If you had to estimate...
polycounter lvl 11
I'm really interested in how long it takes some of you guys to complete a project and how you go about estimating how much time it will take you.
I know in order to estimate something you need specifics so...
If I were to ask you to do a Nex Gen (Normal Mapped) model of the Queen Alien from the alien movies and have it textured and ready to be rigged/exported to Unreal how long would you estimate this project would take you?
I know this model is very complex, but that's the point, given a project that has many unknown variables and potential new challenges how exactly do you go about estimating how long it will take?
Any other other examples from projects you have worked on in the past are welcome as well. I am really just looking for general advice on scheduling and trying to see what kind of standards/thought processes there are regarding this subject.
I know in order to estimate something you need specifics so...
If I were to ask you to do a Nex Gen (Normal Mapped) model of the Queen Alien from the alien movies and have it textured and ready to be rigged/exported to Unreal how long would you estimate this project would take you?
I know this model is very complex, but that's the point, given a project that has many unknown variables and potential new challenges how exactly do you go about estimating how long it will take?
Any other other examples from projects you have worked on in the past are welcome as well. I am really just looking for general advice on scheduling and trying to see what kind of standards/thought processes there are regarding this subject.
Replies
- Concept artist concepts an over arching image of the environment as a whole.
- From that concept, we dissect it to come up with an asset list of objects needed in order to achieve said environment.
- That list is prioritized from "BIGGEST IMPACT" to least impact.
- From there each artist puts down how many days they believe it will take them to model, texture, create damage states, setup for Havok and/or rigging for each asset they'll be working on.
- We then begin to work on our assets from the biggest impact to the least impact.
Generally its 1 artist per environment. And by artist I mean someone like me who pushes polys & pixels. There's generally 1 other person on the environment with us who's job is to dress the world, work on terrain, and get started on lighting.Hope this helps.
http://www.gameartisans.org/forums/showthread.php?t=8592
consensus ended up being about 2-3 weeks. Depending on if you can kit bash.
It's really interesting to hear about the process at relic. However, I am also wondering how someone would respond in a specific situation where they are facing a large project that has many unknowns and they need to provide an estimate, and what exactly that estimate would be.
I suppose my question is as much as about the process as it is the actual time it takes people to make art. There are so many parts of the process that consume time and make a project take longer, and I'm just curios what peoples experience has taught them about scheduling and what are some reasonable expectations for some of these complex projects like non-human creatures.
This seams to really become a particular issue when your doing freelance work.
@Dekard: Thanks for the link that pretty much covers what I was asking
But if anyone else has any other tidbits of wisdom they wish to bestow please don't' hold back ^_^
Concepts > Discussion > Asset list > Estimates > Prioritize
Only difference with Turn10 is the nature of their environments means more people are required to work on them. Usually it's the main lead, working with 1-3 production artists in-house, then off-site teams (no clue how many artists they had). Though the leads are often in charge of several levels each, so they don't get to do a ton of work on any one environment.
3-4 weeks.
Highpoly: 1 work week
Lowpoly/UV/Bake Normals & AO: 1 work week
Texturing: 1 work week
I say these times because of the following - an asset like the alien queen doesn't really have any hard surface stuff which can be more time consuming than organic surfaces to model (since you could do this all in ZBrush, potentially even start from a base mesh of zspheres to save more time). I would maybe up this estimate to 2 weeks if it was a "hero asset" or intended to be used in close-up cinematics etc. since you'd need to put in a lot more detail and make everything really clean.
A week for all lowpoly work, then UVs, and finally baking good normals and AO is maybe a little high, especially if you start from your "base mesh" and optimise it down. It would depend on how well your bakes went - assuming you didn't run into any issues this step could probably be done in 3-4 working days. I'm basically rolling this up with any buffer time to fix unforeseen issues with either the highpoly or lowpoly at this stage.
Texturing an asset like this probably wouldn't take all that long purely because if you have a good bake of AO and normals, you can grab a crazybump pass for the highlights/shadows on edges, overlay a few simple noise/dirt textures to get a decent base over the flat colour, then start painting in detail where necessary. The alien queen is a pretty consistently coloured and detailed creature, so there's really not much variation to consider. Again the time estimate for this would depend on how detailed it needed to be and how big the textures were. I'm assuming around a 2048x2048 texture sheet.
Then again, that's just me. Some other artist might estimate the same and produce something twice as good as I could manage, or others might estimate a week more and produce something mediocre.
Edit: Also, doesn't really hold any "unknown variables" for me, so I feel fairly confident in those estimations, assuming you want a quality, triple-A game asset out of it in the end.
Revisions are where ya get killed on time frames though.
That's basically 6 working days... the alien queen... are you sure?
