[x3d-public] Problem with inverted polygons/normals: actually texture transparency

Sat Nov 11 10:07:27 PST 2017

Sorry Andreas, am looking but still not understanding whether inconsistent rendering (for image textures that contain alpha channel information) is a specification issue or not.  Patrick indicated (below) that the specification is ambiguous, i.e. "none of the browsers is wrong or right."

Either the specification needs to be more precise, or else some browser implementations are erroneous.

Opinions by X3D browser implementers please, which is it?

17.2.2 Lighting model
http://www.web3d.org/documents/specifications/19775-1/V3.3/Part01/components/lighting.html#Lightingmodel

18.2.1 Texture map formats
http://www.web3d.org/documents/specifications/19775-1/V3.3/Part01/components/texturing.html#TextureMapFormats

18.2.2 Texture map image formats
http://www.web3d.org/documents/specifications/19775-1/V3.3/Part01/components/texturing.html#TextureMapImageFormats

On 11/9/2017 7:06 AM, Andreas Plesch wrote:
>> Date: Wed, 8 Nov 2017 08:04:52 -0800
>> From: Don Brutzman <brutzman at nps.edu>
>> To: 'Patrick D?hne' <pdaehne at gmail.com>, Michalis Kamburelis
>>          <michalis.kambi at gmail.com>, Roy Walmsley <roy.walmsley at ntlworld.com>
>> Cc: 'X3D Graphics public mailing list' <x3d-public at web3d.org>, Mario
>>          Nagamura <mnagamur at lsi.usp.br>, Marcello Zuffo <mkzuffo at usp.br>, Barb
>>          Helfer <blhelfer at gmail.com>, <support at okino.com>
>> Subject: Re: [x3d-public] Problem with inverted polygons/normals:
>>          actually texture transparency
>>
>> Patrick: really great.  Thank you.
>>
>> Your replacement version is uploaded and checked in (25% smaller), Shay D Pixel has become less shy and is fully with us now.  8)
>>
>> http://x3dGraphics.com/examples/X3dForAdvancedModeling/ShayDPixel/pixelModel_BaseColor.png
>>
>> The first .png version was converted from original .tga Targa image using latest GIMP.  If further image scrutiny is desired by anyone, all original .tga images now available at
>> http://x3dgraphics.com/examples/X3dForAdvancedModeling/ShayDPixel/images/originals/
>>
>> (No good deed goes unpunished!) Request: who is writing up the specification feedback issue?
>>
>>          Web3D Standards Comment Form
>>          http://www.web3d.org/content/web3d-standards-comment-form
>>
> 
> To me the issue does not seem to impact the standard. Each browser
> does its best to render transparency.
> 
> It is an implementation and authoring issue. Perhaps it is a case for
> adding guidance in the ImageTexture tooltip: "If the texture is
> opaque, omitting the alpha channel in the image file can help avoid
> rendering artefacts related to transparency and reduce file size."
> 
> -Andreas

Excellent, thank you, now added to ImageTexture and (slightly edited) to PixelTexture.  Also improved hints for TextureProperties node.

http://www.web3d.org/x3d/tooltips/X3dTooltips.html#ImageTexture
http://www.web3d.org/x3d/tooltips/X3dTooltips.html#PixelTexture
http://www.web3d.org/x3d/tooltips/X3dTooltips.html#TextureProperties

