POV-Ray : Newsgroups : povray.unofficial.patches : MegaPov Motion_blur newbie question Server Time
18 Apr 2024 11:42:30 EDT (-0400)
  MegaPov Motion_blur newbie question (Message 1 to 8 of 8)  
From: danyolgiax
Subject: MegaPov Motion_blur newbie question
Date: 18 Apr 2006 11:15:00
Message: <web.4445023c145106a7a8eb78fb0@news.povray.org>
Hi!

Using motion_blur with a box runs along x axis I've seen this effect is
bidirectional! I've got a blur in front and at the end of the box!

There's a way to make it uni-directional?

Thanks!

Daniele


Post a reply to this message

From: Windell H  Oskay
Subject: Re: MegaPov Motion_blur newbie question
Date: 18 Apr 2006 14:20:00
Message: <web.44452d8e99d124d6959113f40@news.povray.org>
"danyolgiax" <d.o### [at] gmailcom> wrote:
> Using motion_blur with a box runs along x axis I've seen this effect is
> bidirectional! I've got a blur in front and at the end of the box!

Look in the documentation about how motion blur works--
( http://megapov.inetart.net/manual-1.2.1/megapov0121.html#motion_blur  )
it averages over a bunch of possible locations for your object.  There is
nothing in this procedure to give an "arrow of time," so if you want one,
you have to add it in yourself.  A good way to do this is to draw in
another copy of your object, slightly displaced to one end of your blur.

Here's an example:

#version unofficial megapov 1.1;
global_settings {
  assumed_gamma 1.0
   motion_blur 20, (1/10)
 }

light_source { <100,100,100>  color rgb 2  }

plane { y, -10  texture { pigment {rgb 0} } }

#declare MySphere = sphere { 0, 1 texture { pigment {rgb 1} }};

//Copy with motion blur:
 motion_blur {    type 0
object {MySphere translate -x*clock*30}
 }

//Copy withou motion blur:
object {MySphere  translate x*1.5}

camera { angle 10
  location <0,100,0>  look_at <0,0,0>
}


Post a reply to this message

From: Warp
Subject: Re: MegaPov Motion_blur newbie question
Date: 18 Apr 2006 15:57:00
Message: <4445448c@news.povray.org>
Windell H. Oskay <windell(at)oskay.net> wrote:
> A good way to do this is to draw in
> another copy of your object, slightly displaced to one end of your blur.

  But why would you want to do that? It doesn't correspond to any effect
which happens in photography. IOW, the result will be physically incorrect
and doesn't correspond to anything that happens in real life in photography.
It will only be an artificial special effect.

-- 
                                                          - Warp


Post a reply to this message

From: Warp
Subject: Re: MegaPov Motion_blur newbie question
Date: 18 Apr 2006 15:57:14
Message: <4445449a@news.povray.org>
danyolgiax <d.o### [at] gmailcom> wrote:
> There's a way to make it uni-directional?

  What do you mean?

-- 
                                                          - Warp


Post a reply to this message

From: Trevor G Quayle
Subject: Re: MegaPov Motion_blur newbie question
Date: 18 Apr 2006 16:35:00
Message: <web.44454d0399d124d66c4803960@news.povray.org>
"danyolgiax" <d.o### [at] gmailcom> wrote:
> Hi!
>
> Using motion_blur with a box runs along x axis I've seen this effect is
> bidirectional! I've got a blur in front and at the end of the box!
>
> There's a way to make it uni-directional?
>
> Thanks!
>
> Daniele

Actually motion blur works exactly the way it is supposed to.  In reality,
for an object that is moving at a constant speed, each instance of the
object has equal exposure time and equal weighting in the image.  This may
give the impression that there is blur in front and in back, when really
it's not, the the back of the blur is the object at the opening of the
shutter and the front of the blur is the object at the closing of the
shutter.
What you are assuming that you should see would be for an object that was
moving when the shutter opened, but stopped for some period of time while
the shutter was open, giving that 'position' a greater weight in the
exposure.  To do this (I assume this should work as I haven't actually
played with motion blur in megapov), you would need to apply a function to
the object tranform factor that gives a non-linear movement over the period
of time the blur is calculated.