I could get a base mesh in 5 hours. Sculpt for 15. Retopo in 6 or so. Unwrap in another 5 or 6. Do all the maps in another 10.
Ok, so more into the 40+ side of things.
Still, to me, that doesn't seem unrealistic.
Still, to add a big "HOWEVER," I never get the awesome detail into my models that you get into yours MoP. Maybe I just don't spend enough time on my work.
EDIT: also, that's kinda if this was a personal project. If I was working for someone, you can bet it would take longer as I fix according to their notes.
Mop, 30 hour work weeks!? I want to work with you, hehe
Gav
2 weeks minimum for a decent work (included weekends and working more than 12 hours per day).
If you include revisions, tweaks... i'd say more time. And QUALITY = more time. A model done from a simple base mesh, can be done faster, no doubt. Zbrush does magic with an ugly box.
what the fuck!!! ive heard of companies that procrastinate, but a quarter of a week, yer shittin me
Also people almost always under-estimate.
Probably 4 weeks to be rigged and in-game and forgotten
I would probably try to achieve the same sort of thing at home, though, assuming I was a pure freelancer.
Edit: Also, since jocose originally said "ready to be rigged/exported", I haven't included any sort of rigging or exporting time in my estimates (as some others seem to have). I'm just assuming that I hand off a textured, normal-mapped lowpoly model.
It depends on the model! the alien queen has finished concepts and references. The highpoly doens't need any time consuming hard surface work and the texture would be pretty simple. Only one material, nearly black...
e: and no fur, I HATE fur/hair/feathers! The fucking cat at dominancewar took ages!
Interesting that most time estimations are about the same.^^
Exporting implies animations are ready to go as well? Or where you thinking as a immovable test prop?
btw aesir, off topic but google chrome thinks your portfolio is dangerous
"The website at www.artofscottjonsson.com contains elements from the site updatedate.cn, which appears to host malware - software that can hurt your computer or otherwise operate without your consent. Just visiting a site that contains malware can infect your computer."
I was planning to take it offline anyways for a much needed overhaul. I'll just light it on fire when I get home.
I was thinking ready to be rigged, but not yet rigged. So I guess an immovable test prop with the potential to be animated later. Because the rig is an entire other factor (that is dependent on the types of animations) so I didn't want to introduce it and muck up peoples answers.
So... Using any sort of normal method for estimating time(8 hour days, 5 days a week), your two weeks is actually about a month. And thats being generous, 6 hour days are more realistic as mop says.
=P
Quality takes time.
Between 3 and 4 weeks to have really cool looking model for me. But Once again, it's the Alien Queen, not a puppy...
Yep, it could be a month easily, but i think that's working on a studio, relaxed. Being freelance we are asked to make things faster, we work for hours.
This reminds me some studios (publicity) asking me to have a model in 2-4 days (they contact me in a friday, and they say: for tuesday as too late, can you?). If you can't do the job, job you lose.
Long time ago, i accepted a work, one model, one week (7 days)... and it was the worst experience in my life XD, full of stress and working more than 12 hours per day. It wasn't a normal mapped model, but it had its time, because the job was a detailed subdiv model for animation (teeth included) + sculpting for details + shading & texturing + morphs targets... ready for the TD/animator.
A really polished model could be easily more than 1 month of work. In Gran Turismo, it's a car... and they spent more than 2 months.
basically your 12 hours workday, 7 days a week is very dangerous because this is the maximal limit your body could handle on a long term contract (let's say 3/4 months) then, you burn out, start to do shit and lose everything as if you were a wow addict, one rule we were taught at my school about managing time is that a task always takes 20% more time than scheduled, even (especially) if they already include these extra 20%.
as mop said, that 6 hours thing is a safe frame and if you put that rule next to it, you'll notice it's respected, the job will probably take 8 hours, maybe 10 depending on how fast the guy is, how disturbed he is while working if a freelancer (groceries, wife, kids, friends, anything that can distract him)
Any chance could get some more estimates for working on environmental stuff in the work place. For example, a facade building for a first person game with a distribution of street assets near it. Wish I could be more exact, but anything environmental really.
4 days high(1 day base mesh, 3 sculpt?), 2 low, 1 uv, 1 bake, 3 texture, + 2 days slack = 13 days or 2-3 weeks
Personally i think its doable in 2 weeks if you're focused. That is of course under perfect conditions without needing major reworking for the client. Otherwise the timeline of about 3 weeks most people are using is very safe/accurate.
Not sure of the topic, though. Are we talking about a rushed production with tight constraint conditions or a more relax/quality-focused one ?
As a person who just started freelancing I was wondering how long it took people to do things, and I would guess around 3 weeks also (give or take a few days) for me. I would probably spend nearly half my time sculpting, but the texturing would be pretty easy since the bake would give you most your texture details for an alien queen.
You know what she looks like, yeah?