Author Topic: Bucket rendering testing needed  (Read 16454 times)

2013-08-23, 00:27:00
Reply #15

cecofuli

  • Active Users
  • **
  • Posts: 1577
    • View Profile
    • www.francescolegrenzi.com
Ok, I did some strong test. I did these two test. Unfortunately, I had to change slightly the shader. But for this purpose isn't so important.
Resolution: A3 300 DPI = 5000 x 3500 px

Progressive = 24h
Bucket= 22h

 I don't remember exactly the Progressive parameter. I think: PTS: 8 or 4 and LSM 2 or 4 .
Bucket: Int samples 32 - Pass: 10 (5 reached in 22 hours)  -  Adaptive threshold: 0.05

As you can see, with bucket I have a little bit more noise in the wall and in the uniform area, but the DOF is much better, in less time.
I hope Ondra find the way to solve this problem (DOF) in progressive mode.





« Last Edit: 2013-08-23, 00:28:37 by cecofuli »

2013-08-23, 10:09:30
Reply #16

maru

  • Corona Team
  • Active Users
  • ****
  • Posts: 12711
  • Marcin
    • View Profile
I don't remember exactly the Progressive parameter. I think: PTS: 8 or 4 and LSM 2 or 4 .
Bucket: Int samples 32 - Pass: 10 (5 reached in 22 hours)  -  Adaptive threshold: 0.05

Isn't int samples 32 an overkill?
Marcin Miodek | chaos-corona.com
3D Support Team Lead - Corona | contact us

2013-08-23, 10:12:28
Reply #17

cecofuli

  • Active Users
  • **
  • Posts: 1577
    • View Profile
    • www.francescolegrenzi.com
I don't know, I haven't time to do more test =) But, for me, I want to obtain the best quality with the current Corona version.

2013-08-23, 11:08:47
Reply #18

maru

  • Corona Team
  • Active Users
  • ****
  • Posts: 12711
  • Marcin
    • View Profile
Ok, but I think you could achieve same results in much shorter time, especially when rendering image at such high resolution. I'm not sure, because all these controls combined - pts, lsm, int samples - are becoming hard to comprehend.
Marcin Miodek | chaos-corona.com
3D Support Team Lead - Corona | contact us

2013-08-30, 18:53:33
Reply #19

maru

  • Corona Team
  • Active Users
  • ****
  • Posts: 12711
  • Marcin
    • View Profile
Why does a scene rendered with exactly one pass in progressive mode look different than the same scene rendered with exactly one pass with 1 initial sample in bucket mode? What's the difference? Can someone explain this?
Ok, it was because of HD cache.
« Last Edit: 2013-08-30, 19:14:19 by maru »
Marcin Miodek | chaos-corona.com
3D Support Team Lead - Corona | contact us

2013-09-15, 01:19:10
Reply #20

Stan_But

  • Active Users
  • **
  • Posts: 526
    • View Profile
    • https://www.behance.net/archdizs
i found when render with bucket mode that some bucket can run very slowly in very simple places without reflection with diffuse solid color. Even if override with default corona material.
Is it a bug?

2013-09-15, 10:29:04
Reply #21

rafpug

  • Active Users
  • **
  • Posts: 722
    • View Profile
Ok, I did some strong test. I did these two test. Unfortunately, I had to change slightly the shader. But for this purpose isn't so important.
Resolution: A3 300 DPI = 5000 x 3500 px

Progressive = 24h
Bucket= 22h

 I don't remember exactly the Progressive parameter. I think: PTS: 8 or 4 and LSM 2 or 4 .
Bucket: Int samples 32 - Pass: 10 (5 reached in 22 hours)  -  Adaptive threshold: 0.05

As you can see, with bucket I have a little bit more noise in the wall and in the uniform area, but the DOF is much better, in less time.
I hope Ondra find the way to solve this problem (DOF) in progressive mode.



Good Morning

Progressive = 24h
Bucket= 22h


1 day in rendering?
All this time, but is it possible?

Greats
Raf