I hope this made sense to you and was helpful.  Coincidentally, I just came
across an article about this very subject the other day that was very clear
on the reasoning, but for I can't locate it at the moment.

-tgq


Post a reply to this message

From: Roman Reiner
Subject: Re: MegaPov Motion_blur newbie question
Date: 18 Apr 2006 17:00:00
Message: <web.4445527e99d124d62593a4600@news.povray.org>
"danyolgiax" <d.o### [at] gmailcom> wrote:
> Hi!
>
> Using motion_blur with a box runs along x axis I've seen this effect is
> bidirectional! I've got a blur in front and at the end of the box!
>
> There's a way to make it uni-directional?
>
> Thanks!
>
> Daniele

I'm not sure what you mean. Maybe you want to use type 1?

"Type 0 will oscillate the motion blur around the current clock value. Half
of the SHUTTER-TIME value is subtracted from, the other half added to that
clock value. Type 0 is the default.

Type 1 will add the full SHUTTER-TIME value to the current clock value."

(from MegaPov Documentation 2.5.2. Motion blur)

Regards Roman


Post a reply to this message

From: Windell H  Oskay
Subject: Re: MegaPov Motion_blur newbie question
Date: 18 Apr 2006 18:00:00
Message: <web.444560ab99d124d6959113f40@news.povray.org>
Warp <war### [at] tagpovrayorg> wrote:

>   But why would you want to do that? It doesn't correspond to any effect
> which happens in photography. IOW, the result will be physically incorrect
> and doesn't correspond to anything that happens in real life in photography.
> It will only be an artificial special effect.

Exactly right-- In a photograph you can never tell the direction that an
object is moving by its blur.
This is a special effect, and a cartoony one at that.

Christoph Hormann has given us an excellent example of how motion blur
"should" be used
( http://megapov.inetart.net/demo/mblur.png ), but there's a lot to be said
for cartoons and illustrations as well.  The question was about how to make
it look that way, not *whether* it should look that way.  =)

-Windell.


Post a reply to this message

From: Veggiet
Subject: Re: MegaPov Motion_blur newbie question
Date: 13 Jul 2007 11:17:22
Message: <46979782$1@news.povray.org>
If this is the effect you want, and your motion command is "translate 
x*clock" then you could just square the value of the clock. like this: 
"translate x*clock*clock*2 "

"Trevor G Quayle" <Tin### [at] hotmailcom> wrote in message 
news:web.44454d0399d124d66c4803960@news.povray.org...
> "danyolgiax" <d.o### [at] gmailcom> wrote:
>> Hi!
>>
>> Using motion_blur with a box runs along x axis I've seen this effect is
>> bidirectional! I've got a blur in front and at the end of the box!
>>
>> There's a way to make it uni-directional?
>>
>> Thanks!
>>
>> Daniele
>
> Actually motion blur works exactly the way it is supposed to.  In reality,
> for an object that is moving at a constant speed, each instance of the
> object has equal exposure time and equal weighting in the image.  This may
> give the impression that there is blur in front and in back, when really
> it's not, the the back of the blur is the object at the opening of the
> shutter and the front of the blur is the object at the closing of the
> shutter.
> What you are assuming that you should see would be for an object that was
> moving when the shutter opened, but stopped for some period of time while
> the shutter was open, giving that 'position' a greater weight in the
> exposure.  To do this (I assume this should work as I haven't actually
> played with motion blur in megapov), you would need to apply a function to
> the object tranform factor that gives a non-linear movement over the 
> period
> of time the blur is calculated.
>
> I hope this made sense to you and was helpful.  Coincidentally, I just 
> came
> across an article about this very subject the other day that was very 
> clear
> on the reasoning, but for I can't locate it at the moment.
>
> -tgq
>
>
>


Post a reply to this message

Copyright 2003-2023 Persistence of Vision Raytracer Pty. Ltd.