# Reducemesh - Rhino BETA - different output compared to Rhino 8

**URL:** <https://discourse.mcneel.com/t/reducemesh-rhino-beta-different-output-compared-to-rhino-8/224231>\
**Category:** Serengeti (Rhino BETA)\
**Tags:** windows, reducemesh, planaronly\
**Created:** [November 3, 2026, 12:32pm UTC](https://discourse.mcneel.com/t/reducemesh-rhino-beta-different-output-compared-to-rhino-8/224231 "2026-11-03T12:32:04Z")\
**Posts on this page:** 13\
**Page:** 1

<div class="post-metadata">

**Author:** ![Erdem](https://sea3.discourse-cdn.com/mcneel/user_avatar/discourse.mcneel.com/erdem/32/762552_2.png) [@Erdem](https://discourse.mcneel.com/u/Erdem)\
**Post date:** [November 3, 2026, 12:32pm UTC](https://discourse.mcneel.com/t/reducemesh-rhino-beta-different-output-compared-to-rhino-8/224231/1 "2026-11-03T12:32:04Z")

</div>

Hello guys,

I recently noticed that ReduceMesh with planaronly option behaves very different in RH-Beta compared to RH-8. Here is a comparison. Does anybody else have that also?

This is the expected result from RH-8:

 ![image](https://global.discourse-cdn.com/mcneel/uploads/default/original/4X/1/2/2/12219c15efc765092545732e40cabff7f87689fa.png)

And this is what I get on RH-BETA:

 ![image](https://global.discourse-cdn.com/mcneel/uploads/default/original/4X/4/a/3/4a37c3aa984c808172665b42f15c0f0c2dd69c54.png)

---

<div class="post-metadata">

**Author:** ![davidcockey](https://sea3.discourse-cdn.com/mcneel/user_avatar/discourse.mcneel.com/davidcockey/32/771921_2.png) [@davidcockey](https://discourse.mcneel.com/u/davidcockey)\
**Post date:** [November 3, 2026, 12:55pm UTC](https://discourse.mcneel.com/t/reducemesh-rhino-beta-different-output-compared-to-rhino-8/224231/2 "2026-11-03T12:55:13Z")

</div>

Post a .3dm file with your geometry.

---

<div class="post-metadata">

**Author:** ![Erdem](https://sea3.discourse-cdn.com/mcneel/user_avatar/discourse.mcneel.com/erdem/32/762552_2.png) [@Erdem](https://discourse.mcneel.com/u/Erdem)\
**Post date:** [November 3, 2026, 12:56pm UTC](https://discourse.mcneel.com/t/reducemesh-rhino-beta-different-output-compared-to-rhino-8/224231/3 "2026-11-03T12:56:45Z")

</div>

Rhino 8 or BETA?

---

<div class="post-metadata">

**Author:** ![davidcockey](https://sea3.discourse-cdn.com/mcneel/user_avatar/discourse.mcneel.com/davidcockey/32/771921_2.png) [@davidcockey](https://discourse.mcneel.com/u/davidcockey)\
**Post date:** [November 3, 2026, 12:57pm UTC](https://discourse.mcneel.com/t/reducemesh-rhino-beta-different-output-compared-to-rhino-8/224231/4 "2026-11-03T12:57:48Z")

</div>

Post a V8 .3dm file with the geometry. That file can be opened in both V8 and Beta/V9.

---

<div class="post-metadata">

**Author:** ![Erdem](https://sea3.discourse-cdn.com/mcneel/user_avatar/discourse.mcneel.com/erdem/32/762552_2.png) [@Erdem](https://discourse.mcneel.com/u/Erdem)\
**Post date:** [November 3, 2026, 1:04pm UTC](https://discourse.mcneel.com/t/reducemesh-rhino-beta-different-output-compared-to-rhino-8/224231/5 "2026-11-03T13:04:56Z")

</div>

[Reducemesh-planaronly.3dm](https://discourse.mcneel.com/uploads/short-url/iD6VYhsS9kcey9QIW5omPGJwJGq.3dm) (5.1 MB)  
I cannot upload the geometry in the image but I can reproduce it with anything. Even with a simple sphere… See the file.

---

<div class="post-metadata">

**Author:** ![Japhy](https://sea3.discourse-cdn.com/mcneel/user_avatar/discourse.mcneel.com/japhy/32/487647_2.png) [@Japhy](https://discourse.mcneel.com/u/Japhy)\
**Post date:** [November 3, 2026, 1:19pm UTC](https://discourse.mcneel.com/t/reducemesh-rhino-beta-different-output-compared-to-rhino-8/224231/6 "2026-11-03T13:19:14Z")

</div>

Hi Erdem,

Thanks for the file, i was able to repeat it here. The same sphere goes to 360 faces with Planar only in the 9 BETA and stays smooth in 8. Logged as [RH-99976](https://mcneel.myjetbrains.com/youtrack/issue/RH-99976) ReduceMesh Planar only distorts curved meshes in 9

---

<div class="post-metadata">

**Author:** ![piac](https://sea3.discourse-cdn.com/mcneel/user_avatar/discourse.mcneel.com/piac/32/769366_2.png) [@piac](https://discourse.mcneel.com/u/piac)\
**Post date:** [November 5, 2026, 4:50pm UTC](https://discourse.mcneel.com/t/reducemesh-rhino-beta-different-output-compared-to-rhino-8/224231/7 "2026-11-05T16:50:22Z")

</div>

Hey @Erdem

thank you very much for the sample. The problem that the Rhino 9 change was tackling is coming from the fact that the tolerance value for PlanarOnly was entirely fixed, in a way that does not scale with the size of the object. Therefore, for some small objects, the entire function failed once the “PlanarOnly” option is turned on.

