Animation Add-On Fee?

[particlerealities]

An old, very good client (with relatively shallow pockets) wants a previous restaurant project animated via panning cameras, stitched together into a short movie, maybe 1-3 minutes long. I've been asked to guesstimate how much money it will cost to animate it. I will render everything internally. To do so, I will have to go back through the scene and convert it to GPU and optimize for render speed per frame. I'm not really concerned with charging the client for computer render time, but I will need to buy another GPU (or two).

I'm sure this is pretty vague with a lot of unanswered questions that I'm not even thinking about, but animation is not something we typically output.

I'm curious to hear what you all would charge for something like this? I'm thinking a 2-3 days to convert the model & materials to GPU, 1 day to animate the cameras, then a few hours to stitch together in Premiere. So 4-5 days of labor, plus $1,000 for hardware upgrades. I already have a dedicated GTX980Ti, and hope to stick a GTX1080Ti in the case, if we can find the power (I have a 1300W PSU in my Dell 7910, but no leads currently visible unless there are some unused ports on the PSU that's buried in the case). I would appreciate hearing what you'd charge for this.

[stupendousman_dk]

How about $ 5,000 - Labor is paid for and you get a new GPU. Seems about fair.
Problems will turn up in your conversion and you will properly punch in some more hours than estimated though.

[particlerealities]

Yeah, I'm already running into problems. LOL Render times are way too long, and my workstation doesn't have the physical room for another dual-slot card, and I don't have the available power leads to even run a third GPU (only have 2 8-pins, and 2 6-pins). So, looks like I'll either be paying for an additional render node license and sticking a GPU in a coworker's workstation, or buying an additional CPU for my workstation and rendering in CPU mode (my Dell has a dual CPU motherboard, but I.T. only purchased one CPU).

[stupendousman_dk]

Forget RT and render with Adv. at fairly low settings. Make use of VrayDenoiser and 'speed' your camera moves so motion blur seems natural, add it in post. Render at a lower resolution and upscale your render a little bit in post, maybe add a bit of subtle grain. Remember clients are not professionials, even though it hurts your professional pride you can pull it off, and client will never tell the difference.

[infranea]

I'm thinking a 2-3 days to convert the model & materials to GPU, 1 day to animate the cameras, then a few hours to stitch together in Premiere. So 4-5 days of labor, plus $1,000 for hardware upgrades.

I can't help you much with costs estimation, you should think about how much time will you need for every task, and then use some multiplayer based on risk factor. Like for example 1.0 for camera animation, as there is not much that can go wrong with that, and maybe 1.25 for scenes conversion, because you might have some issues here with materials/lights etc.

The one thing that I'd like to mention here is that charging your client 100% for your hardware investment doesn't seem to be fair. It's not that you will deliver animation with this hardware included, its investment that you do for your company and it will for sure help you out in the future.

[particlerealities]

The one thing that I'd like to mention here is that charging your client 100% for your hardware investment doesn't seem to be fair. It's not that you will deliver animation with this hardware included, its investment that you do for your company and it will for sure help you out in the future.

Agreed. This client in the past has been willing to help pay for various upgrades in the past, like a recent project they needed us to edit for them that was created by a previous firm with V-Ray for Sketchup. We use SketchUp, but never had V-Ray for it. The client covered the license cost. There's only so much I can work into the fee, but I have to at least try to wrap some of the upgrades into the project cost.

[stupendousman_dk]

If you hire a carpenter you expect him to bring his tools, at the agreed hourly wages. But of cause there is some wiggleroom for the costs of material supply for the project.

I think software / hardware goes somewhere in between. It is fair to say we have the drilling machine, but we need a new drill for the specific task. Or to compare, we have v-ray but need another gpu.

Anyway it is difficult, you could also just charge a higher hourly and then suddenly be able to afford the gpu out of the budget.

[particlerealities]

Here's my thinking... In the past, we outsourced our animation processing to a render farm. Did our test frames, then submitted the job. Went home, and came to work the next day to check the render farm progress, only to find most of their output was inconsistent, random black frames, essentially unusable. And it cost us thousands of dollars overnight. We switched to another farm that "held our hand" to help us generate the output, but in the end, we still incurred around $25k in overall farm fees by the time the project was complete. I am trying to save that headache by keeping the frame processing internal. The client is paying for an animation, and part of the overall cost of an animation is core-hour processing time. Does everyone just eat that time? Or do they guess at the farm fee months prior to submitting the job to the farm and include it in the proposal to the client? Or do they pass the farm's invoice on to the client once farm-services are rendered? I'm essentially charging the client up front for a portion of the render farm fee, and eating the rest of the cost. The animation won't be done processing nearly as quickly, but there is a known, up-front cost for that part of the service. Seems reasonable to me.

[andybot_cg]

I've found that it's just not realistic to expect to render frames for an animation internally. Last time I calculated the time to render an animation based on some test frames, it would have taken about a week and a half running my couple of render nodes full capacity. With Rebus, the whole thing was only a few hours total. My animation was just under 3 minutes in HD, and the farm cost was under $2,000. When was the last time you rendered on a farm? With the changes in V-Ray to LC+BF - the frame consistency seems fairly reliable nowadays.

[particlerealities]

Last time I used a farm was 2011. That's also the last time we generated an animation because it was such a horrible experience and our clients just couldn't afford the expense.

[andybot_cg]

Last time I used a farm was 2011. That's also the last time we generated an animation because it was such a horrible experience and our clients just couldn't afford the expense.

That's interesting. I was in a similar boat - I had last done an interior walk-through animation in 2013, and it was a pretty lousy experience. All the precalculation woes and frame inconsistencies were there. I then did one last year, and the experience was so much better. I would suggest trying a few dozen test frames and see if it is easy enough or not. I think the benefits will far outweigh the cost, especially in terms of saving your time.

[anthonyh]

The last(and only) time we tried Rebus we crashed their farm apparently. We had problems using movies as textures in max so were using image sequences and their upload process went nuts trying to cope. The operator rang us up and wasn`t very pleased. Fun times. We use a small internal farm, it`s coped with everything we`ve needed for years although it`s getting a bit behind the times what with Vray wanting us to stay faithful to BF/LC.

[andybot_cg]

The last(and only) time we tried Rebus we crashed their farm apparently. We had problems using movies as textures in max so were using image sequences and their upload process went nuts trying to cope. The operator rang us up and wasn`t very pleased. Fun times. We use a small internal farm, it`s coped with everything we`ve needed for years although it`s getting a bit behind the times what with Vray wanting us to stay faithful to BF/LC.

Ouch! Impressive, didn't know that was possible
I think it's still worth a test, and there are other farms that are worth looking at too. I did have to spend a little bit of time fiddling with my model and some references, but the time savings overall were quite worth it.

[particlerealities]

Interesting turn of events - I have an email sitting in my inbox this morning from the client. Seems they decided to build a model from scratch using their latest design standards, instead of simply incorporating a few updates into our old model. So, I will be modeling from scratch and have the opportunity to optimize the scene for GPU from day one if I decide to render with GPU. I think I'm still leaning towards rendering internally since I'm not up against any hard deadlines. Now the question is - do I upgrade my CPU (add a second processor to my dual-capable Dell T7910), or squeeze in another GPU? Guess I should start a new thread with an "upgrade" related title in the appropriate section.

[anthonyh]

Would a separate render box be a better option, you can use DR or have it rendering frames while you`re working ? Assuming you have the render node license of course.