Linear interpolation of two vector arrays with different lenghts

Started by
54 comments, last by ManitouVR 4 years, 6 months ago

I made it yes.

6 minutes ago, Zakwayda said:

The only difference is that in my example the time information is reduced to a single value for each point

But isnt that exactly what i have at the moment? Dont we need more resolution on the 3rd dimension to preserve the timing data?

10 minutes ago, Zakwayda said:

but rather information about changes in speed

How would you store this information in one value??

12 minutes ago, Zakwayda said:

If that's the case, it sounds like maybe your problem is essentially solved

Yes sir! It seems to be solved for now if (pressuredata == LooksGoodAsWell){return true;}

?

3D Artist

Advertisement
20 minutes ago, C3D_ said:

Yes sir! It seems to be solved for now if (pressuredata == LooksGoodAsWell){return true;}

?

Great - I hope that conditional evaluates to true :)

Quote

But isnt that exactly what i have at the moment? Dont we need more resolution on the 3rd dimension to preserve the timing data?

Quote

How would you store this information in one value??

The resolution in the 3rd dimension would be the same as that of the original curve - that is, one value per point. Currently you do have data for each point - the absolute and delta times. What we'd need though for each point is information about the change in speed between the segment preceding the point and the segment following it (this is all assuming my idea makes sense and would actually work). In my previous post I included some ideas about how this value could be computed.

In any case, I think it's an interesting problem, and if you decide you want to pursue this idea, I can try to elaborate on what I mean. But, if you're reasonably happy with the results you have now, it may be that there'd be little point in pursuing more complex solutions. As we've established, the current solution generally only loses timing information along relatively straight parts of the path, which may not be something the user is likely to be concerned about or even notice.

19 minutes ago, Zakwayda said:
information about the change in speed
Like 0 to 100mph in 3 seconds. That would be one value but still linear. I dont get your idea... but, never mind. Let's stop here for now. :)
23 minutes ago, Zakwayda said:
In my previous post I included some ideas about how this value could be computed
I have to read it again and many times more to grasp it. But i will try to get it! (I was able to grasp the Chaikin algorithm by the way... hehe! ^^)
24 minutes ago, Zakwayda said:
I think it's an interesting problem, and if you decide you want to pursue this idea, I can try to elaborate on what I mean.
Okay. Thanks a lot. I think its a very interesting problem as well.  I will now include the pen pressure and if this wont work out -> back to drawing board. But i really really hope this wont be the case... although the idea of straight lines might destroy my hopes...
Because making a straight line, the timing might go through... but the pressure... we will see.
I'll post back asap.
Thanks to everybody involved. @Zakwayda @alvaro
One last question... Would i open a new thread here or would i continue this thread?

3D Artist

1 hour ago, C3D_ said:
One last question... Would i open a new thread here or would i continue this thread?

I think only the mods can answer that question. I'd think any continuation of this discussion would belong here (I'd imagine future readers would appreciate having all the information and discussion in one place), but that's just my guess as a user of the forums.

Okay,thanks! Then i will continue here if necessary.

@Zakwayda Thanks man! Second time you helped me out succesfully! Couldnt have done it alone. :) Also thanks to @alvaro. Great postprocessing algorithm! Still amazed how it does its job with crappy data out of thin air! Thanks for your effort and time!

 

3D Artist

This topic is closed to new replies.

Advertisement