.xx drivers or something.

[madhatter300871]

Well, I discovered something new tonight. madVR bypasses the Nvidia video brightness/contrast/hue etc so when I slide them in the nvidia control panel nothing happens. OK, no big deal because I never used them anyway. However, the nvidia brightness/contrast/gamma/hue/digital vibrance settings for the desktop DO affect the rendered video. I never realised this, I had them set to some arbitrary value, my desktop looked OK and I wasn't really that bothered.

I have now tweaked them a little and made a vast improvement to the video, mainly lowered the contrast a touch to get much more detail in dark areas, and upped the digital vibrance a tiny amount which gives, to my eye at least, a more vibrant picture.

In response to 'nand chan' I played with the yCMS calibration values and to be honest I couldn't see a difference, so I am running with not calibrated mode for now. To be honest, I have no idea what so ever how all those calculations work, I have never read up on it and find myself being very lazy indeed in doing so. I'm prepared to accept that it's beyond me for the time being.

[madhatter300871]

On a different note ... I pressed control-j to get the debug info displayed and I am getting a few dropped frames when the video starts (which I have read is normal) and 1 frame duplicated every 20 minutes or so. Is this normal ?

My projector accepts 50Hz and 60Hz input.

With 23.976, 24 and 25 fps source, with projector/nvidia card set to 60Hz I get 1 frame repeat about every 20 minutes or so.

With 23.976, 24 and 25 fps source, with projector/nvidia card set to 50Hz I get 1 frame repeat about every 1 second or so and hundreds of 'render glitches'.

I can't use reclock because I send audio direct to my amp, which reclock will not let me do (I dont think). I was expecting the 25fps video to play perfectly at 50Hz, but not so !

Do my figures look normal ? Again, madVR is new to me so I am still getting familiar with it.

[nand chan]

In response to 'nand chan' I played with the yCMS calibration values and to be honest I couldn't see a difference, so I am running with not calibrated mode for now. To be honest, I have no idea what so ever how all those calculations work, I have never read up on it and find myself being very lazy indeed in doing so. I'm prepared to accept that it's beyond me for the time being.

Well, first of all, it's important to know what the exact chromatic signature of your projector looks like, then you can adjust video to whatever to get accurate colors.

To demonstrate the difference, it's probably best to display something very green (I love to use the grass scene at the start of Madoka episode 1, shortly after the OP) - our eyes are most sensitive to the green receptors so the difference is more obvious, compared to something like blue which we basically can't detect very well at all.

I can't use reclock because I send audio direct to my amp, which reclock will not let me do (I dont think)..

The only thing ReClock can't do is stream to ASIO, but if you use Windows 7, it can use WASAPI (and if Windows XP, kernel streaming), to directly stream PCM / DTS / AC3 / whatever to your receiver / amp / sound card / HDMI connection, bypassing the entire Windows audio stack etc.

You need to use an audio filter that can bitstream the audio (like ffdshow audio), and enable the streaming options in ReClock + set the correct output device.

[Mikey2]

I noticed that once in a while the following lines fleetingly appears in the OSD:

Composition Rate: 23.976Hz
Clock Deviation: -.1xxx%

(where xxx is presumably the percentage difference between the actual display frequency [the top line of the OSD] and the composition rate (?) ])

Why does this appear and disappear at seemingly random intervals? (While it often displays when I do something like turn subtitles on or off or switch between exclusive and windowed mode. this does not need to be the case.)

I am especially confused since my display value is never exactly 23.976Hz.

I should note that I am using ReClock - with it I see a "1 frame drop" on the order of hours, if not days. In addition, from what I understand, the other values appear solid (once reset, I do not see any frame-drops nor delays and my queues are all full...)

Is this simply some sort of logic written into the OSD to occasionally show these values, or is there something improperly configured in my system that is still causing sync/frame-rate issues?

Thanks much! Once again I love this renderer!! :)
MikeY

[madhatter300871]

The only thing ReClock can't do is stream to ASIO, but if you use Windows 7, it can use WASAPI (and if Windows XP, kernel streaming), to directly stream PCM / DTS / AC3 / whatever to your receiver / amp / sound card / HDMI connection, bypassing the entire Windows audio stack etc.

You need to use an audio filter that can bitstream the audio (like ffdshow audio), and enable the streaming options in ReClock + set the correct output device.

I use windows XP, and I use ffdshow. OK, I'll set re-clock to use kernel streaming, I have already set ffdshow to send DTS and AC3 direct to the amp .... all I got was noise ..... I'll try again tonight.

However, if I do this am i not defeating the object. The video will play at a different rate but the audio will not ?

Like Mikey2, I also get 1 frame drop / frame repeat in the order of about 1 every 20 minutes or so, I have seen it go up to an hour or so.

[mzso]

Hi!
I had my screen rotated and I wanted to check out something in on of the videos. But the player hung. After that I tried with other players and madVR hung all. (potplayer, kmplayer, mpc-hc)
All were fine with EVR.

[Xaurus]

I do like AMD myself, due to good performance per watt, due to *not* having those presentation glitch problems and due to a freely available frame packed HDMI 3D output SDK. Unfortunately neither NVidia nor AMD are really perfect at this time. I wish it were possible to combine the best of NVidia and AMD into one new HTPC graphics card.
Hi madshi, been 1 month since your last post in your thread here... hope all is well. :)

[SamuriHL]

Hi madshi, been 1 month since your last post in your thread here... hope all is well. :)

He's on vacation...at least from madVR development. He'll be back. :)

[pirlouy]

I remember the 0.10 version, for one year without Madshi news. Everybody became crazy, starting to ask for Open Source, and insulting Madshi. Was fun. :p

[JarrettH]

Yep, expected. It's nice that when he takes a break it's still at a fairly stable state :thanks:

[nand chan]

I remember the 0.10 version, for one year without Madshi news. Everybody became crazy, starting to ask for Open Source, and insulting Madshi. Was fun. :p

I really do wish madVR was made open source, though. :(

I'd really love to play around with the renderer, unfortunately there's no good open source renderer out there.

