How to Bill for Render Time

[Shibui Design]

Looking for some advice here.

 

Over the past year I've been utilizing the Maxwell Render plugin to render still images for my clients. As a FormZ user for 17 years, I've gotten pretty adept at creating high quality Renderzone images. Clients are expecting better quality these days, so I've turned to this route. My struggle is, since I can produce nice images using Renderzone with very short render times (never more that 10-15 minutes), I've never charged separately for "render time." I will define render time as the time the computer spends developing the images. That time was simply tracked with my active design time and then billed accordingly. Now since I'm using Maxwell and render times can be anywhere between 3 and 8 hours each. I've purchased a very speedy machine just to render out images, often overnight. I know I could turn to a render farm, and I have on occasion, but I don't want to have to depend on this if I can do it myself effectively.

 

Does anyone have any advice on how I might charge for strictly render time? Any thoughts would be welcome. Thanks!!

[jonmoore]

Charging for render time is less important than charging accurately for creative development and change requests. I've often found that change management can be the difference between a healthy profit or a nasty loss.

 

Educating the client about the increased complexity of physically correct rendering workflows is vital too. And be transparent, clients often think they're making relatively minor change requests but these might necessitate length render times and more importantly scene set up changes.

 

If you mainly deal in still images there's little value in providing a 'time & materials' style breakdown of render time. However if you're rendering animations (which are usually only feasible via a farm/cloud solution) it's often best to work on a 'time and materials basis' for those occasions. A reasonable markup on what you are being charged usually suffices but be sure to let your client know animations usually involve multiple takes/test renders etc.

 

Hope this helps a bit.

[Shibui Design]

Thanks for your input Jon! The true value I bring my clients is creative and this is my focus. Renderings and drawings are simply the output of creative, in that they are used to communicate the design intent. One can produce a very low-quality rendering that is high in creative merit or the inverse, in which case one might best focus on simply creating nice renders. My trouble at the moment is getting this output takes me far longer utilizing other render methods, Maxwell, in this case. Do I simply not charge for the time my hardware and software are processing in the background and only charge for the time I invest in developing the creative? I struggle with this scenario, as it is not "free" since I pay for hardware, software, electricity and oftentimes, me managing the queuing of render jobs. Furthermore, I need to recognize the lost opportunity to deliver more creative (and income) if my time is spent wrangling long renders.

 

Curious how others might manage this.

[Andrew West]

I too felt the need to step up my game on render quality.  Especially when it came to interiors with glossy surfaces and lots of refraction.  So about a year ago I started using Maxwell in the hopes that it would help.  The transition was not as expected.  The additional time needed to both set up these renders, convert materials, play with the lighting and cameras and then painfully wait for a very low res Maxwell Fire preview was a little surprising.  It eventually led me to purchase a $7000 dual xeon 20 core processor machine in the hopes of speeding up the process.  It did not help.  As a result I often turn to a render farm  to process my images in any kind of a timely fashion. As an example I did an interior last week that would have taken 67 hours on my machine and Rebus Farm did it in 20 minutes for $200.  Because of this I tend to write into my contract a set amount for render farm services regardless of whether I use them or not.  I explain that these fees are in addition to any modeling or creative time and any changes made after the final render will incur additional render farm fees.  If I can use my machine then I make more to cover the cost of my equipment.  If I use Rebus I charge their fee plus the additional hour or two to set up the job, submit it and then re composite it into the final.  And of course you have to charge for any additional modeling time for changes.  This is where your profit or loss is really significant.  I work in architecture so I explain to my clients that changes to the model are exactly like changes to the final building.  A contractor charges for design changes against what was originally bid and so do I.    It really puts the responsibility of making proper decisions up front on the designers and owners.  If the design turns out to be a work in progress then I charge hourly for changes and renders until a design is finalized. 

One other piece of advice is to communicate constantly with your clients about changes and costs.  To all parties involved!  It is all too easy to just want to work in an isolated  void creating pretty pictures.  Take the time to make sure that you know who is making the decisions and who is paying the bill and make sure they are communicating with each other.  Then every change has to have a cost associated with it.  No matter how minor. I keep a ledger by my keyboard and write down every model change and its associated render time.

