Getting the most out of your VFX budget
If you were to ask most directors, producers and studios how to get the most out of their VFX budget the knee jerk reaction would be to hire the cheapest VFX company they could find. If you planned to build an elaborate house and invest a lot of time and money into it would you really want to choose the cheapest contractor who hired the cheapest workers and used the cheapest supplies?
Well no, they'd say they wanted something good enough.
This is all false economy. In an effort to save money it's not unusual for the final results to not only cost more but to look poorer.
Work smarter, not 'cheaper'.
It's not unusual in a major vfx movie to see enough money wasted to make up the difference of cheap to good vfx and those involved in the decisions rarely realize that.
Newsflash: Each vfx company, each VFX artist, and each VFX supervisor is different and will make a thousand different creative and technical decisions than one another. The concept that any place will do the same work and achieve the exact same final results is misguided at best. VFX does not produce a standard product. There is no commodity to be sold by the lowest priced distributor.
Rather than approach the issue as if it were simply a matter of where to buy the cheapest product the real question is how to get the most of whatever money and time is available.
In order to do this one of the key factors is to look at VFX as another department in the film making process. Treating it as a 3rd party process separate from the rest of the project only hurts the process.
If VFX were approached like live action there can be a number of advantages in terms of making decisions.
1. Figure out the shots needed to tell the story. Decide on the priority of the shots.
Look at each vfx shot as the equivalent of a whole new setup on a live action shoot. There's time and money involved for each shot. In some cases the shot may be the same as moving the entire film company to another location. Make sure the shot will be worth it.
If you're working on a graphic novel you only have so many pages and panels available so you make the most of each one. Approaching VFX this same way will trim some of the potential excess shots.
Example: 2 people talking while crossing the street will likely require 4 shots- wide, establishing - a 2 shot and a shot of each actor. Replacing one of the actors with a CG character or replacing the background should not require a dozen shots of zipping cranes, ant view shots, shots from the 42nd floor window, etc.
There will be times when it may be worth turning a scene into a visual feast and times when it's not necessary. Money and time will always be limited so choose when it's worth it.
2. Design the shots so they work with the rest of the film and make sure they tell the necessary story for that shot
If there is a gag or some action that needs to be shown to the audience then make sure the correct camera angle and timing are used. This would seem to be obvious but there are times where other angles and camera motions are selected that completely lose the reason for the shot. That's a wasted VFX shot. A shot that should have had a lot of power visually and narratively was turned into filler material and whistles right by the audience.
Example: If you wanted a shot of a literally flat character you wouldn't shoot the character straight on where the effect would be lost. A 3/4 view where the motion and twisting of the character makes their flatness obvious would be a much better choice.
If you had a skilled action choreographer such as Jackie Chan you'd want to avoid the shaky cam, quick cut, random angle editing that make up some action sequences. You'd need the audience to see the setup and the action clearly to make the most of it.
A good visual effects supervisor will work with you to design the shots for the most impact.
Some directors leave the design of big vfx sequences to their 2nd unit director. This only works if the director is part of the design process since the 2nd unit director will not be part of the post-production (unless they're also the vfx supervisor).
3. Use storyboards for your main VFX shots
For standard live action it's possible to shoot a film without storyboards but if there's any complex shots that require pre-planning then storyboards are a must. Storyboards are used to communicate visually. The director is able to make it clear what they want by a simple storyboard. Stunts, special effects, visual effects and the rest of the team now have an understanding of what's desired. Miscommunication is a very costly mistake. VFX also uses the storyboards as a method to budget. Without them the shot may be budgeted higher than it should be or lower than it should be. Any errors in the budget need to be avoided, especially if the budget is tight.
Concept design of what new creatures, worlds or other new imagery will have to be done to allow the team to know what the final target is to look like.
More complex scenes can be prevised to capture the timing and action in more detail than a storyboard can convey. Just know the level required for the previs to avoid tweaking them beyond what's necessary.
4. Hire a good vfx supervisor early in pre-production.
To save money it's not uncommon for independent films to bring on the supervisor at the start of shooting or not hire a supervisor and turn all the footage over in post to a vfx company. This usually turns out to be a costly mistake.
This loses any possibility of taken advantage of the experience of the supervisor to help design the shots. It loses the ability of the supervisor to flag potential ways to save on location shooting or set building. There's no prep available at that stage and no planning so the likelihood of getting the best footage for vfx is lower.
Bringing footage to a vfx company with no supervision is like bringing a cassette recording someone happened to record during filming to a sound mixing company. It may have seemed like a money saving approach not to hire a professional sound recorder to be on the set but the result is a likely to be a very expensive and time consuming ADR process that would have more than paid to have the correct person there to begin with.
The DP may have done a VFX project a few years before or the director may have watched the making of segment of a DVD but neither of these can make up for the expertise and experience a good vfx supervisor brings to make sure the film is shot correct for the vfx required.
End result: you've just spent far more than a supervisor would have cost to begin with. And you will have lost much bang for your buck.
5. Hire the right supervisor for the project
That means meeting with 3 or 4 supervisors. The supervisor will likely be involved longer than either the DP or the Production Designer since the VFX work starts in pre-production and goes all the way to the release print. On a vfx heavy project do you really want to select someone that will be that involved based only on a resume? The studio may want to make this decision but the director and producer should certainly both be involved as is likely in the other key roles.
You want someone that can speak to as a creative lead and won't need a technical translator. You'll want someone you can trust with your vision and getting it done.
Some directors have a fear the vfx supervisor is going to be doing their own thing. The vfx supervisors are there to serve the director and the story - same as the other creative keys. The other fear at times for a new director is to work with 'old timers' who have real world experience. This can also apply to their view of DPs and Production Designers as well. As a result they hire someone relatively new like themselves who like the same band or football team. Tip: Hire the best people you can.
When reviewing the resume it's worth noting the supervisor credits more than simply the number of films. Are there a variety of projects? Don't get caught up in the notion that the supervisor hasn't done exactly this specific picture before. Realistic versus non-realistic effects plays less of a role here than imagined simply because the same techniques and approaches are used. The only difference is the reference the vfx team is trying to match to. A good supervisor is flexible. On the flip side I wouldn't suggest a commercials supervisor for a 1500 shot feature film or visa versa. In these cases it's a much different beast with different creative workflow, schedules, budgets and approaches.
A vfx supervisor assigned by a vfx company may be based more on contracts or budgets than who is right for the project. One other issue with company supervisors is they find themselves trying to please the director on the one hand and on the other being pushed by the company to sell the shots as finaled. It's not an easy position.
6. Minimize ‘fix-it’ shots
Most films these days require some amount of vfx or post work to fix a shot or two. A boom mic in the shot or blurring out a sign that didn't get cleared by the lawyers. However the effort should be made to keep these down. If the supervisor or vfx producer suggested adding $2 million dollars for fix it shots the executives would think they were crazy. And yet that has happened and continues to happen. Make sure to occasionally check the dailies at high resolution (film or HD) on a big screen. The use of Avids means that that wig netting, makeup issue or other small problems are missed until late in post-production. The extra 5 or 10 minutes to finish the makeup or other fixes o nset can save enormously in post.
7. Shoot efficiently - no more, no less
During filming of vfx scenes the supervisor will likely need to get measurements and references. The vfx crew will be prepped to do this as quickly as possible. The supervisor knows how much it costs to shoot on the set but the loss of information on the set would mean a large increase in the time and money to the shot later and more importantly could jeopardize the quality. These will be the balance the supervisor will have to determine during shooting. And just as a stunt coordinator will flag a dangerous stunt or a DP will flag an expensive setup, the supervisor may do the same. You've hired them for their expertise so you should take advantage of it.
8. Budget time and money - production versus effects
Productions frequently deal with the live action budget differently than the vfx budget. This disparity in views and accounting can cause problems. If a big set doesn't need to be built because it's determined vfx will be doing it as a post effect, it can be difficult to move the allocated budget to vfx. This can also be a issue going the other way.
The other problem here is that to save live action time and money entire shots or sequences may be pushed onto vfx during shooting. That's fine if the time and budget can expand to accommodate it but frequently the budget doesn't move to vfx and the post-production time is likely to already be too short.
9. Techniques
It's easy to get caught up in discussions of physical models versus computer models or other issues but most vfx shots can be done using multiple techniques. Which one is very dependent on the particular film and what its requirements are. Rather than force a specific technique to work on a project it's usually best to select the most appropriate one and modify as needed.
10. Avoid having too many vfx companies
A frequent approach to saving money is to parcel the work out to multiple companies. Likely also due to time schedules. This can all make sense when selecting specialized companies but at a certain point it provides diminishing returns and can even cost more.
Just as it's useful at times to visit a couple of food markets to get the best deals and products, if you were to visit a dozen it's likely the extra gas and time would exceed any savings. The extra communication issues, mistakes and entanglement of shots can be problematic if the distribution of shots and the number of companies involved exceed a natural threshold.
11. Schedule the time accordingly
If a production really wishes to save money they may want to work with the vfx companies to determine a realistic time schedule. There is a tendency to try to squeeze this time shorter and shorter but the increased overtime pay (with no improvement to quality) means that you end up paying more for less, even from a lower priced vfx company.
