Thanks. Even though most of this is on the website, it's great to find everything together in one handy spreadsheet.

Daryl Dunlap
Download Redshift3d, a GPU renderer for Houdini which has Particle support. Using a GTX 860M I was able to render 10 million particles with motion blur with no problem.

Ever the Redshift advocate Daryll.

But Daryll's right Redshift is able to cope with large particle counts and is far faster than Mantra. Krakatoa is still king of the hill with it's specialist shading options, super fast motion blur and optimised cache loading but Redshift is an ace native tool in Houdini that works directly with BGEO particle caches and more importantly provides easy access to particle attributes for nifty shading options.

However, with regard to the OP's original question Redshift doesn't contain anything that will multiply the point count, that's where Wedging will be your best friend.

Just replied to you in the siLib thread as I agree arrays are a real weakness in VOPs when compared to equivalent ICE workflows. Obviously it would be great if SideFX could provide native VOP solutions but one of the aims of the siLib library is to fill those gaps; so any ideas you have would be greatfully recieved.

@probiner, we have a Google group set up where any ex XSI users can contribute to the project which Andy Nicholas is leading on. Looking through those VOP assets suggests you have many worthy candidates for inclusion. Especially promising to see some VOP alternatives for working with arrays. The array nodes in XSI where excellent and to achieve equivents workflows in Houdini means it's Wrangle time (intimidating for many artists coming over to Houdini from XSI).

https://groups.google.com/forum/#!forum/houdini-silib

By the way, you know me as @LateralPixels over on your excellent Discord board.

A very good way to do something similar with the same end goal, is to use Wedging (look it up in the documentation) to write separate ‘wedges’ (partitions in Krakatoa speak) with different seed values on the emitter. Mantra will never provide the same per frame speed as Krakatoa but breaking up your sim into 5-10 million particle chunks will ease the stress of loading and it's far faster than the point replicate shader and more importantly provides more natural and visually pleasing results.

If you have access to Krakatoa via another DCC there is a PRT exporter available so you can utilise Houdini to create the sim but then render it in Krakatoa. The only downside to this approach is that the PRT exporter is single threaded. Houdini particles are fantastically threaded so under normal circumstances they write a BGEO cache in approx 1 second a frame (based on a5-10 million particles and a 32 thread workstation). With the PRT exporter being single threaded it will be far slower. Having said that if you're happy with your sim you can wedge overnight and be ready to move things on in Krakatoa.

On that basis Mantra might be the faster workflow even though it's per frame render time won't match Krakatoa's (which is highly optimised for particle sims). You'll need to know your way around the VOP shading networks too if you're attempting similar results to Krakatoa, so the benefits of each approach will vary on your personal skillset.

ikoon
As regards locking of the horizon … isn't it possible to script it somehow? The Z component of rotation should be zero, but simple “lock parameter” doesn't work. It is deeper somewhere, does it really require Sesi to implement it?

Yes it is something that needs to be implemented properly but the good news is that 3dConnexion provide a template that SESI can work from. Ideally we'll end up with the other camera modes too which help enrich the user experience.



Being able to change between target camera mode and camera mode is really useful too. Especially when setting up cameras in your scene (the camera mode is more of a first person fly through mode rather than rotating around an object).

I must admit I'm very disappointed by the 3dConnexion implementation in Houdini; it's as basic a level of support as you can get and it suffers from bad frame rates on occasions with no real explicable cause.

The main reason I stick with it is that I've set up my Radial Menus to be triggered by buttons on my SpaceMouse Pro. Really useful when modelling as I don't have to go near the keyboard.

Just wanted to report back that in this instance @tamte's response gave me what I wanted (an ability to set detail attribute values in ‘Run over Numbers’ mode). But @Enivob's response taught me a few tricks ref setting detail attributes whilst running a Wrangle over Points.

So good day at the office to all concerned.

Daryl Dunlap
H calculates an intrinsic bbox for you automatically, which will give you min/max in all 3 axes.

Aware of that Daryl. What I'm looking for is the exact point position within any arbitrary point cloud (the mountain distorted grid is just a measurable test case). My original wrangle code works as it should but it's single threaded because it's running in details mode (see image below). My original request was to get a better idea of how to utilize the wrangle in Numbers mode (introduced in 15.5). This allows you to explicitly tell Houdini how many threads you want the job to use (whilst still working at Detail level).