On a similar thread I looked into one of these overseas illustration studios that boasted that they could produce top quality images for $195 with a 48 hour turn around time.  I called them and grilled them on each image in their portfolio about what the actual cost was.  As it turns out every good image they had cost many thousands of dollars due to model complexity,  render times, render farm fees and of course design changes. 

I hate to think that I nickle and dime my clients to death but the opposite scenario is death by a thousand paper cuts.  

[jonmoore]

Thanks for your input Jon! The true value I bring my clients is creative and this is my focus. 

 

This is actually the crux of my point. I have a design/advertising agency background and everything in that world centres on the perceived value of your creative output. The final production output may be a scan of a charcoal sketch but reaching that final creative deliverable is more often than not the result of many weeks work. Putting too much focus on the costs of physical production can take away from the true value you provide.

 

Andrew has some sage advice for those clients that do attempt to nickle and dime you on production costs. But just be careful those production costs doesn't become the focus of the discussion.

[pylon]

Andrew,

What is the Geekbench 64-bit multicore score of the rendering computer? (That's really the only number you need to look at; forget number of cores, clock speed etc. when comparing machines worthiness for Maxwell rendering.) http://www.primatelabs.com/geekbench/

What is the resolution, in pixel dimensions, of the renderings?

I have never had a Maxwell rendering take nearly that long, and am curious if there might be something in your materials, lighting, or scene parameters that is causing it to take longer than necessary.

[DennisA]

Right or wrong I do not include rendering time as part of project costs. Saying that currently all my renderings are done in Render-zone (generally takes anywhere from 50 seconds to 5 minutes) & I am on the clock working on the project in the post. A giant pixel size rendering I will bill a premium because more time is spent on larger maps & any post enhancing (I will let this render overnight or in the imager).

On a 2nd note - at one point in the past when I used Maxwell (earlier versions of fmz)  it changed my Render-zone lighting, effecting the final rendering - has this been fixed & not a current issue?

[Andrew West]

My final renders are usually about 4200 pixels wide.  I have studied all the available information online that deal with  materials and lighting optimization.  As a result I minimize my material layers, never use displacement (unfortunately), but I do use emitters and fake light planes.  I also never use physical sky and just use a simple light dome for the exterior.  I do turn off multilight (unfortunately). 

I am open to any suggestions you might have to make this faster

 

my geekbench scores are

Single processor: 2651

Multi processor: 28811

[Martin Malinski]

Interesting question.

 

I also work in Architecture and Interiors.

 

All my 'work in progress' is presented via tablet, laptop, projector.... never print.....so I only need to render at screen res. (huge time saving)

The only boards I provide are sample and mood.... for tactile purposes.

 

The ONLY time I render for print is for final documentation. (unless the client specifically requests otherwise, at an agreed cost)

 

I submit final documentation, standards and manuals (print ready) costs separately from design development. (render costs are included within this proposal) As at the production stage it is much easier to plan and provide fixed costs rather than trying to impose an unknown within the design development stage.

 

This allows for much greater flexibility and speedier turn around.

 

Martin

[Justin Montoya]

If we are talking about Maxwell, shouldn't we refer to it's own Benchwell benchmark for comparison? - http://www.maxwellrender.com/benchwell

 

Andrew's computer has a very high score, so I suspect his computer is more than adept, and that Maxwell may not be the right tool for his purposes, but without other options it just has to work for now.  RenderZone is great for what it is, but unfortunately the lack of updates means it's no longer cutting it for many professionals.

 

I tried Maxwell as well and it is just too slow on my 3.46 ghz 12/c 24/t Mac Pro.  I tried using very simple lighting techniques suggested online to no avail (dome and emitters only).   I have to make many client revisions day by day, and I cannot afford to lose my computer while Maxwell is rendering for hours and hours.  FYI, I primarily render at 17x9.5" @150 which has proved to be a good compromise for digital or print communications.  

 

My main goal with another renderer other than RenderZone is emitter lighting.   LED's are everywhere and people want glowing details, graphics, logos, etc.  Maxwell does this, but at a speed cost I just can't justify in my workflow.  When not using emitters, the renderings from RenderZone are very comparable but so much faster.  If RenderZone did progressive rendering and emitters, I don't think I'd be looking at other options so closely.  But as is, it's dated, doesn't scale up to modern workstation power and is not evolving with any new features to stay current.   I avoid post production touch ups as much as possible, since it means that any revision would require the same tedious touch up work.  It works for our business, but I understand that may not work for everyone.

