Payment

[Osteel]

Hello artists!

I had actually wrote a thread about this before, but once again I'm asking after a little research. I want to ask what you would want, if you were hired to do a job in creating isometric tiles. The quality of these tiles are similar to A2:Grimoire of the Rift, which you can see here.

With that in mind, which three options would you specifically demand?

1. Paid per hour, and how much?
2. Paid per tile, and how much?
3. Assuming the game was near complete, and confident it would be, would you rather share earnings?

So let me know what you would really ask for when in this hypothetical situation, based on the quality of the tiles and on your own personal experience. Also for you interest, this is a game that is a Flash game being hosted online, though not an MMO or multiplayer game.

Thanks guys!  

[sakket]

if it were a flash game or something similar getting ad revenue I'd probably go with a percentage of the profits, otherwise probably a fair price per tile and perhaps a verbal contract to keep me in the loop if it turns into a huge success.


EDIT; also depends heavily on if I can work well with the project director.

[Osteel]

Thanks for the reply. The hypothetical game here is a Flash game yes. Have you ever done a job where you were paid for tile, or in your opinion, what would you charge if you went with that approach over shared equity?

[sakket]

Thanks for the reply. The hypothetical game here is a Flash game yes. Have you ever done a job where you were paid for tile, or in your opinion, what would you charge if you went with that approach over shared equity?

goodness, well my experience is limited to a few projects (im an animator not so much pixel artist so my situation might be a little different) but granted a simple but well-marketed game can bring in over $1k, I'm usually content with perhaps 20 dollars per reasonable unit of contribution, but I'm notorious for underselling myself.

[Ryumaru]

1. $8-10 per hour
2. $ 10 per unique tile ( some tiles are edited from others and may be priced lower)
3. I would prefer hourly payment and a low percentage of the earnings ( as opposed to splitting 50/50)

[Osteel]

Thanks for that Ryumaru.

For an isometric game that required, lets say 150-200 unique tiles. How long would you estimate it would take assuming you worked on it as a job?

[Ryumaru]

I've never worked on a pixel project full time but if I were to work say, 8 hours a day, 5 days a week It would probably take 2 weeks to complete 150-200 tiles.

[Osteel]

Two weeks, that's amazing. Assuming we're not looking for industry standards and just nice pixel art as seen in this image here:



What would be estimated costs? I found this online, so credit to whoever made it! 

[ptoing]

I think 2 weeks would be slow if anything really, depending on how big the tiles are.

8 hours * 10 days = 80 hours
200 tiles / 80 = 2.5 tiles per hour

That would be 20 tiles per day. Which really is not that much if you are working fulltime.

[Osteel]

Wow, that's really short timing. Maybe 200 tiles is far less than what you could expect for a simple tactical RPG? In terms of tile size, I think we're looking for something like this:

[ptoing]

Thing with tiles is that you do not need 100% unique tiles as such.

What I mean by this is that you need you need tiles which transit from dirt to grass for example. and if you already have a grass and a dirt tile doing some grass/dirt mix tiles will be pretty fast usually.

And depending how pure pixelart you want to go you can do stuff very fast if you play if a bit loose.



These are 5 64x64 floor/ceiling tiles for Doom I made for a level I am working on when I got the time.
Anyway, the tiles on the far left and right were existing tiles from Doom and the 3 inbetween are simply blended using a function in Promotion and then cleaned up a tiny bit. This would work reasonably well in a grid based stategy game, if you want to have clear borders for each tile.

[Osteel]

Yeah I understand what you mean, meshing to form the variants needed for the ground. However within this ~200 tile list, there are also objects needed which will probably be very unique from one another. But with what you guys have said, I'm surprised at how short the time estimates are!

[DaCarterMan]

Dont mean to hijack your thread but how long do you estimate it should take to make 40 - 50 tiles, not all unique?  It probably differs on style in which it is a cartoon-like style.

[Geti]

1 week tops, Could be done easily in 2 days full time or 4 days part time, especially if  "cartoony" and therefore low colour count, where you can just focus on getting the form described and each cluster as beautiful as possible.

