PDA

View Full Version : Modeling, fast or slow?



Talc
07-16-2008, 05:58 PM
I have little comparison to know weather im slow or fast at modeling other than the deadline for the job. Is there a benchmark of speed for how long it should take to create something.

It's been years since i last worked with lightwave & im only just getting used to the tool set. Each new job is creating unique challenges that is forcing me to go back to basics but im sure that once i've fully re-equainted myself im going to speed up. I know there's always a pressure to find a faster way but any ideas off what i should aim for?

*Pete*
07-16-2008, 06:11 PM
i usually modell as fast as i can, and if i still have time when everything is ready to render i return to the modell to add details.

how long a modell should take is very variable from object to object..if you head over to the wip forum you will find my excavator there, i have spent close to 20 hours of modelling of it by now..but im in no hurry since its for my own demoreel and not for a client.

if in a hurry, real hurry..you could be able to crank something like that up in one day.

as for benchmarks..its all about the deadline given to you by your client.

Talc
07-16-2008, 06:15 PM
Any chance you could put a link up because i would like to see it!

Cheers

*Pete*
07-16-2008, 06:32 PM
http://newtek.com/forums/showthread.php?t=86316

its slightly updated now, im preparing it for animation.

anyway...to further answer your question on how long it should take to modell things...always add more time than you think for yourself when negotiating with clients, you never know when you hit a problem that is difficult to solve.

like in my case, i was asked to make a house built by empty plastic bottles...fine i thought, easy job..modell a bottle and copy it untill you form a shape of a house with it.
but what i didnt count on was the huuuuge polycount and the need of maximum amount of refractions...i was able to deliver, but i still today feel bad over the quality of it, if i just had one more day i would have done a better job.

so, if you think you can create something in a day or two, make sure you get the double time for it and push hard from the very first moment to finish it..if you were correct, you are ready 1-2 days in advance, if not then you didnt waste any time to beginn with.

Talc
07-16-2008, 06:51 PM
Im impressed:) Like the molding on the tyres. (Im finding curved geometry to be tricky & all to easy to break)

You say that's 20 hours. How much thinking time & did you work it from photos, etc..

Brunson
07-16-2008, 06:54 PM
I've been working with Lightwave for about 3 years. I'm not sure exactly what amount of time is reasonable to model a given object, but I know that whatever it is, I'm wayyy slower. I have a human head model in the WIP section that I've been working on since Friday. Probably have put about 35 hours into it so far. Still have a lot left to go. So, if you feel slow, take comfort in knowing that there's others out there that are VERY slow. Part of my slowness is a result of something good, however. I'm a perfectionist. Most of my time is spent refining very small details on my models so that they end up as close to perfect as possible. The downside is that I get so bored spending so many hours on a particular model that I give up and move on to something else, never finishing anything.

When it comes to determining what you should charge, though, I think of it like its the United States Postal Service. There's Express Mail, which is fast, offers a delivery time commitment, and is pretty expensive. Then there's Priority Mail, which is moderately priced and fairly quick. Then there's Parcel Post, which is slow and cheap. Me, I'm like an illegal substance sent via Media Mail stopped by customs for inspection.

Talc
07-16-2008, 07:09 PM
I should read further down threads before posting, sorry *Pete* yes you did use photos. However how much prep work or even false starts did you have to go through?

I think sometimes i get hung up on poly flow & not wanting to break the geometry even though i know what comes out in the render is more important than looking at a wire cage.

Surrealist.
07-16-2008, 07:23 PM
Well a wire cage is like a great indicator to go by with what you are modeling. Healthy wires mean your model is cleanly built and lean which is good for production.

Talc
07-16-2008, 07:30 PM
I know what your saying but for a while i got stuck with the idea that all polys needed to be quads which is not true with catmul. But yes good design should be reflected in the wire frame:agree:

*Pete*
07-16-2008, 07:33 PM
I should read further down threads before posting, sorry *Pete* yes you did use photos. However how much prep work or even false starts did you have to go through?

I think sometimes i get hung up on poly flow & not wanting to break the geometry even though i know what comes out in the render is more important than looking at a wire cage.

prep time was short actually..i wanted to do something like this for some time and had a futuristic cargo lifter propelled by jet engines (i shot photos of an abandoned jet engine a few months ago) in my mind..and i had been drawing sketches trying to figure out exactly what i wanted to do..and then one day after work i walked past an wheeled excavator, ran home, grabbed the camera and the same evening i was already modelling it.

i didnt have no false starts, but a lot of revisions of the stuff i thought i was finished with.

as for polyflow..trust me when i say that it looks horrible beneath ;)
parts of my object is subdivided, parts are pure poly stuff.

what slows me down with this project is the animation part, becouse i havent animated in a few years..its wonderfull to explore that part of LW again, but it goes slowly.

*Pete*
07-16-2008, 07:36 PM
one more thing..since we are speaking about polyflow and subd modelling...get tarons video on organic modelling!!

helps a lot, speed up a lot..very intiutive, very fast...you feel totally in controll over what you are doing.

