 |
|
|
|
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Invisible escribió:
> Nicolas Alvarez wrote:
> > Note that morphs cannot be played backwards and make sense.
> Yes they can. But MPEG-2 doesn't make that real easy...
I know MPEG doesn't make it easy, but they wouldn't make sense anyway-
Try it. Make a video of A, A, B->A backwards, B, B. See if the morph
looks OK.
>> A->B and B->A videos are different, so both are needed if you wanted
>> all possible combinations.
>
> They *can* be different, but no mathematical law that says they must be.
> (However, ES seems to implement it this way.)
In the way ES works, the loop is "rotating". The morph is interpolating
sheep A to sheep B and *also* rotating. If you play it backwards, it
would be rotating backwards, and the sudden jump in "direction" is
really noticeable.
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Nicolas Alvarez escribió:
> Try it. Make a video of A, A, B->A backwards, B, B. See if the morph
> looks OK.
Actually, don't. I'll try to do it for you.
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
>> They *can* be different, but no mathematical law that says they must
>> be. (However, ES seems to implement it this way.)
>
> In the way ES works, the loop is "rotating". The morph is interpolating
> sheep A to sheep B and *also* rotating. If you play it backwards, it
> would be rotating backwards, and the sudden jump in "direction" is
> really noticeable.
I see. So there's a discontinuity in the 1st derivative then? ;-)
--
http://blog.orphi.me.uk/
http://www.zazzle.com/MathematicalOrchid*
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Nicolas Alvarez wrote:
> I used to, but stopped because the MPEG video showing at fullscreen
> (plus the fact that you may end up rendering frames for them) needs some
> CPU time.
<I'm old>
I can remember when decoding MPEG-1 in realtime required specialised
hardware - and encoding required a very high-end computer system!
And today, my PC is decoding MPEG-2 in realtime using idle CPU cycles...
How times have changed!
</I'm old>
--
http://blog.orphi.me.uk/
http://www.zazzle.com/MathematicalOrchid*
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Invisible escribió:
> I see. So there's a discontinuity in the 1st derivative then? ;-)
>
On the What?
o_0
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Nicolas Alvarez wrote:
> Invisible escribió:
>> I see. So there's a discontinuity in the 1st derivative then? ;-)
>>
>
> On the What?
>
> o_0
Speed is the first derivating of position with respect to time.
In other words, the consecutive frames match, but the movement does not.
--
http://blog.orphi.me.uk/
http://www.zazzle.com/MathematicalOrchid*
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
>> Try it. Make a video of A, A, B->A backwards, B, B. See if the morph
>> looks OK.
>
> Actually, don't. I'll try to do it for you.
Isn't as noticeable as I expected (maybe it's because of the morph I
picked) but definitely there is a jump.
http://www.wikifortio.com/798356/test.mpg.zip (10MB)
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Invisible wrote:
> I can remember when decoding MPEG-1 in realtime required specialised
> hardware - and encoding required a very high-end computer system!
Bah. I was out of grad school and people were still building and selling
hardware JPEG cards, let alone MPEG. :-)
The 6 minutes of mpeg we had to demo the system took 14 weeks to encode
at a service beureau.
--
Darren New / San Diego, CA, USA (PST)
It's not feature creep if you put it
at the end and adjust the release date.
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
"Invisible" <voi### [at] dev null> wrote:
> Rune wrote:
>> "Invisible" <voi### [at] devnull> wrote?
>>> Now, if I could make it download some more sheep... (You'd think 350
>>> would be enough. However, it seems to be more like 15 actual images,
>>> plus 15! morphs between them. Which isn't so interesting.)
>>
>> Why 15! ?
>>
>> For X images, wouldn't the necassary morphs between them be only
>> X*(X-1)/2 and not X! ?
>
> Hmm. Your knowledge of combinatorics clearly exceeds mine considerably...
> :-}
Sounds odd, given that I don't understand most of the math you normally post
about here...
Anyway, since both forward and backwards morphs are needed, as well as the
loops themselves, that would be exactly X^2 animations if all loops should
be connected to all other loops (X loops and X*(X-1) transitions between
them).
Rune
--
http://runevision.com
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
>> I can remember when decoding MPEG-1 in realtime required specialised
>> hardware - and encoding required a very high-end computer system!
>
> Bah. I was out of grad school and people were still building and selling
> hardware JPEG cards, let alone MPEG. :-)
Ooo... LOL!
> The 6 minutes of mpeg we had to demo the system took 14 weeks to encode
> at a service beureau.
ADVANCED!!
[Hmm, POV-Ray service beureau, anyone?]
--
http://blog.orphi.me.uk/
http://www.zazzle.com/MathematicalOrchid*
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