[big brother]

How can you assert your estimation so confidently without even knowing how big the tiles are? They can be done "easily" AND "as beautiful as possible?"

It seems like a lot of artists/clients forget about the revision process. Most producers double the time of an artist's estimate to account for this, since even good artists rarely nail exactly what the client wants on the first try.

[Lizzrd]

It depends a lot on style, size and most importantly the artist.
Not everyone has the same workpace.

[Helm]

Yes, it is prudent to not calculate final times based on how much time you spent pixelling tiles for your own projects. I'm not saying that's what Geti did above, but just in case. When you draw something for yourself you're faster because there's no misconception as to what you're trying to do. With a client, there might be many revision cycles.

[Atnas]

Very much so. My first few jobs, I wound up completely underestimating the revision process. Thankfully I had an understanding client who adjusted the price based on my new time rather than my uninformed estimation. It's more economic, I think, to give the artist one assignment to begin with, for both of you to get a proper idea of how long it will take.

Also, not all freelancers can work full time, so what is a 2 week job for some full time freelancers may double if they have another job or school to balance. I think Geti's original 1 week estimation for 50 tiles is more realistic than 2 days, however it really should be said that at least another half week is the safest timeframe imo. This is coming from a relative beginner, I am sure that ptoing could get it done in two and a half days. But work speed is VERY variable, and I doubt I would be able to do 2.5 original tiles per hour (more like 1.5), and full time is also not an option for me.

[Geti]

How can you assert your estimation so confidently without even knowing how big the tiles are? They can be done "easily" AND "as beautiful as possible?"

By assuming the worst case scenario (32x32, few recolours) and making a realistic estimation from there.

Cartoon styled tiles are excellent to work quickly on because you can block out a large chunk of the tileset in one go (say, all the tiles for a grassy area), quickly outline and two-tone shade any required forms (for our example, rocks sticking out of the ground, larger clumps of grass, some larger stone blocks and bushes, etc etc) so you can build a small mockup using all required tiles. From there it's easy to pick what forms don't work and where more variation is needed. From 1-2 hours work, you've then got a rough draft you can send to the client, and while you're waiting for feedback you can start on another part of the tileset.

The only thing that would make that take longer than a week is if he required some larger objects with various facing directions to match the tiles, or animation in some of them.

I can't work full time either (full time uni student), but I can draft ~8 original 32x32 tiles in an hour and polish them in 3, so I suppose I'm just under the 2.5/hr level. For smaller tiles (What I'm better at) I expect that count could be quite a bit higher than 2.5/hr though.

Getting feedback before you've done the hard yards (optimising those clusters, tweaking colours, etc etc) lets you cut down on wasted pixelling time a lot, so I think 1 week is reasonable for 50 tiles with some copypaste variation tiles, but because of possible revision issues I'd quote a longer time to a new client to give myself some breathing space.
If I was working with someone I had experience with (and therefore knew more accurately what they wanted) I'd give them a less lenient time frame.

It's all pretty context sensitive, but from what I've heard from DaCarterMan in a Personal Message just now I wouldn't have been too far off the mark with his requirements.

[CodeGeorge]

I'm really questioning what people are saying here.  I'd challenge anyone to make like 8 freakin' great quality unique tiles that work seemlessly with all past present and future tiles. 

I can easily spend hours on just a few tiles to make 'em really pop. 

Granted if it's really toony it would not take nearly as long but in order to put out a really good tree or other objects and have great diversity you've gotta spend a heck of a lot more time on them then these people are saying.  At least I do.  I'd wonder what quality of tile would go out at a rate of 8 per hour quiet honestly...


I'd usually charge probably $5-10 per unique tile mattering on style.  If it's super cartoony or simple less (like NES style or Zelda grass).

[Tourist]

I'm not a professional pixel artist.  I do however, have some experience in program management in an unrelated technical field.  It is very common for the expert to underestimate the time (and budget) required for any chunk of work.  I see the same thing in this thread.  It's not that the expert doesn't know the stuff (the pixel artist has art skills, the engineer understands the technical aspects of the project, etc). 