Tamte's original reply should answer my needs but I haven't had a chance to test it as I've got a deadline looming.

Enivob
Hmm.. maybe you have an object level transformation in play? Try setting top level transformation to 0,0,0 and scale to 1,1,1.

No object level transforms or scaling. Thanks for the suggestion anyway.

I've been drawn into other things and won't be able to investigate further till this evening. The approach you've laid out has opened my eyes to more efficient ways of working and I'm sure I'll figure out what's going on here.

@Enivob, had a play with your single Wrangle approach and the good news is that it works and it's lightning fast but the result it provides isn't the true min and max as you can see here. I'll try and debug it myself as I'm sure it's something simple but any thoughts from your end will be appreciated too,

Thanks too Enivob.

Liking the look of that. Why take two wrangles when one will do the job.

I'll have a play and report back on both approaches.

I don't come from a programing background but have taught myself what I can of VEX off the back of previous experience of Python and Processing (and funnily enough, it's my knowledge of Processing that has more helpful in trying to get to grips with VEX). I hate the idea of having to learn a C like language to get more out of Houdini, especially when I mainly only use wrangles. But Houdini seems to becoming more and more Vex focussed with every release so I may have to bite the bullet.

Thanks for your reply Tomas. I do use Attribute Promote on the next step of the tree.

I think my problem was that I was expecting the detail attribute to be written automatically in the same manner as when the Wrangle is set to ‘Run over Detail’. With the documentation mentioning that "no attributes will be bound, except read only Detail attributes I thought I could run the same snippet without amendment. SetDetailAttrib certainly sounds like it will do what I want in Numbers mode. I'll report back ref how I get on.



The purpose of the scene was strictly research as I've never used the ‘Run Over Numbers’ mode since it was introduced in 15.5 and I wanted to find a better way of wrangling point clouds in the multi million range seeing as Detail mode is single threaded.

I understand what “Run over Numbers” is meant to achieve but not how it works. I have a simple network where I'm looking to find the high and low value of a point array on the x, y or z and the majority of the work is done via two Attribute Wrangles running in Detail Mode, meaning they're being run single threaded.

This is what the code looks like:

Wrangle 1

Wrangle 2


I get the result I want in a test case of 100K points scattered over a Mountain Sop deformed Grid. My worry is that this solution won't scale very well so I've been investigating “Run over Numbers” but I can't get the Wrangles to write the array info to Detail attributes in the same manner as running the Wrangles in “Run over Detail(once only)” mode.

I'd ideally like to suss this out as it seems a perfect opportunity for splitting the job up to multiple threads so the solution will scale better, but there's no usage documentation anywhere other than the cryptic tooltips and associated help card info.

Any guidance would be most appreciated.

Read through this thread. It explains why there can be compatibility problems with earlier versions of Houdini.


polysplit in h16 generates different point order? [sidefx.com]

It's not the same.

When you disable tilt and rotate round your object you get a tilt as part of the rotation that you can't correct because you no longer have any control over the tilt.

It's not just Autodesk and Modo that provide the lock horizon option and they all do it for good reason.

I'm fine with the current driver in Houdini. The only thing I change is to reverse all of the options. I think that comes from a misspent youth spending far too much of my spare time playing flight sims.

X_H_N
OK I find solution when I use shop instead of mat from mats export work correctly.

You should report that errant behaviour to support. The new Mat context should work exactly the same way as the Shop context for FBX exports.

I have an enhancement request to allow the locking of the horizon (this is an option on all Autodesk applications and Modo too). It helps keep things far more fluid when using a 3dConnexion device.

Artye

Jonathan Moore2
It's there, on your local system help, scroll down:

/ref/windows/pythonpaneleditor

Thanks. That's very helpful indeed.

The initial build didn't have the Load button, but the 16.0.529 nightly has that remedied.

I'm finding the QT5 help a tad shaky on these early builds but I'm sure it will settle before too long.

Thanks again.

Thanks for the suggestion Artye. Unfortunately Python Viewport Color editor doesn't appear to be included in the H16 examples. It's included in the last 15.5 production build but the ‘load example’ button is missing in H16.0504.20.

I'm sure it's just an oversight in the new H16 help pages so I'll investigate again when the H16 nightly builds are available.

No problem, I'll post ER's for the lock horizon and grid opacity requests.