However, it seems that this made the values too relaxed for larger objects.

I’ve tackled this so that the tolerance is reduced for “small” objects, but it still is never increased more than in Rhino 8. This should both cover your case and also other cases where geometry is small. We will see if more tweaks are necessary.

This will be present in the next 9 SR1 candidate.

---

<div class="post-metadata">

**Author:** ![Holo](https://sea3.discourse-cdn.com/mcneel/user_avatar/discourse.mcneel.com/holo/32/496872_2.png) [@Holo](https://discourse.mcneel.com/u/Holo)\
**Post date:** [November 6, 2026, 9:40am UTC](https://discourse.mcneel.com/t/reducemesh-rhino-beta-different-output-compared-to-rhino-8/224231/8 "2026-11-06T09:40:39Z")

</div>

Wish:

Please show a tolerance setting to PlanarOnly if it is tagged on.  
default = use document tolerance  
custom = numeric value

Meshes are often built with less tolerance than 0.001 mm so to me this would be a nice addition.

---

<div class="post-metadata">

**Author:** ![piac](https://sea3.discourse-cdn.com/mcneel/user_avatar/discourse.mcneel.com/piac/32/769366_2.png) [@piac](https://discourse.mcneel.com/u/piac)\
**Post date:** [November 6, 2026, 10:37am UTC](https://discourse.mcneel.com/t/reducemesh-rhino-beta-different-output-compared-to-rhino-8/224231/9 "2026-11-06T10:37:24Z")

</div>

> [@Holo](#):
>
> a tolerance setting to PlanarOnly

[RH-100075](https://mcneel.myjetbrains.com/youtrack/issue/RH-100075/Tolerance-setting-to-PlanarOnly) filed.

We wonder still, how this tolerance default should be computed on the actual mesh: fixed, linear, following curvature, etc? This is the last step missing in this area IMO. We will think about this and come back with something useful, I hope “soon”.

---

<div class="post-metadata">

**Author:** ![Holo](https://sea3.discourse-cdn.com/mcneel/user_avatar/discourse.mcneel.com/holo/32/496872_2.png) [@Holo](https://discourse.mcneel.com/u/Holo)\
**Post date:** [November 6, 2026, 12:36pm UTC](https://discourse.mcneel.com/t/reducemesh-rhino-beta-different-output-compared-to-rhino-8/224231/10 "2026-11-06T12:36:57Z")

</div>

If the vertices on the plane of one face deviate more than tolerance from the other face’s plane then not planar within tolerance?

---

<div class="post-metadata">

**Author:** ![piac](https://sea3.discourse-cdn.com/mcneel/user_avatar/discourse.mcneel.com/piac/32/769366_2.png) [@piac](https://discourse.mcneel.com/u/piac)\
**Post date:** [November 6, 2026, 6:03pm UTC](https://discourse.mcneel.com/t/reducemesh-rhino-beta-different-output-compared-to-rhino-8/224231/11 "2026-11-06T18:03:57Z")

</div>

@Holo yes it’s a good idea, but with something small enough, faceted small enough, that could easily encompass the entire mesh or too large portions. The main question is how to restrict this to meaningful areas.

---

<div class="post-metadata">

**Author:** ![Holo](https://sea3.discourse-cdn.com/mcneel/user_avatar/discourse.mcneel.com/holo/32/496872_2.png) [@Holo](https://discourse.mcneel.com/u/Holo)\
**Post date:** [November 6, 2026, 6:15pm UTC](https://discourse.mcneel.com/t/reducemesh-rhino-beta-different-output-compared-to-rhino-8/224231/12 "2026-11-06T18:15:05Z")

</div>

I understood the concept and tried to answer the question 🙂 I think that if two planes are within the documents tolerance they should be considered planar… It kind of is the point with a tolerance and affects whether we can join curves and surfaces if ends/edges are within that tolerance, so to me it makes sense to use the same value for coplanarness.

But, how does it work for nurbs mergecooplanarfaces?

---

<div class="post-metadata">

**Author:** ![davidcockey](https://sea3.discourse-cdn.com/mcneel/user_avatar/discourse.mcneel.com/davidcockey/32/771921_2.png) [@davidcockey](https://discourse.mcneel.com/u/davidcockey)\
**Post date:** [November 6, 2026, 6:45pm UTC](https://discourse.mcneel.com/t/reducemesh-rhino-beta-different-output-compared-to-rhino-8/224231/13 "2026-11-06T18:45:12Z")

</div>

> [@piac](#):
>
> We wonder still, how this tolerance default should be computed on the actual mesh: fixed, linear, following curvature, etc? This is the last step missing in this area IMO.

Test the angle of the normals between the faces with a tolerance for the maximum acceptable angle. I’m working on a simple way to calculate the angle, though perhaps Rhino already has a suitable method.

The angle test could be combined with a maximum distance deviation test.