But when making an estimate, the expert assumes:

• Ideal working conditions, with no distractions or interruptions
• no time allocated to meetings and communication, including waiting for email responses and so forth
• ignoring or underestimating rework
• everything is planned out at the start; no time is spent figuring things while in the middle of the work

What happens in reality is that on longer projects:

• There are some interruptions that not only eat up the schedule, but disrupt the work flow.  You lose 'the zone' and it takes some time to get it back.
• There are some meetings that eat up the clock.  Or time spent waiting for a response that has to cross several time zones.  The example would be sending one set of images and having to wait for feedback.  Do you start the second set at the risk of additional rework, or do you stop for the day?  Both can cost time.
• The person starts the work and there are quite a few things that are undefined. "I know what I'm doing!  I'll figure it out when we get there! "  This can be ok, but you have to account for the time spent filling in the unknowns.  Also, doing this in the middle of work, with time pressures, can result in bad decisions.
  
  An example would be if the project said, "make some tiles with an ancient Egyptian look, maybe some hieroglyphs or statues."  When estimating what this would take, do you remember to include the time researching actual hieroglyphs or Egyptian symbols?  Most people forget that this sort of thing takes a little time, even if it's just a 20 minute read on Wikipedia, and it adds up.

Assuming the expert is an expert and has the skills (art in this case), they are usually off by a factor of three on the overall time estimate.  Sometimes it's as much as five.  A good plan should have equal chances of getting done ahead of time or of taking extra time (but almost nobody plans like that). 

That's not to say all the extra time is billable, but it does eat into the time available to complete the overall work,

Tourist

[Geti]

I would call 8 tiles "a few" tiles though, CodeGeorge. Polishing that many tiles of a similar style in 3 hours seems like a fairly reasonable estimate, but I could very easily have a different workflow or be talking about a different type of tile. The aim with any project is barely to get tiles that "work seemlessly with all past present and future tiles", as that seems fairly impossible. The aim is to produce tiles that work cohesively together (and with existing assets), are aesthetically pleasing in the context presented, and do not interfere with gameplay (by removing contrast or w/e)

Regardless, I suppose I could likely be jumping the gun a little, as Tourist has suggested (though, I'd hardly consider myself an expert). I might take you up on that challenge tonight though.

[Helm]

George, a whole NES-style tileset could be 5-6 hours of work from start to finish but as I've said this isn't an accurate description of how one works under the direction of a client. The issue isn't that there aren't people fast enough to make your eyes bleed. There are. I'm a very fast pixel artist, for example. The issue is revision fatigue, interruptions, stop-starts, backtracking and all the other things that occur as Tourist said, in a managed relationship.

[Tourist]

Ehh, to be clear, by expert I meant someone who has sufficient skills to accomplish the work.  If you also have to spend time learning skills (example: making a seamless tile), then the time required greatly increases. 

To get a better measure of your productivity, I suggest we make this a proper challenge.  The goal is not to compete for speed, but to give yourself a better idea when estimating for jobs.

Step 1: Note the current time on your computer. Write it down somewhere.  Write down the date, too (in case this takes a long time).

Step 2: Choose a tile size: 16x16, 32x32, or 48x48.  You will make 8 tiles at this size.  For optional added fun, make 7 tiles of this size and one tile at 64x64, because the game needs UI and/or title screen.  Or maybe the customer doesn't really know what he needs.

Step 3: Head to one of the random generator sites:
http://nine.frenchboys.net/index.php
http://www.seventhsanctum.com/
or any other random source that you like.

Select one place generator.  Generate no more than two places.

Select one character/monster generator.  Generate no more than 4 monsters or characters.  Select three of the four.

Select two categories of items, clothing, or treasure.  Generate no more than two items from each.  Select no more than one from each category.