12. Decisions
As with other departments, vfx requires a number of decisions. It's likely vfx will require even more decisions due to the flexibility and scope. However it's a myth to think that all vfx decisions can and will be done in post-production. Decisions on designs (such as creatures or virtual environments) have to be locked into in pre-production to insure that the work will be done. Delaying these types of major decisions until post means compromising all the original photography since there is no locked plan. Any design work will then be necessary to squeeze in the limited post-production schedule at full and overtime rates.
13. Changes
Changes are natural in the creative world of filmmaking. Frequently small changes in vfx can be accommodated very easily. However be aware that large changes are no different than changes in live action. If you’re on location and choose to shoot in a direction where no set has been built then that will likely be a costly new set. The fact that it may be done on the computer doesn’t necessarily mean it’s a lot faster and cheaper to change. If you’ve shot an entire sequence and have made a major change in the edit this may need a large reshoot (or in the case of vfx, may need tossing out a lot of work and starting all over again) The cost could well exceed the cost of a reshoot of a sequence.
(Posting written with Notes To Store for the iPad)
Update: I can't believe there are still student films that put up postings for vfx artists to work for free on something they already shot. (Just check craigslist.org any day of the week) I guess ignorance starts early. Hello. Pre-planning is even more critical when there's little or no budget. You could have made a great effect efficiently but instead chose to see if someone could take the bits you shot and make something with it.
Thanks for posting this.
ReplyDeleteI suggest every company adopt John Dykstra's model for VFX:
"There are three ways to do any shot. There's fast, there's good and there's cheap. But you can only work in combinations of two.
You can have it cheap and you can have it fast, but can't have it good;
you can have it fast and you can have it good, but you can't have it cheap;
you can have it good and you can have it cheap, but you can't have it fast. True words."
There are a number of variants on this (I don't think John originated it)
ReplyDeleteFast. Cheap. Good. Which 2 do you want?
There also used to be a cartoon many places had taped on their wall. It had a character laughing in hysterics and the title was "You want it when?"
Those were the days.
Scott-
ReplyDeletethanks for summing this up all in one place. If more and more productions thought of VFX as a department instead of a vendor the process would be much smoother for everyone.
I hope this gets forwarded around to many studios, producers and directors.
-David
Well put. Many others have come at this in a less rigorous or efficient way. Nice work Scott
ReplyDeleteExcellent article! I always enjoy your commentary and find comfort in your pragmatic, matter of fact perspective.
ReplyDeleteI completely agree with every point you make but would like to suggest that VFX facilities, more so than production, are the biggest offenders when it comes to inefficient use of resources. It is ironic considering the commoditization of VFX services in recent years.
You either have time and no money, or money and no time.
Time. Production schedules continue to accelerate as does the demand for quality and innovation. Proprietary workflow technology remains relatively unrealistic for many facilities. But equally as important as the VFX being created, is the pipeline supporting it. Many facilities fail to capitalize on the savings that efficiencies within the core business can provide. The value that development personnel add to VFX production continues to be seen as removed from the bottom line. An assumption that could not be further from the truth!
In 1998 my colleagues and I spent about three months tracking the amount of time we spent on specific tasks; creating directories, maintaining naming conventions, etc. We found that roughly 38% of our time was spent on things unrelated to the skills for which we were hired!
Since then i have exploited every opportunity to alleviate waste within production workflows by creating pipeline policies, procedures, and toolsets that empower artists, alleviated redundancy, and streamline resource allocation and communication.
“Envy,” as it was referred to by our staff, enabled 10 artists over a three year period, to complete an average of 2.8 shots per production day, per artist: amortized over that same period.
The cost of hiring Developers is quickly offset by the benefits of pipeline automation.
Money. Just as important as hiring the right Supervisor, facilities need to practice a similar concept and hire the right artists. Too often production schedules are front loaded with less expensive, less experienced artists in an effort to meet what later proves to be an arbitrary, disconnected, and under bid project. Like a dog lacking Pavlov's response, management increases staff with highly skilled and experienced,and (expensive) artists in an attempt to get the project back on track. Overtime is merely diminishing returns with exponential cost.
Find the best people you can. Assemble a team of experts in their respective discipline. Maintain a manageable team size with respect to a project's scope and estimated COGS. You will be amazed! I have been practicing what i preach for several years now and our costs have come within 10k of original COG estimates. Furthermore, we have adhered to a policy of "never more than 12 in a day, never more than 6 in a row" even during crunches.
You will find an increase in the quality of your work (and your lives), greater overall productivity, as well as have a fighting chance at making a profit (if you think 5-15% is a good profit margin).
My two cents...
Ciao'
Good points. In addition to the driving force of the filmmakers there can be cost and time savings within the vfx companies that at least they can control.