>> On 11/8/2017 7:51 AM, Roy Walmsley wrote:
>>> Patrick,
>>>
>>> Great spot. Thank you.
>>>
>>> That also explains the behaviour of my Coin3D viewer. It saves transparent objects to render after opaque objects, and I had it set to sort transparent objects by distance.
>>>
>>> All the best,
>>>
>>> Roy
>>
>> On 11/8/2017 7:45 AM, Michalis Kamburelis wrote:
>>> 2017-11-08 16:27 GMT+01:00 Patrick D?hne <pdaehne at gmail.com>:
>>>> As soon as the texture file contains an alpha channel, X3D browsers classify shapes with that texture as transparent shapes. Rendering transparent shapes is difficult. Browsers handle transparent shapes differently. Instant Player and X3DOM write transparent objects to the depth buffer. Most other browser don?t. In this special case, the result of Instant Player and X3DOM looks better than result of other browsers, but that is pure coincidence. None of the browsers is wrong or right.
>>>
>>> And view3dscene looks at the texture contents to decide whether to use
>>> alpha testing (which doesn't have problems with ordering) or alpha
>>> blending (which is indeed tricky, but can represent partial
>>> transparency). It can be overridden by "alphaChannel" field,
>>> https://castle-engine.sourceforge.io/x3d_implementation_texturing_extensions.php#section_ext_alpha_channel_detection
>>> .
>>>
>>> Regards,
>>> Michalis
>>
>>> -----Original Message-----
>>> From: x3d-public [mailto:x3d-public-bounces at web3d.org] On Behalf Of Patrick D?hne
>>> Sent: 08 November 2017 15:27
>>> To: X3D Graphics public mailing list <x3d-public at web3d.org>
>>> Subject: Re: [x3d-public] Problem with inverted polygons/normals
>>>
>>>> Am 07.11.2017 um 16:17 schrieb Don Brutzman <brutzman at nps.edu>:
>>>>
>>>> Initial draft version of an interesting new model is available at
>>>>
>>>>       Shay D Pixel Version 2
>>>>       http://x3dgraphics.com/examples/X3dForAdvancedModeling/ShayDPixel/ShayDPixelVersion2Index.html
>>>>
>>>> This version was exported from an .fbx version using Okino Nugraf.  Original at
>>>>
>>>>       http://x3dgraphics.com/examples/X3dForAdvancedModeling/ShayDPixel/pixel_v2.fbx
>>>>
>>>> It looks OK in several players (such as Instant Reality and X3DOM) but polygon ordering appears reversed in X_ITE.  Adjusting the solid field hasn't helped.  Perhaps a normal issue?
>>>>
>>>> Wondering if anyone else can figure out what is going on here, screenshot attached for clarity.  TIA for all help.
>>>
>>>
>>> The texture of the model contains an alpha channel. Remove the alpha channel from the texture, and everything looks right.
>>>
>>> As soon as the texture file contains an alpha channel, X3D browsers classify shapes with that texture as transparent shapes. Rendering transparent shapes is difficult. Browsers handle transparent shapes differently. Instant Player and X3DOM write transparent objects to the depth buffer. Most other browser don?t. In this special case, the result of Instant Player and X3DOM looks better than result of other browsers, but that is pure coincidence. None of the browsers is wrong or right.
>>>
>>> Bye,
>>>
>>> Patrick
>>
>>
>> ------------------------------
>>
>> Message: 4
>> Date: Wed, 8 Nov 2017 11:15:52 -0500
>> From: Russ Kinter <pyth7 at verizon.net>
>> To: x3d-public at web3d.org
>> Subject: [x3d-public] Instant Player Sound Issue on Windows 10
>> Message-ID: <15f9c6a8c9c-c09-144bc at webjas-vad212.srv.aolmail.net>
>> Content-Type: text/plain; charset="utf-8"
>>
>> Hi All,
>> Instant Player 64 bit will not play this simple test file on a new laptop with Windows 10:
>> http://www.deepmatrix.org/Sound_test/Sound_test.wrl
>>
>> All other VRML browsers, including FreeWRL will play the file on the same laptop.
>> My wife's laptop with Windows 10 won't play the file either using Instant Player 64 bit.
>>
>> However, an older version of Instant Player WILL play the file on my XP midtower.
>>
>> Does anyone else with Windows 10 have sound issues with Instant Player 64 bit?
>>
>> Thanks!
>> Russ Kinter

all the best, Don
-- 
Don Brutzman  Naval Postgraduate School, Code USW/Br       brutzman at nps.edu
Watkins 270,  MOVES Institute, Monterey CA 93943-5000 USA   +1.831.656.2149
X3D graphics, virtual worlds, navy robotics http://faculty.nps.edu/brutzman