 

Currently, the only time I charge extra for Rendering time is animations.   I typically set them up to run at the end of the day for a simple fly through.   The way I look at it, my workstation is an extension of me as designer and if it's busy working for me, I bill for it.  It reminds me of a Bob's Burgers quote from Teddy "My therapist said that if I don't value my time, no body else will."

 

For those sticking with Maxwell, and using render farms, have you looked into the new Compute Clouds from Amazon and Google?   Basically you are renting computing power via an IP address. 

Example of farm using their cloud servers - 

 

http://www.rendicity.com/ 

 

https://cloudplatformonline.com/RenderMore_Homepage.html

 

or you can actually setup your own virtual farm for a fraction the cost of a typical render farm, so they say:

http://cosmocyte.com/animation-techniques/

 

I don't have any experience with this new cloud based rendering, but the theory is sound.  Way less overhead than hosting your own rendering farm, and you just pay as you go!  I would love to do some experiments with this once our other rendering engine options get better...

 

 

I do have experience setting up a couple of rendering farms and it could be worthwhile for you if you prefer to manage everything in house.  There's the computer costs and the electrical costs involved, plus the overhead of managing the systems.  It's definitely not for everyone, but if you like that sort of thing, you can get very reasonable 8c/16t workstations off lease on eBay for under $600 shipped.  Buy as many as you can afford now, and you can always add more.   There's also blade servers if you have room for a rack and a proper place for it where the noise and heat won't be a bother.  Again, there's the other costs involved here, but if you are constantly sending out to render farms, this might serve you better.

 

Deep down, I think I'm a computer geek with some design ability, so these types of discussions are right up my alley!  

[pylon]

DennisA,

 

I don't recall any issues with Maxwell influencing Renderzone lighting off the top of my head. Maxwell only "reads" from the basic formZ Light types (Point, Spot, etc.); at no point has it ever modified them. Maxwell light types can be created (Physical Sky, etc.), but Renderzone should ignore these. Can you refresh my memory, or provide a link if you reported this previously?

 

Andrew,

 

Just curious: Are all of your renderings for large prints? 4200 pixels would be 42" wide printed at 100 dpi.  A 2k rendering would take 1/4 the time... I usually render at around 1500px for screen/projector use.

I am happy to take a look at one of your projects and see if any optimizations can be made. Just dropbox (or equivalent) the formZ project file and all dependencies to

[DennisA]

Pylon... tracking down correspondence with formZ was from 2014... I will give Maxwell another try & see if I have any issues. Had support confirmed with John & email with Paul. If you give me an email I can email correspondence to you- being it is 2 years old it may not be relevant on current versions.

[Andrew West]

Pylon

The reason I render at 4200 wide is that occasionally my clients like to print job signs for the construction site.  These tend to be about 48" wide and they are often displayed at eye level.  Ideally I would render it at an even higher level.  I am also routinely asked for high res images for magazine publications. 

 

Aside from this I have received a test file back from Pylon that has optimized my render times back to something very reasonable.  Essentially my render times went from ~60 hours to ~1 hour.  See this post http://forums.formz.com/index.php?/topic/3716-reflections/?p=13004 and download the test file to see the Maxwell settings.

[pylon]

Hi Andrew,

Thanks for posting back on your Maxwell success. I love a happy ending.

 

Dennis,

Address is in the post just above yours.

[Shibui Design]

Aside from this I have received a test file back from Pylon that has optimized my render times back to something very reasonable.  Essentially my render times went from ~60 hours to ~1 hour.  See this post http://forums.formz.com/index.php?/topic/3716-reflections/?p=13004 and download the test file to see the Maxwell settings.

 

Wow! what are the secrets for going from 60 hours to 1 hour?

[pylon]

Envdesign,

 

Andrew linked to this thread in his post; here it is again:

 

http://forums.formz.com/index.php?/topic/3716-reflections/?p=13004%C2%A0