Either that, or let me customize the pixel shader code madVR uses.

[leeperry]

I really do wish madVR was made open source
madshi has put more sweat and energy in his VR than anyone ever did before....it's funny how every MPC/ffdshow update ends up in KMP/PotPlayer on a regular basis eventually ;)

[nand chan]

madshi has put more sweat and energy in his VR than anyone ever did before....it's funny how every MPC/ffdshow update ends up in KMP/PotPlayer on a regular basis eventually ;)

And? Who cares what knock-off products do? The important bit is that the main product gets continually improved. And after all, who's to say anything against a bit of choice?

The only motive I can see for keeping something closed source is if you want to sell it eventually.

[andyvt]

it's funny how every MPC/ffdshow update ends up in KMP/PotPlayer on a regular basis eventually ;)

That's the point of FOSS. Free is free.


The only motive I can see for keeping something closed source is if you want to sell it eventually.

There are other considerations, to name a few:

- use of proprietary/NDA API - some vendors won't give the necessary NDAs to OSS
- project control
- use of non-FOSS code/libraries

There is a good case for either side, but ultimately it's not our business. The developer's wishes and their willingness to share should be respected. After all, given the choice b/w nothing (i.e. no sharing) and something (binary) the benefit seems obvious; why not just say "thank you" and leave it alone?

[SamuriHL]

madshi has said in the past that once he gets to version 1 he's going to decide whether or not to charge for it. It's also the reason he's not taking donations right now...he simply doesn't know yet what he wants to do. Honestly, if he did charge for it, I'd pay for it cause quite frankly, it's the best renderer out there IMO. For now, let's enjoy what we're given and wait to see what plans he has for the future.

[nevcairiel]

That's the point of FOSS. Free is free.

Actually, the point of FOSS is not that its free, the point of FOSS is that when you take, you also have to give back, which is something those korean knock-offs have not learned. :p

[andyvt]

Actually, the point of FOSS is not that its free, the point of FOSS is that when you take, you also have to give back, which is something those korean knock-offs have not learned. :p

We'll have to disagree (http://www.gnu.org/philosophy/free-sw.html) on that.

[nevcairiel]

We'll have to disagree (http://www.gnu.org/philosophy/free-sw.html) on that.

That text specifically states that when you take free software and modify it, the result should also be free software (specifically everything about copyleft)
What good is FOSS when there is one FOSS product, but everyone uses it to create their closed things, and no other FOSS emerges.

[nand chan]

That text specifically states that when you take free software and modify it, the result should also be free software (specifically everything about copyleft)
What good is FOSS when there is one FOSS product, but everyone uses it to create their closed things, and no other FOSS emerges.

I don't actually consider the GPL to be “free”. For me, truly free software includes BSD and derivatives (eg. CPOL). I think the GPL is too restrictive.

Also, the usage of proprietary APIs can be fixed by including binary blobs. I'm not one of those purists that thinks everything should be 100% free and GPL, I just care about being able to tinker with software when I want to, eg. the shaders / transformation code.

[andyvt]

That text specifically states that when you take free software and modify it, the result should also be free software (specifically everything about copyleft)
What good is FOSS when there is one FOSS product, but everyone uses it to create their closed things, and no other FOSS emerges.

Oh, do they modify the filters and not disclose source? If so, forget what I said earlier about them specifically.

For the other part, I do think there is value in FOSS integration in closed source projects; whether you can [legally] do that is controlled by the developers intent communicated via the license (some do, some don't).

[nevcairiel]

The GPL may not be the most "free", but i would at least like recognition of my work when its used in another project. Just taking it, improving some aspects, and claiming how you're better is not a good system. :)
If someone approached me and asked if he could use my stuff in his project which is incompatible with the GPL, i would give them a relicensed version - of course its easy for me, everything being by one developer. :p

[SamuriHL]

My company has an entire legal department devoted to these issues. We have a whole procedure to follow if we want to use ANY kind of open source in our products, and I can tell you, the license makes ALL the difference on that one.

In any case, madshi will ultimately decide what he wants to do with his product. Clearly he's chosen to keep it free but closed source for the time being. As the developer, that's absolutely his choice to make. ;)

[andyvt]

The GPL may not be the most "free", but i would at least like recognition of my work when its used in another project. Just taking it, improving some aspects, and claiming how you're better is not a good system. :)

I can agree with that. The GPL and other viral OSS licenses are a good way (in theory anyway) to reduce the risk of appropriation without attribution in commercial SW.

[nand chan]

I can agree with that. The GPL and other viral OSS licenses are a good way (in theory anyway) to reduce the risk of appropriation without attribution in commercial SW.

You can easily write up a license that forces you to include the original author's name and license in derivative works (but otherwise offers no restrictions based on open source-ness).

[TheShadowRunner]

Little bug:
When using On2 VP7 VP70 video decoder (http://videoff7.free.fr/vp7dec.ax) along with madVR, the video (http://samples.mplayerhq.hu/V-codecs/VP7/) is upside down.
When using any other renderer, EVR, VMR9 etc, video is OK.

[nehcrow]

My madVR is stuttering on my XPS 15 laptop.
It only stutters in D3D Fullscreen but it's fine in windowed mode.
My specs are:
i7 2630QM processor
Nvidia GT525M

Help?