If you chose to make a larger 64x64 tile, then go to http://chaoticshiny.com/tarotgen.php and generate no more than two tarot cards.  Select one.

This is to simulate someone else controlling what needs to be made.  You get to make some selections to avoid totally bad results, but you don't get to pick and choose until you get really good results.

4) Draw the tiles.
Note the time to complete a first draft of all the tiles, and the time when they are all finished.  You may also track the time spent actually working separately.

For the map locations,  draw either one tile for a map icon for each place, or 2-3 tiles as a sample for one of the places.
For the characters, make one tile each of them.
For the items, make a tile for each.
For the tarot card, if you chose that option, make a larger tile that incorporates the result.  It doesn't need to be a tarot card, but it needs to include most of the elements.

Optional fun: If you just want to restrict this to terrain type tiles to emphasize seamless tile building, then select one of the locations and build all 8 tiles based on that.  Even the larger tile if you chose that option.

5) Tally the times.  Compare time spent making pixels vs total calendar time, including eating, sleeping, and whatnot.  Time spent on the web sites above count against the total time spent. 

If you think, "this might be fun, but I'm not sure I have the time right now," then this challenge is for you.  Real life interferes.   If you set aside dedicated time to pixel and isolate yourself from distractions, you distort the results a bit, because you generally can't do that on larger projects.

If the generators give you results that are incomplete, you can 'ask for clarification' by supplementing the results with the results from a different table, blending the results.  Character + clothing for example, or equipment + room.  If you do this, generate no more than two results and select one to include.  Also, add at least an hour to the total time, up to half a working day (3-4 hours) for a more realistic response time.

If you really want to add to the fun, after you have made the first draft of the tiles, add a half a day to the total time, head back to the generator sites, generate another round of choices, and blend one additional separate result to each of the tiles you have just made.  This simulates the customer requesting a revision or changing his mind.  Don't replace what you have done entirely, but incorporate one or more elements for the revision.  If round one says 'man with red shirt' and the revision is 'woman with blue shirt' then change the sex or the color, but not both.  For places that only generate names or titles, add some detail based on a different generator instead.

Again, the goal is not to compete to see who can pixel the quickest, it's to give yourself a more realistic estimate of your productivity when responding to job requests.

Tourist

PS  Should I post this to the challenges board? 

[Gemini166]

Tourist, that post just made my day. I am SUPER indecisive when it comes to pretty much anything, so those generators are just what I needed to jump-start my imagination.
Thanks a TON. 

[Geti]

Now this is a challenge I'm quite keen on. Let's see how it goes.

[Geti]

Well, I did a small sheet as a practice for this.

The colours are off (too dark on my monitor, probably too saturated on others), but I've got to get faster at that as well so it's relevant. Probably getting closer to three with upload times and a cup of tea. Time for bed, this was a relaxing end to the week. I'm looking forward to some textual inspiration.

I can't decide whether to work on just environment or do the whole kit and caboodle with some characters and items. I haven't done map icons in a while, so I might go for the latter.

[big brother]

Those certainly are 8 tiles, but I'm not sure what I'm looking at. They look so noisy. I don't think the 2x2 blocks of pixels work in this case. Lots of banding, too, which definitely won't fly in tiles so small. There are also contrast issues with the palette, which further exacerbate the readability problem. I think they fall short of the "as beautiful as possible," easily-attainable goal you mentioned earlier.

By assuming the worst case scenario (32x32, few recolours) and making a realistic estimation from there.

I wasn't aware that 32x32 tiles are the biggest usable size. Even just looking at this thread alone, you will find that Ptoing posted some 64x64 sized tiles.

[Geti]

Those certainly are 8 tiles, but I'm not sure what I'm looking at. They look so noisy. I don't think the 2x2 blocks of pixels work in this case. Lots of banding, too, which definitely won't fly in tiles so small. There are also contrast issues with the palette, which further exacerbate the readability problem. I think they fall short of the "as beautiful as possible," easily-attainable goal you mentioned earlier.

I wasn't aware that 32x32 tiles are the biggest usable size. Even just looking at this thread alone, you will find that Ptoing posted some 64x64 sized tiles.

Cheers for the criticism. I agree with you, but do understand that I wasn't talking about making the tiles as beautiful as possible, I was talking about making the clusters as beautiful as possible, which with 3 hours to work and a cartoon focus (something like this or even this), which is very feasible.
With a more detailed focus, where you're using small clusters much more often and have more clusters in total to manage, the task becomes quite a bit more difficult. I've fallen short there as well, and I'm well aware of that. I'm going to have to step up my game, surely.

However, keep in mind that aesthetic beauty is in the eye of the beholder, for example "green" on TIGSource, while I would by no means consider him an authority, posted

I love these.. Those colors as well.

If he was my client, I'd be able to simply clean up a few of the more glaring issues from what I can see (some of the banding, though I think sometimes with vegetation at this scale the banded solution is the only way to prevent it being so harsh, as well as tweaking the greys a little more) and send them off as final. If you were my client I'd be in strife, obviously, as you don't like them, but from my experience a client generally has some images at hand for a style "target", or can at least mention a few games they like the look of.
Cluster beauty comes from your placement of pixels and is a little less subjective, which is what I was aiming for.
You seem to have latched onto that phrase from the beginning and it's been bugging me, so I wanted to clear it up.

In response to your comment on 32x32 not being worst-case, that's true, but due to most of my experience being with smaller assets I'd likely turn down a job offer if it were at 64x64. Surely it'd be more practical to paint the tiles once they hit that size?


Regardless, I'm quite keen to get some fresh inspiration. If I get time, I'll tackle this tonight, but I suppose I start counting from now. 9:20 AM 5th March 2011.

I decided to use the "Lost Civilisation generator" on seventhsanctum, as it gives you a more verbose description.

This just wide-ranging imperium was noted for its advanced tactics. It was destroyed by a neighboring country because of the people's extreme immorality, leaving behind only relics and methods of transportation.

Second click. I'm going to do 8 tiles, as before. 6 terrain tiles, 2 relics, and 1 "Set Piece" transport object which will be 64x64 at largest.
I decided to "ask for clarification" on the transport object and got

Rockdigger

so I guess the civilisation was subject to rocky terrain. The relics will probably be broken weapons and coins.

I'm not sure whether to do 32x32 or 16x16, what are your thoughts? I'm going to do some sketch design and research now, and then I'm going out for the day, so I'm anticipating this to take the weekend, possibly into monday. We'll see how it goes. I'm leaning towards the smaller size at the moment.

Off to have breakfast I do hope some other people do this, it's quite an interesting challenge.

[Geti]

I've still got some cleaning to do but it's coming along. Not sure where to go for the "relics", I'm going to have to ask for clarification on that one, or should I do a character instead? Choices to make, though if someone here "assigns" me a task there I suppose I'll go with that, as not having much leeway on the choices is part of the idea.
I asked someone online to point me to a random page of Arnes work as an "artistic sample" and they chose EXILE, so that's where the sci-fi influence has come from (I've interpreted "imperium" and "relics" in a more scifi sense than general fantasy sense because of it)

5 hours of pixelling/sketching in, a day and a half in on the calendar. This exercise is fun

8 16x16 tiles and one 128x64 set piece.

Issues:
-The rockdigger is a bit slapdash. I'm glad I played around sketching possible designs first, (though that took a good chunk of time) or it'd be even more so. Cleaning time, there, especially on some of the plantlife.
-The mockup layout doesn't make much sense, I'll have to shift it around a bit.

I'll likely build some background tiles so I can have a cave entrance in the mockup, to make the set more versatile. Probably in the realm of 12 tiles at most by that stage.

Criticism welcome.

[Tourist]

I'd say what you've got is fine so far.  Eight tiles plus a larger image, first draft complete.  I don't think the mockup needs to make much sense because you're not doing a full tile set, just enough to show that these few pieces can fit together.  Looking rather good, actually.