ReplyDeleteOne example is a proprietary system that is brought up to just a ‘good enough’ level. The company is likely to shift the team to other more sexy projects. Meanwhile the current system could save each artist an hour a day if it had one extra feature. It would seem to be a no brainer that saving an hour a day for each of the 30-50 artists would make it worthwhile to have a developer or two continue to work on it. Unfortunately these no brainer decisions are sometimes unable/unwilling to be made by the vfx company.
The other thing is all vfx artists have to be aware of is their own time and effort. It is very easy to lose focus on the important items in a shot and get caught up with the ‘fun’ items in the shot. VFX shots are like fractals - you can continue to work on finer and finer details endlessly. When focused on the pixels of the current shot it’s very easy to forget the bigger view of the sequence.
Is the take you did today a big improvement from the version the day before?
Will the director or audience notice?
Will the 100th smoke element make a real difference?
Is it worth re-rendering the whole shot for a small glitch?
On a 1 second cut is it worth spending 3 man days (animator, compositor, lighter) on a new variation?
Should the plug be pulled on a given technique now and plan B put into place?
Are you polishing something that should be redesigned or redone from scratch?
Are the models being over built and over painted?
Is the software being built to solve every problem, no matter how unlikely?
Are you trying to create something that could have been filmed faster and easier on a consumer HD camera than developing new software and spending days letting it run?
Time is the real gold to work with in vfx and the less that can be put into the shot (and still make it work), the better.
In an effort to do it the ‘right way’ with full simulations and dynamics it’s easy to forget that we’re illusionists. We can cheat in many cases (if it’s at the right point in the pipeline). Just as productions can get away with constructing just the facade of a building we can get away with simplifying.
Hand painting pixels has been used to save a number of render glitches.
Roto has saved a number of poor greenscreen shots.
A strategically placed element can work wonders on some shots.
If production focuses on the shots that need to be done and the vfx company and artists themselves constantly re-evaluate their approaches and streamline wherever possible, there can be significant savings.
Could you possibly post some examples of what you think a good VFX company -to- Production company contract looks like?
ReplyDeleteFor example one can find a standard photographer's "Model Release" form from many sources, but I've never seen anything even remotely like such a basic form for visual effects companies.
By creating a basic shell or kernel, (or simply posting some rough examples), it would be helpful for readers to move towards a common standard. Even producers and production companies reading this might be better prepared for what is expected of them as well.
And a sincere thank you Scott for your insightful writings!
Contract example - Unfortunately I'm not a lawyer (and I like not being a lawyer).
ReplyDeleteThis might be an interesting project for the VES or even better would be for the VFX Trade organization (if that were to happen).
I think there are various film production contracts online but would suggest at this time seeing an entertainment lawyer. Getting it right is important and a good lawyer who specializes in this should be worth their money. Some vfx companies have sales reps which may have a good source of knowledge.
Some of the issues (I think I cover some of this in budgeting posting):
1. Deliver schedules and deadlines- Production turnovers to VFX company. Shots or sequences from the vfx company to the production.
Penalties? If production turns over a sequence a month late how do you make up for it? (extra time, money,etc) What if the director doesn't buy off on the concept art before filming starts?
2. Credits - how many, where, what type.
Company credit. Credit on poster or other materials?
3. Payment schedule - weekly? monthly?
Some studios want 90 - 120 day payments. You don't want that.
4. Milestones?
5. Budget and what the budget is based on. Which storyboards? Which previs?
6. How are changes dealt with? Change orders? Who determines when a change order is required and what's the approval process?
7. Demo material (photos, footage) approval for the vfx company. Many times you can't show what you're working on. If it's dependent on the films release what happens if the release is delayed a year or never released? It happens.
8. PR approval - can you state you worked on the project? Can the company run ads? Articles?
9. awards - It's up to the producer who they put up for most awards but you wouldn't want to do most of the work and not be listed somehow.
10. Temp screenings - Are those in the budget and schedule?
11. What if someone at the company releases photos/footage to the outside world?
12. Who owns the models (physical, CG)? (typically the production company paid for it as a work for hire but there can be variations on this).
13. Who owns the additional element footage? If you spend a day shooting smoke elements does production own that or is it now part of the stock footage? (i.e. does the contract only specify the final images or do they own all resources to make those final images)
14. What are the technical specs? (i.e. how are the images/footage delivered to the company, for the DI from the vfx company, for editorial, etc)
15. What's the coverage for location shooting? (plane fares, room, board, per diem, etc)
16. If the company has bartered money for points in the project then there's a whole new level of terms and conditions to cover gross points, etc.
I'm sure there are dozens of other key issues I'm not even thinking about. As mentioned a good entertainment lawyer (get someone who specializes in these types of things) should be hired and you should rack your brain to include as much as possible. I hate to say it but assume the worst from the people you're contracting with. (i.e. don't assume they'll do the right thing from your perspective if it's not in the contract)