processing / p5.js-video

177 stars 38 forks source link

Greenscreen visible in Chrome? #40

Closed GanWeaving closed 8 years ago

GanWeaving commented 10 years ago

capture

GanWeaving commented 10 years ago

It works fine in my Firefox browser but not in Chrome.

shiffman commented 10 years ago

OS / browser version?

GanWeaving commented 10 years ago

Windows 7 Chrome 38 Firefox 31

scottgarner commented 10 years ago

Thanks, @Xeronimo74. I tested with Chrome 36 on Windows 7, but not 38. I'll grab it and give it a shot.

Any errors in the developer console?

scottgarner commented 10 years ago

Hmm. I tested Chrome (36), Chrome Beta (37) and Chrome Canary (38) in Windows 7 without any problems. Would you mind taking a look the demo on the homepage of http://seriouslyjs.org to see if that works?

GanWeaving commented 10 years ago

Hi, yes, seriouslyjs seems to work correctly. Still the same green screen problem with hello.p5js.org though ... The console gives me these errors: 18 : CS -> BG : FAILED closepopuptoplevel chrome-extension://hdokiejnpimakedhajhdlcegeplioahd/onloadwff.js:77 21 : CS -> BG : FAILED closepopuptoplevel chrome-extension://hdokiejnpimakedhajhdlcegeplioahd/onloadwff.js:77 Font from origin 'http://fonts.gstatic.com' has been blocked from loading by Cross-Origin Resource Sharing policy: No 'Access-Control-Allow-Origin' header is present on the requested resource. Origin 'http://hello.p5js.org' is therefore not allowed access. The response had HTTP status code 403. hello.p5js.org/:1 Failed to load resource: the server responded with a status of 403 (Forbidden: header 'Content-Type' value denied) http://fonts.gstatic.com/s/inconsolata/v8/BjAYBlHtW3CJxDcjzrnZCI4P5ICox8Kq3LLUNMylGO4.woff2 Font from origin 'http://fonts.gstatic.com' has been blocked from loading by Cross-Origin Resource Sharing policy: No 'Access-Control-Allow-Origin' header is present on the requested resource. Origin 'http://hello.p5js.org' is therefore not allowed access. The response had HTTP status code 403. hello.p5js.org/:1 Failed to load resource: the server responded with a status of 403 (Forbidden: header 'Content-Type' value denied) http://fonts.gstatic.com/s/inconsolata/v8/BjAYBlHtW3CJxDcjzrnZCIbN6UDyHWBl620a-IRfuBk.woff

scottgarner commented 10 years ago

Thanks for testing, @Xeronimo74.

It looks like your LastPass extension is throwing some errors, but i don't see how that could affect the keying. Those font security errors are unusual, too, but also seem unrelated. Are you running any other exertions that might be causing problems?

Also, would you mind trying out this simplified chroma test: http://hello.p5js.org/test/chroma.html

Just hit play and see if playing with the sliders affects the key. Thanks!

Jdakotapowell commented 10 years ago

Seeing a green screen with the Chrome browser.

scottgarner commented 10 years ago

Hi @Jdakotapowell. Can you provide your browser and OS version along with any errors in the console?

Jdakotapowell commented 10 years ago

Chrome - Windows Vista - as per console errors - can get bk to you on Sat - on a deadline & sweating it out!

On Thu, Aug 7, 2014 at 4:55 PM, Scott Garner notifications@github.com wrote:

Hi @Jdakotapowell https://github.com/Jdakotapowell. Can you provide your browser and OS version along with any errors in the console?

— Reply to this email directly or view it on GitHub https://github.com/scottgarner/p5.js-video/issues/40#issuecomment-51491771 .

scottgarner commented 10 years ago

@Jdakotapowell Thanks for the feedback and take your time.

nolanbjohnson commented 10 years ago

Chrome (Version 36.0.1985.125 m) most up to date -- Windows 8. Don't see any console errors. Tried the chroma test above as well, but sliders have no effect.

scottgarner commented 10 years ago

Thanks for testing, @nolanbjohnson. Would you mind trying the demo at http://seriouslyjs.org to see if that works?

nolanbjohnson commented 10 years ago

the demo does work for me.

scottgarner commented 10 years ago

So strange! @brianchirls, mind if I rope you in here?

nolanbjohnson commented 10 years ago

Correction...I was not getting the full effect on the demo. The buttons applied a filter but did not completely replace the green screen. Just watched it on my mac and saw the difference. Sorry for the confusion.

brianchirls commented 10 years ago

This is a really tough one. Works fine on my Windows machine. And there are no obvious errors in the console.