It's your choice if you'd rather make a full tile set, but if you'd rather follow through with the challenge then complete these tiles instead.  Polish them up until you think they are good enough to get paid for, and check the times. 

If you want data to revise and add to what you have, maybe use one of the object/treasure generators to add a relic or two.  The Serendipity site has a treasure generator (http://nine.frenchboys.net/treasure.php) that produces results like:

Lying here is a statue made of rosewood, depicting a series of interlocked circles.
You unearth a stone plaque depicting a battle. It is unnaturally cold to the touch.
You have found a candleholder made of battered steel.

Mostly fantasy stuff, so you' might need to adapt it for your realm.  A sci-fi light lamp instead of a candle holder, for instance.

Finishing up these tiles up is good enough to say challenge complete without the extra rework though.  The after-challenge activity is to figure if this was a reasonable simulation of your real work flow, figuring out what was important to getting the tiles done (like the sketching you mentioned), and maybe revisit the original request that started the thread as if it were a sample job.

Tourist

[big brother]

Not to go off-topic with crits --

The orange is too saturated, it clashes with the tiles for attention.

I'm not a fan of the line of rocks between the grass/rock tiles and the little rock tiles.

The stones look a little crate-like, they read more as individual objects. I feel like they should at least be able to tile with themselves. Using both the brown and green highlights on those tiles seems overkill. It would look more consistent with the established look if those green parts were grass (drawn like the grass tiles) growing in the rocks' cracks.

As for the rockdigger, it looks a good but a bit boxy. I would try to deviate from rectangular shapes to de-emphasize the implied tile grid. If that darkest shade was a little darker and with a slight blue tint, it would make the forms pop a lot more. Also, the vehicle's function is not immediately apparent. Initially I thought those grinders were a tank tread. The scoop that sticks out in front is difficult to distinguish from the grass and rocks.

Overall, this set is much more concrete and cohesive. Usually, sketching on paper is a good way to plan ahead without wasting much time.

[Geti]

I'll get this finished tomorrow I expect, I've got to get on with getting a game from start to finished by next tuesday pretty snappily O_o

In response, before bed:
@Tourist: I think I will just polish these ones up. I don't have the time to go gallivanting off developing this for much longer, so I'll stick to the task and evaluate.

@big brother: My monitor has adjusted its colours (I'm using flux) and it shows the orange as much harsher. The green sticks out a bit more than I'd like on the brickwork too, I'm going to tweak the various forms there; the larger stone doesn't read too well, and the grass topped dirt needs tweaking.

The rockdigger crits are very valid, I'm concerned about the readability but the spec was that everything was derelict and as the machine had to double as a "method of transportation" I thought it could be a good idea to have a bit of ambiguity.

Anyway, I'll push and pull everything around tomorrow amidst a short day at uni (yes!) and a bunch of paperwork from blackberry (boo!). Thanks for the feedback, you've been a help through this regardless of  some earlier harshness.

[Geti]

it's already updated itself upwards on the page, but here's the final product:

I toned down the saturation of the background colour and added some small relic tiles to add some noise to the surface.

6 hours work total, 2.5 calendar days total.

I'd say this is heading towards representative. It's how I'd "behave" if I was a week or two from the deadline, as I've been busy with a lot of other things. I'm glad I've got it done.

Now I've just got to do all the assets and code for a game to go onto a platform I've never developed for. No fear

[Helm]

I added another dirt halftile, tell me if you think it makes a difference in breaking up the grid a liiiittle just for the dirt sections. I don't suggest doing this with every gradiation, but it pays to do a little of it.

I also wanted to say I love the vibe of your tileset and that I think you've gotten a bit of a raw deal as far as reaction go so far because people thought you were boasting. Your art is good, you seem fast enough, please stick around in Pixelation and help others/help yourself

[Geti]

I certainly don't plan on leaving any time soon, Helm.
The half tile does help the transition (particularly under the set piece) but I'm not sold on exactly how useful it would be. I'll have to play around with that idea a little more, it's interesting to think of antialiasing at a much larger scale.