Hello! I’m currently working as a lead technical artist in the product visualization sector for a larger outfit. The team’s mission is to produce static imagery of our product, at scale, to assist the company marketing effort. Animations are a secondary effort. The team/dept is new to the company.
Looking for some higher-level advice from anyone here familiar with such product visualization efforts and have dealt with coordinating a small to medium team of artists (6-12+ people). Below is background information followed by advice I’m hoping to find here -
Background Information:
Developing a scalable engineering CAD to USD pipeline is currently my main effort (something akin to digital twin generation from manufacturing product data). While I’ve been building up this pipeline, the artists have been using off-the-shelf Cinema4D to produce renders from USD data I’ve been able to provide. We have a lot of product, a lot of CAD and a lot of PLM data alongside a ~2-year product lifecycle. Pipeline (python) and automation will be key to keeping up.
A big draw of USD data is the ability to inject engineering metadata directly into the geometry. We know which parts are which, what colors need to be represented, what the engineering assembly hierarchy should be, etc… We can even trace back to the source CAD this way. Very cool.
After engineering CAD => USD pipeline has been pushed up to the point of diminishing returns, I’ll be focusing on what I’m calling an “Integrations” effort. Integrations involves taking the robust USD datasets and building around them to drive process in the artist DCC.
Pipeline:(CAD => Houdini => USD).
Heavy use of Solaris, TOPs, some SOPs. Planned mtlx development with USD.
Integrations: (USD => Artist DCC => Render Output).
Interpreting the USD data, all of it’s metadata to drive process within the DCC. Minimizing artist touch points as much as possible.
I’m trying to identify hazards and avoid them. Building up operations around Keyshot, for example, is an obvious hazard for the operation as Keyshot is not really a production solution for a scalable 3D rendering operation.
Thinking ahead to this “Integrations” effort, I’m considering two artist DCCs - Blender or Cinema4D. Maya is off-the-table or I’d just go with Maya (familiar).
Cinema4D:
I’ve taken a gander at Cinema4D and it’s API and found it clunky from a technical perspective (coming from Houdini which is my primary and Maya). Although Cinema4D is very simple to grasp from the artist perspective, I’m concerned that it’s a bit too basic, off-the-shelf and will ultimately hinder scalability of the operation from a technical perspective. A lot of presumptions on my part, please input if you disagree.
Currently the artists are using Cinema4D as that’s what they know predominately. I do not suggest a DCC change to blender lightly.
Blender 4.2LTS:
I’m not a Blender fanboi, but I have been watching development progress of the solution for the last 6 years on/off. It looks like Blender has matured to the point of being a valid DCC to have in production (to build pipeline around). I have not, but will be testing out it’s capability and API soon.
Blender = No license restrictions, what looks to be a sensible API, less black-boxed than Cinema4D.
Questions (get to the point already):
Thank you staying with me. A bit to read. Here are my questions.
-
Input on building up an operation around Cinema4D? Is C4D intended for a team and a more complex operation as I’m describing? Is the API as weirdly structured as I think it is or should I be giving it a deeper analysis?
-
Input on building up an operation around Blender? Is Blender meant for a team and a more complex operation as I’m describing? How stable has the LTS versions been for you?
-
What is your understanding of the current and growing talent pool of artists that use Blender vs Cinema? Seems like a lot of college graduates I’ve talked with are aligned with Blender predominately. By comparison, it’s been more difficult to find Cinema4D artists for product visualization (not just mograph).
-
USD interoperability: Right now I’m underwhelmed by both C4D and Blender regarding USD support. Maxon recently joined the AOUSD, which is neat? Blender’s scene graph construct doesn’t work like USD and 100% support for USD for Blender probably will not happen… However, geometry nodes look a whole heck of a lot like Houdini’s attribute spreadsheet. What are your thoughts on the future of USD interoperability for Blender, for Cinema4D?
-
Ultimately, between Blender or Cinema4D, which solution makes more sense to build a fairly complex, USD and Python heavy, high-volume production product visualization operation around that has a Houdini / Engine backbone?
I’m leaning towards Blender, but I’m not familiar enough with either Blender or Cinema4D to be super confident, which Is why I wanted to post up here.
If they are both just fantastic solutions, I’ll dive into Cinema4D and learn it’s API in-depth. If Cinema4D is more of a off-the-shelf solution meant for cookie-cutter productions in MoGraph land (which seems like the case), I’ll dive into Blender and it’s API in-depth.
Thank you for any advice / input. Please ask additional questions if more clarification would help.