So let's see if we can determine if it's Seriously.js in general or just the chroma key effect. @nolanbjohnson, @Xeronimo74 and @Jdakotapowell can you try out this page and let us know what you see? http://brianchirls.github.io/Seriously.js/examples/animate/

There should be a canvas with an animated wipe effect going back and forth between some colored pencils and a toy robot with a purple pinwheel. Also, please check the console and let us know if anything shows up there.

And if you have the time, maybe you could list which browser extensions you have installed? We could see if there's anything you all have in common.

Thanks for your debugging help, everyone.

GanWeaving commented 10 years ago

Hi, I see that swipe effect. Plus a fine line of white dots at times (screenshot: http://i.imgur.com/3L8XYxT.jpg) ... No errors in the console.

As for the browser extensions: http://i.imgur.com/jwgju64.jpg , http://i.imgur.com/sJYl60h.jpg

GanWeaving commented 10 years ago

As for the Chroma test page, I'm seeing a green screen too. And the sliders don't seem to have an effect.

brianchirls commented 10 years ago

Ah those white dots are definitely not supposed to be there. Let me see if I can fix that...

I'm not too hopeful that the extensions will prove anything, but let's see what the others post and see if there's any overlap. thanks again.

nolanbjohnson commented 10 years ago

Ditto @Xeronimo74 on the swipe effect and the white dots. Don't see any overlap in extensions. image

brianchirls commented 10 years ago

Okay, I think we can rule out the extensions. I suspect it has something to do with the way your GPU/drivers handle the branching in the shader. Can you guys try out this test for me? It's a modified version of the split effect that I hope will not have the white dots. If that's the case, then I have an idea of how I can fix the chroma effect too.

http://jsbin.com/zecoq/1

For comparison, here's the same thing using the original split effect, which should have white dots. http://jsbin.com/hadej/1

Also, can you tell us what GPU hardware you're using?

Thanks, B

GanWeaving commented 10 years ago

Yes, the second one has the dots while the first one does not.

I've got an integrated graphics card from Intel (this is my work PC) ...

brianchirls commented 10 years ago

Ah, excellent. I'll make an issue for this over at the Seriously.js repo and will see if I can fix Chroma key. May take me a day or two.

scottgarner commented 10 years ago

Thanks for jumping in here, Brian. I really appreciate it.

On Fri, Aug 8, 2014 at 9:41 AM, Brian Chirls notifications@github.com wrote:

Ah, excellent. I'll make an issue for this over at the Seriously.js repo and will see if I can fix Chroma key. May take me a day or two.

— Reply to this email directly or view it on GitHub https://github.com/scottgarner/p5.js-video/issues/40#issuecomment-51610847 .

brianchirls commented 10 years ago

No problem. I want to see these issues solved for Seriously as well. It's looking like a long list of effects that may be affected.

We should probably do a call one of these days to go over a few different things that may affect your video piece, what could be fixed and what could be improved.

brianchirls commented 10 years ago

Here's an example using a modified chroma key effect. Does it work?

http://jsbin.com/dijog/1

nolanbjohnson commented 10 years ago

I don't see a green screen (it's a full white background) but no visuals either. Not sure if it's helpful, but there's a lot of aliasing around the edges.

brianchirls commented 10 years ago

Thanks. Do you hear the audio?

brianchirls commented 10 years ago

Updated to make video controls visible, so you can make sure it's playing. http://jsbin.com/dijog/2

nolanbjohnson commented 10 years ago

Yep, audio is good.

brianchirls commented 10 years ago

@nolanbjohnson just to be thorough, did the modified split effect work for you? i.e. is there the line of white dots here? http://jsbin.com/hadej/1

nolanbjohnson commented 10 years ago

yes, I do see the dots.

darkyen commented 10 years ago

@nolanbjohnson can you tell me your output for chrome://gpu and then windows information for the gpu and driver via dxdiag ??

nolanbjohnson commented 10 years ago

Will do tomorrow morning @darkyen. Don't have my work laptop with me.

nolanbjohnson commented 10 years ago

chrome://gpu output: Graphics Feature Status Canvas: Software only, hardware acceleration unavailable Flash 3D: Software only, hardware acceleration unavailable Flash Stage3D: Software only, hardware acceleration unavailable Flash Stage3D Baseline profile: Software only, hardware acceleration unavailable Compositing: Software only and threaded. Hardware acceleration unavailable. Rasterization: Software only. Hardware acceleration disabled. Video Decode: Software only, hardware acceleration unavailable Video Encode: Software only, hardware acceleration unavailable WebGL: Software only, hardware acceleration unavailable Driver Bug Workarounds clear_uniforms_before_first_program_use disable_d3d11 exit_on_context_lost Problems Detected GPU rasterization is whitelisted on N4, N5, N7 and Moto X: 362779 Disabled Features: gpu_rasterization Some drivers are unable to reset the D3D device in the GPU process sandbox Applied Workarounds: exit_on_context_lost The Mali-Txxx driver does not guarantee flush ordering: 154715, 10068, 269829, 294779, 285292 Applied Workarounds: use_virtualized_gl_contexts TexSubImage2D() is faster for full uploads on ANGLE Applied Workarounds: texsubimage2d_faster_than_teximage2d Clear uniforms before first program use on all platforms: 124764, 349137 Applied Workarounds: clear_uniforms_before_first_program_use Using D3D11 causes browser crashes on certain Intel GPUs: 310808 Applied Workarounds: disable_d3d11 Accelerated rasterization has not been enabled or is not supported by the current system. Disabled Features: rasterization Version Information Data exported 8/12/2014 11:02:34 AM Chrome version Chrome/36.0.1985.125 Operating system Windows NT 6.3 Software rendering list version 8.7 Driver bug list version 6.1 ANGLE commit id d3d870647655 2D graphics backend Skia Command Line Args Files (x86)\Google\Chrome\Application\chrome.exe" --flag-switches-begin --flag-switches-end --disable-gpu --high-dpi-support=0 Performance Information Graphics 5.2 Gaming 5.2 Overall 5.2 Driver Information Initialization time 0 Sandboxed false GPU0 VENDOR = 0x8086, DEVICE= 0x0416 Optimus false AMD switchable false Desktop compositing Aero Glass Driver vendor Intel Corporation Driver version 10.18.10.3496 Driver date 3-11-2014 Pixel shader version
Vertex shader version
Machine model name
Machine model version
GL_VENDOR
GL_RENDERER GL_VERSION
GL_EXTENSIONS
Window system binding vendor
Window system binding version
Window system binding extensions
Direct rendering Yes Reset notification strategy 0x0000 Diagnostics ... loading ... Log Messages [8000:3720:0808/155921:ERROR:gles2_cmd_decoder.cc(6078)] : [.WebGLRenderingContext]RENDER WARNING: texture bound to texture unit 0 is not renderable. It maybe non-power-of-2 and have incompatible texture filtering or is not 'texture complete' [8000:3720:0808/164104:ERROR:gles2_cmd_decoder.cc(6078)] : [.WebGLRenderingContext]RENDER WARNING: texture bound to texture unit 0 is not renderable. It maybe non-power-of-2 and have incompatible texture filtering or is not 'texture complete' GpuProcessHostUIShim: The GPU process exited normally. Everything is okay.

nolanbjohnson commented 10 years ago

@darkyen and windows: image

brianchirls commented 10 years ago

@Xeronimo74 Can you tell me if the modified Chroma effect works?

http://jsbin.com/dijog/1

brianchirls commented 10 years ago

@nolanbjohnson: Thanks for posting that info. I talked to @darkyen and he wasn't able to replicate your problem. We're running out of options here...

The last thing I can think of is to try running the WebGL Conformance Tests. If you can paste the results here, I can try to see if it's picking up any quirks in your particular implementation. It will take a few minutes to run. When it's done, you can click "display text summary" and paste in the list of failures.

Chrome on my machine doesn't quite pass ALL of the tests, but maybe I can see if yours fails different ones from mine.

brianchirls commented 10 years ago

@nolanbjohnson I notice you're running Chrome with the --disable-gpu option. Did you set that up yourself? How come?

Maybe if you try running Chrome without that, you'll see different results. Or maybe it will crash. (That would count as "different," I think.)

GanWeaving commented 10 years ago

@brianchirls yes it works, white background

brianchirls commented 10 years ago

Great! Thanks, @Xeronimo74. So I'll fix this under the ticket over at my repo, which you are welcome to follow and comment on. I will also take a stab at refactoring other plugins to hopefully avoid the same problem.

GanWeaving commented 10 years ago

ok, cool!

brianchirls commented 10 years ago

@scottgarner The chroma effect is updated on the develop channel. Feel free to try it out. It should work the same for most people and solve the problem for @Xeronimo74.

https://github.com/brianchirls/Seriously.js/blob/develop/effects/seriously.chroma.js

nolanbjohnson commented 10 years ago

@brianchirls, when I run the WebGL test it goes fine for a bit and then everything starts failing.

This seems to be the recurring error msg: FAIL Unable to fetch WebGL rendering context for Canvas failed: context does not exist.

nolanbjohnson commented 10 years ago

Not sure why --disable-gpu option is there, @brianchirls. Though, when I try http://jsbin.com/dijog/1 it also works for me--white background.

scottgarner commented 10 years ago

Everything looks good so far on my end with the chroma effect from develop. Thanks, @brianchirls.