Surrealist.
07-16-2008, 08:16 PM
I know what your saying but for a while i got stuck with the idea that all polys needed to be quads which is not true with catmul. But yes good design should be reflected in the wire frame:agree:

Yeah. Also realize that CC's produce a different math on the subdivision. Take a look at the beginning of my tutorial if you want a clear explanation of it. And also be forewarned that there are some potential drawbacks production-wise to CC's. I just went through a whole slew of my models and changed them all back to quads because of it. Simply put, when doing certain APS functions, CC's will crash - not because of the higher subdivision - because it has been tested at similar resolutions. I am not sure why this is but it is. And it is something to be aware of. There are a few other things too I think with different layout tools. Some of them have been corrected, but none the less it is something to watch for. It may not be a factor for one model but the higher your polycount and or the number of models the more problem it presents.

In short, to be safe, model in quads. That way you are covered no matter what.

*Pete*
07-17-2008, 08:12 AM
Like the molding on the tyres. (Im finding curved geometry to be tricky & all to easy to break)


modelling is like cooking, you know a number of recipies that help.

previously i also had problem specially with tyres..there are a lot of different methods to do this, but i found out that the array tool was superb for my purposes.

first i lathed a tyre shape, then i cut out 1/15th (or something) of it, used that small part to modell the molds, arrayed 15 clones of ot (circular), welded points where it mismatched, cut out the 1/15th part again, then array and merge points..done.

would i not know that recipy by memory/experience..id spend a lot of time to find out how to do it, thus adding loads of time for the modelling.

most of the other things ive done have been ever more basic, make a box, round it, stencil holes, extrude and cut holes...repeat untill done.

i also modelled the whole machine in the shape it is..the crane arms were modelled as positioned, i just put nulls as new pivot points for animation...its so much easier when you do not always need to modell after the pivot points, as i used to do before.

the only thing i modelled at xyz center was the tires becouse i didnt want to bother with changing the center of lathe and array tools...

so, modelling can be really fast once you find what fits you..there are no real secret shortcuts, but its more like knowing the next step of everything you do..as in my case, the pivot point used to be a timewaster for me..untill i understod that its not necessary to modell after it, obvious for some but it took time for me to find out.

in general..the more you modell, the faster you become..watching modelling videos helps a lot becouse you see more than just the tutorial.you see the small things that nobody would even consider to advise to you becouse they are so obvious..such as the pivot points ;)

Netvudu
07-17-2008, 08:43 AM
It also depends a lot on where your area of expertise lies. In my case, after 11 years+ using Lightwave, I consider modeling as that tedious thing that I have to do to be able to work in the fun stuff :rolleyes:

Talc
07-17-2008, 07:12 PM
"Surrealist" I've started reading through your ecxellent tutorial although i havent had time to complete it. It was great to see ideas explained with such apparent clarity. Although some of the illustrations involved sophisticated concepts that im not too sure if a beginner would understand without expanding?

I am reminded of the first communication exersize given to me "Explane to an alien how to boil an egg when they don't know what a tap, water & kettle are"

This is fantastic work, helping me to understand (in greater depth) how Lightwave functions & for that I thank you!

Surrealist.
07-17-2008, 07:30 PM
Talc,

Thanks for your feedback. I am definitely interested in what concepts you think may need better explaining. Always good to have that kind of input and I will gladly expand on it. :)

Good to hear you are getting something out if it. The very first parts are mainly theory and then after that there are more tutorial-based sections with steps and examples.

And of course if you have any questions let me know.

Talc
07-17-2008, 08:02 PM
My ISP must be playing up at the moment, i can't get back to check what i was about say. Please take this as constructive & I may be very easily wrong.

The illistration of increasing <> decreasing mesh complexity, whilst talking about 5 poly vertex is a big concept. I frequently struggle with controling the mesh, stepping up where detail is needed and reducing when not. As i said in and earlier post about getting hung up on quads & yes i do try to keep to the rule, it's a hard discipline that is easy to break.

My latest frustration is accidently creating full polys that look like two points that arn't detectable when you hit "W" & bug the hell out of you when you hit tab....

Surrealist.
07-17-2008, 08:40 PM
Yeah I can't get the lightwiki to show up either.

But I see what you mean. I'll see if I can notate that or something. It is not really intended that you would completely have a practical understanding of it at that point. More or less a general road map. I'll see what I can do to edit it in such a way that explains the context a little better. It is something that comes with time. But also if you understand and apply the concepts in the tutorial if kind of takes care of itself for the most part.

Regarding your other problem. Maybe you can post a sample. I do know that this can happen a lot if you edit in subD mode which I do not recommend. I usually set my mode to CCs and edit in polymode toggling only to see where I am. Then when I finally have all quads I will switch to standard subd mode.

*Pete*
07-18-2008, 04:06 AM
My latest frustration is accidently creating full polys that look like two points that arn't detectable when you hit "W" & bug the hell out of you when you hit tab....

any chance it is when you have more than 4 polys attached to a point, making the edges to form a star shape?
when you hit Tab you might get a scratched looking line where its all supposed to be smooth and fine...Spinquads (ctr +k) are your friends in those situations...

Talc
07-21-2008, 06:23 PM
Apologies for not responding sooner. It's been the weekend and today has been a real killer, however i think i found the source of my 2 point poly prob, it was USER (meaning thicko me) error.

I had resizised polys in one axis however leaving new polys along one edge that looked like 2 points but where not, thus creating problems with subD but not detectable when hitting "W", problem solved:thumbsup: