Project Future - how it is assembled

[Tapewolf]

How Project Future is coloured

EDIT:  I'm not sure offhand what the rating of the Library is.  One of my brushes has a rather interesting name which is probably Not Safe For Church, (or Work, depending on where that is).


There have been some questions about how Project Future is coloured, and whether it should take as long as it does to complete each page - currently 6 hours plus, depending on complexity.  There is also the separate question of why the comic on the site has an aliasing problem (staircase effect).  That issue is dealt with at the end of this post.

As suggested, I am posting a description of the processes which I use to assemble the comic.  Civil and constructive suggestions for streamlining the process and otherwise improving it are welcome, and who knows, some people may find some of these techniques useful.  I can't recommend all of them, though.

Before I start there is one thing I will say - before anyone suggests it, moving to Photoshop is unfortunately not an option.  Among many other reasons, it is far, far too expensive, and I need the capability to run it on three systems - including platforms that Photoshop is not available for.  Other alternatives to GIMP I will consider with an open mind, but they must be able to run on UNIX systems as well as Windows, and GIMP seems to be the only one that fits the bill :-/

From a personal observation - and with no offence to Ren - most of the time is going into cleaning up the line-art as supplied by him.  Comics drawn by Spooks and Kipiru generally take about half the time since there is substantially less cleanup work to do.
It is worth bearing in mind that when the comic began, I took on the colouring side because I wanted to share the workload, so I'm not complaining about it, it's just one of those things.

Anyway, this is how the comic is constructed.

The process as it stands

* In GIMP, import line-art as 'Background'.

* Duplicate to a new layer 'Background copy', keeping the original for reference.

* Add proper panel lines, make sure each panel is separated from each other panel so that a flood fill in the interstice will not leak into any of them.

* Duplicate this to a new layer called 'ink' and deactivate the previous layers.

* Repair outer edges of line-art so that the white background can be removed via the Fuzzy Select tool, without it also gutting the characters as well.  This takes a long time, depending on the quality of the inking and subsequent scan.

* Check that the panthers etc look like they did in their last appearance and modify accordingly if there is a significant deviation (usually to do with the leg joints).

* Erase any background objects, e.g. fragments of chair, guidelines and any extraneous things such as characters from videogames that should not be there.

* Duplicate the layer as 'Colour' and deactivate 'Ink'.  If there were background objects such as chairs or desks, create another layer called 'Backdrop' and draw those using 'Ink' or the original layer as a reference.

* Repair internal regions as necessary and flat-colour the line-art using flood-fill.  Create any character effect layers such as warp-aci underglow, clan marks, decals on the panthers or glasses.

* For the panthers, move their artwork to a new layer, invert it, block-replace the colours (outline to grey, interior to black, make sure eyes are red, pupils are black and that the grey joint sections, ears, nose etc are filled in.

* Remove any white specks by doing a block select of white for each panel (Select By Colour tool) and deselect any genuinely white areas (e.g. eyes, Kris' coat) using the Fuzzy Select tool. Then block replace the remaining white areas in black (shift-floodfill).

* Duplicate layer to 'Colour blended' and deactivate 'Colour'.
For fur transitions, e.g. Dorcan's wrists, Josh's mask and any other fur patterning, mask off the relevant areas using Fuzzy Select and blend them together using a tablet and the Smudge tool.  The right size and density of brush should create a fuzzy, furry effect.  Add the dark muzzle spot on any fox faces at that point.

* For leather clothing, go back to the 'Colour' layer and globally select the base colour of the coat using the Select By Colour tool.  Copy it and paste it into a new layer called 'Leather black', 'Leather brown' or so forth.  Use a different layer for each distinct colour.

* With a tablet, select a fuzzy brush, the dodge tool and set the exposure to 15%, Range = Midtones.  Work out the direction of the light for each scene and highlight the edges accordingly.  Try to follow the contour of the object.

When this is done, duplicate the layer and make any necessary colour correction using the curve menu (different layers for different colours makes this easier, tan leather tends not to need correction).  Remember that the sheen will look different when the image is downsampled to 72dpi.

* On this layer, select empty space, invert the selection to mask just the leather off, and use the Smudge tool to blend it.  Smear the highlights extensively to reduce the plasticky look, unless like Niall's coat, or the panther armour, it is supposed to have a very polished look.

* Repeat for other highlightable materials as necessary.  For the panthers, set the range to Shadows and the exposure to 25%.

* Create a new layer at the bottom of the visible stack for the walls.  Block fill it in the wall base colour and angle it off to produce a sort of two-tone shadow transition effect.  When this is done, duplicate the layer and apply a 20 pixel radius Gaussian blur to the copy.

* Create a new layer for the floor and ceiling if necessary.  Add the skirting boards.  If using chequered tiles, create a checkerboard pattern and try flailingly to get it to match the room layout using the perspective transform tool.

* Add any other background items such as light switches to the backdrop.  Video screens or backdrops behind the backdrop layer go in a new layer called 'Underdrop'.

* If the corners of the room are visible, e.g. a down-the-corridor shot, flounder about with the room perspective for 15-45 minutes, depending on scene complexity and state of mind.

* If there are seated or standing characters where the floor is visible, create a new layer called 'Shadows' set to about 29% opacity.  Use a very large brush, e.g. "Fucking massive" and draw faint shadowed regions as necessary.  You may need to mask off parts of the colour layer, depending on where the shadows fall and the structure of the image.  This is done as a single layer since two shadow layers will interfere constructively where they overlap.

* If there are any lighting effects, create a new layer on top using the subtraction layer filter (apparently Photoshop doesn't support this).  Set it to be white and ease off the layer strength until you have about the right base darkness level.  Using a very large brush (again, "Fucking Massive" is handy for this) cut away into the white to let the light through.  You can use multiple layers of this effect to increase the dynamic range.

* If there are any magic effects, glows, ion beams or suchlike, create multiple layers of the appropriate colour, using the addition layer filter, which will push things towards white and saturate them.  Adjust the layer filters and strengths as necessary.

* When the image is complete, export it to a flattened PNG file.

* Open up the PNG, cut it into separate panels.  Trim off the black borders.

* On a machine running MacOS 10.4 or higher, make sure you have installed the following software:
1. ComicLife 1.2 or higher
2. pngtopnm and pnmtopng (ImageMagick may do at a pinch)
3. optipng 0.6.3 or higher

* Import the panels into ComicLife, add captions and crop as necessary.  When finished, Export to images using a 72dpi PNG file.

* Run the following commands (I use a shell script to automate this):
$ pngtopnm [path and filename of exported image] --background=White -mix > uh.pnm
$ pnmtopng uh.pnm > strip132.png
$ rm uh.pnm
$ optipng -o7 strip132.png


...this will remove the alpha layer from the comic, replace it with white and then optimize the PNG compression, which can reduce the file size by up to about 30%.

* Now upload the file (I use lftp).


The aliasing problem

Comiclife for Mac is used to add the captions and text.  It is incredibly useful, and frankly there isn't really any substitute for it on the market, open source or closed.  The only viable alternative is to manually to it all in GIMP or by importing the frames into Inkscape and drawing the speech balloons as vectors.  I tried this - it's going to take a lot of practice and coaching before I get even mildly competent at it.

When rendering, ComicLife also performs slight colour correction, and when scaling to 72dpi also runs a sharpness filter on the resulting image.
This filter darkens and sharpens the edges of characters etc making them look bolder, but also has the effect of emphasising staircase edges, e.g. on the diagonal panels on this page:

http://www.project-future.org/stuff/things/strip132.png

If I export the comic to 300dpi and then use ImageMagick to perform a lagrange filtered reduction to 72dpi, it improves matters substantially, but leaves the characters looking blurred and washed-out compared to CL's own rescaling algorithm.

e.g.

http://www.project-future.org/stuff/things/strip132cl.png (ComicLife)
vs
http://www.project-future.org/stuff/things/strip132l.png (LaGrange)

...I will probably post a poll in PF forum asking which version is preferred.  It would take a day or two to re-export the archives in 300dpi and process them, but it is entirely feasible.  It would give me a chance to fix the off-colour walls in chapter 2 as well...

[Turnsky]

uh... tape, don't get me wrong here, but your statement on speech bubbles and dialog, almost every image editor out there has a method of creating speech bubbles and dialog..

http://img.photobucket.com/albums/v625/Turnsky/Sketches/speechbubbleshowdotheywork.jpg including photoshop.

[Alteisentier]

uh... tape, don't get me wrong here, but your statement on speech bubbles and dialog, almost every image editor out there has a method of creating speech bubbles and dialog..

http://img.photobucket.com/albums/v625/Turnsky/Sketches/speechbubbleshowdotheywork.jpg including photoshop.

Pretty much the same in Gimp.

New layer, click transparency.
Eclipse or square tool. Tick rounded edges.
Draw your bubble.
Fill it with white.
Take lasso tool, draw the little tail flick. (Make sure to add to selection)
Select, path.
Select, non.
Edit, stroke bar.
Make sure to select black so your line is black.
Click okay!

Takes like, 50 seconds.

Why don't you just create a transparent layer and draw over the line work? You have a tablet. That would remove so much hassle for you.

[RobbieThe1st]

This is all very interesting.
Personally, I'm curious about vectorizing the lineart, sizing it up many times its final size, rasterizing it and then shrinking it. This will produce nice anti-aliased lines, but the problem is in the first step.
Currently I'm messing with VectorMagic(which runs on Linux via Wine, or online version), but it isn't super great when working with fully colored final images.
Tape, would you mind posting a copy of one of your 1-bit linart scans? I'd like to mess with it, and I'm sure others might as well.

Thanks,
-Rob

[Tapewolf]

uh... tape, don't get me wrong here, but your statement on speech bubbles and dialog, almost every image editor out there has a method of creating speech bubbles and dialog..

Yeah, you can build them by hand.  I wasn't aware of the technique Alt suggested and I will look into it, see if I can make it work for me.

But, unless I'm missing something, this isn't the same as Comiclife, which handles panel clipping, font effects and most importantly, has a dedicated speech bubble editor.  Alt says this took almost a minute for a single bubble, and that didn't include the titling.
In Comiclife, for all its faults (and yes, it is something of a crutch) it's more like 5 seconds, which includes typing out the dialogue.

Since the speech bubbles are vectorized they are easy to stretch around and hell, I could translate the entire archive to Russian or something in under a day just be retyping it and repositioning them as necessary.
Basically, it automates most of it - which does result in identikit speech bubbles, and sometimes routing the dialogue is a pain.  But for sheer convenience, you have to see it in action to appreciate it.  Especially if you want to go back and modify the dialogue after the bubble has been created.

This is all very interesting.
Personally, I'm curious about vectorizing the lineart, sizing it up many times its final size, rasterizing it and then shrinking it. This will produce nice anti-aliased lines, but the problem is in the first step.

I have grave doubts about vectorizing things like plasma, Gaussian blur effects or the additive lighting.  Frankly, the vector lines added by ComicLife look pretty decent when the thing is rendered to 300dpi, and IIRC it can output higher than that too.  It's the 72dpi output that's being a problem.

Tape, would you mind posting a copy of one of your 1-bit linart scans? I'd like to mess with it, and I'm sure others might as well.

This is page 130:
http://www.project-future.org/stuff/things/Project%20Future%20Page%20130-original.png

I could post page 132 as well, but I don't have it on me.

EDIT:
For people trying their hand at colouring it, I would be interested to know how long it takes.

[VAE]

Tape... i was thinking of something.... a good possible way would be making prefab bubbles, placing them on a layer of their own, under the text and stretching them to fit the text (first placing text of course)... then perhaps adding the "tail" to them by hand.
Shouldn't take too long... maybe 15-20 sec with text already there, and if you save a duplicate of each comic in something that can  manage layers, changing text is equally trivial as the previous.

[Alteisentier]

With all the amount of time and effort you put into fixing up the comic you may as well just learn to trace the lines with your tablet. It'd produce much better looking lines for a start, make your job easier and quicker. And allow you to add more details into the comic.

I'm not sure why you are using 1bit scans, when you could get a gray scale. change the saturation, then make an additional layer and multiply to make the lines nice and dark, then just clean them up. Since the 1bit scans you are getting look simply awful and blocky as all hell.

I don't even know where your argument with the comic program holds up, you're honestly putting 6-10 hours into this. That's fine, you're dedicated. That's great, so dedicated you should be looking for ways to improve your comic; and that program is just bringing your over all quality down because of 'automation' an extra few seconds fixing up and bubbles in gimp, will result in a much higher quality over all; which is what you should strive for. And once you figure it out, it'll be done just as quick. Gimp can even automate it it's self; I don't remember how to make it do that. But i'm sure you can use google to find a tutorial; if you really want to improve.

Honestly, try using your tablet on a transparent layer to draw over a gray scale comic, use a transparent layer underneath for your base colour. And one above that for all your highlights and shade. Once you get into the hang of it, it'll make things so much easier, faster and better looking.

It's things like "On this layer, select empty space, invert the selection to mask just the leather off, and use the Smudge tool to blend it.  Smear the highlights extensively to reduce the plasticky look, unless like Niall's coat, or the panther armour, it is supposed to have a very polished look."

So overly complicated for the work you are doing. You could skip this entirely and just use a darker gray to colour the jackets highlight, at the resolution and quality of the comic, we can't tell the difference. That's one step you can skip entirely and save time on.

"* Duplicate layer to 'Colour blended' and deactivate 'Colour'.
For fur transitions, e.g. Dorcan's wrists, Josh's mask and any other fur patterning, mask off the relevant areas using Fuzzy Select and blend them together using a tablet and the Smudge tool.  The right size and density of brush should create a fuzzy, furry effect.  Add the dark muzzle spot on any fox faces at that point."

This as well, it's particularly over kill, no one can tell the difference.

[Turnsky]

*no way in heck i'm gonna requote all THAT*

yeah, there's a method surrounding it all.
Alternatively have the panel gutters in a layer above the dialog balloons, and trim off the excess when needed.

Automated processes are good and fine, but sometimes one just wants to streamline the process in favor of something a tad simpler.

i'm fairly unaware of the nuances of comiclife, but if it can export to PSD or something, you could save it in your primary image editor to dispense with the output issues comiclife gives you. Might be worth a try.

as for the lineart... oooh, messy, it might be worth sourcing the proper greyscale scans from ren instead of 1-bit, much easier to tweak in an image editor, then.

[Tapewolf]

i'm fairly unaware of the nuances of comiclife, but if it can export to PSD or something, you could save it in your primary image editor to dispense with the output issues comiclife gives you. Might be worth a try.

What I have been wondering about earlier is extracting the bubbles created by Comiclife and superimposing them on the    lineart as a new layer.  That would be fiddly, though and adds an extra step that can't be automated (whereas exporting at 300dpi and rescaling from a script, can).

as for the lineart... oooh, messy, it might be worth sourcing the proper greyscale scans from ren instead of 1-bit, much easier to tweak in an image editor, then.

Yes, I've been meaning to ask about that.  If nothing else - if I stick with my current technique - it would still give me a lot more control over the B/W conversion.

For things like Dorcan - Road Warrior, Pirate Niall, Adventurer Jakob and various other things, I get just the pencils in greyscale.  Depending on quality of the pencilwork, it may take some time to get those in a suitable state for colouring.  The comic I always get in 1-bit format.

Part of the issue is that I don't always get the raw scan anyway - in some cases, e.g. Chance in P130, he does a lot of editing his end first before I get to see it.  I'm not sure if he's doing that in 1-bit format or not, and what effects a greyscale version would have on his cleaning up stage.

It's something I'll have to ask him about, but yes, my gut feeling is that the 1-bit scans are definitely not helping matters.

With all the amount of time and effort you put into fixing up the comic you may as well just learn to trace the lines with your tablet. It'd produce much better looking lines for a start, make your job easier and quicker. And allow you to add more details into the comic.

It's funny, actually.  On the very first page (late Feb 2008) I did try to do this:
http://i89.photobucket.com/albums/k216/tapewolf/projectfuture/josh_comparison.png
...but I found it looked too much like I'd drawn it.  Kind of like some people hate the sound of their own voice when it's recorded.  In a way, I feel that Ren is getting marginalised enough as it is - redrawing it in some strange mixture of Ren's and my style would make me feel even more guilty about it.

I didn't go into detail with regards to the cleanup and I'm not sure how easy it would be to do.  Probably it would be easiest just to set up an X11 recorder and make a horribly boring film of it that's too long for youtube.

What I am basically doing though is checking the line edges and redrawing them myself where necessary.  I'm not sure that retracing the entire thing from scratch would speed it up, to be honest, but maybe I'm missing something.

I'm not sure why you are using 1bit scans, when you could get a gray scale. change the saturation, then make an additional layer and multiply to make the lines nice and dark, then just clean them up. Since the 1bit scans you are getting look simply awful and blocky as all hell.

I did ask Ren about sending greyscales instead of B/W scans at one point, I will ask him about it again.

...you should be looking for ways to improve your comic; and that program is just bringing your over all quality down because of 'automation' an extra few seconds fixing up and bubbles in gimp, will result in a much higher quality over all; which is what you should strive for. And once you figure it out, it'll be done just as quick. Gimp can even automate it it's self; I don't remember how to make it do that. But i'm sure you can use google to find a tutorial; if you really want to improve.

One of my worst fears is inconsistency.  While there is a lot to be said for switching over to doing everything manually, even just in terms of learning how to do everything, it's going to result in a sudden change in appearance of the comic rather than a gradual one, and that is something I'm not keen on, especially not in the middle of a chapter.
However I will start practicing, and look at doing a manually-titled version of the current comic tonight.  Heck, if it looks that much better I might retrofit it.

Honestly, try using your tablet on a transparent layer to draw over a gray scale comic, use a transparent layer underneath for your base colour. And one above that for all your highlights and shade. Once you get into the hang of it, it'll make things so much easier, faster and better looking.

I think that needs a little more explanation (aside from the fact that I don't have the greyscale version).  Am I right in thinking that this is what you're suggesting?

1. Hollow out the grey line-art.
2. Manually brush in the base colour on a layer underneath the lines, painting it using a tablet or something.
3. Add a new layer for highlights and shade

If so, there are two pitfalls I can see.  For 1, it's going to mean a shitload of cleaning up the pieces of grey that shouldn't be in there.  Ren uses this technique nowadays and that seems to be the bottleneck for him.

The second one is the highlighting.  The only way I've been able to get it to look right is by replicating the base colour and dodging it.  This gives you an additive effect - through multiple passes - and a lot of control over the depth of the effect.
I have tried to do it using filters and just painting in the parts I want to highlight - the slight gloss on Abel's dark fur in this for example:
http://project-future.org/stuff/things/Amber2010.png

...but it's not really giving me the effect I want.  Of course the problem might instead be with what I consider to 'look right'...

So overly complicated for the work you are doing. You could skip this entirely and just use a darker gray to colour the jackets highlight, at the resolution and quality of the comic, we can't tell the difference. That's one step you can skip entirely and save time on.

Ye-e-es, but one of the things I've been very careful to do is keep the full-res versions around.  People who have been to AC have probably seen the 300dpi version in my folder, or been given a business card with the URL to download them from.  If PF was only ever supposed to be viewed a 72dpi, it wouldn't be such a problem.  I've been trying to make it look decent-ish in 300dpi and 72dpi, and with the exception of the odd secret message that is only visible in the full-res prints, I don't really want to compromise it by using a technique that relies on the DPI being low.

Heck, if they didn't take so much bandwidth and if I wasn't slightly paranoid about someone trying to sell prints of it, I'd have a little download link on the main page...

By the way, on the topic of inking via tablet:
http://project-future.org/strips/comic/strip094a.png
...This was supposed to look shit, but somehow the lines ended up kind of crinkly, which I did not intend.

http://project-future.org/strips/comic/strip094.png
...This one was done because the earlier version looked a bit too good.

[Alteisentier]

One of my bugbears is consistency.  While there is a lot to be said for switching over to doing everything manually, even just in terms of learning how to do everything, it's going to result in a sudden change in appearance of the comic rather than a gradual one, and that is something I'm not keen on, especially not in the middle of a chapter.
However I will start practicing, and look at doing a manually-titled version of the current comic tonight.  Heck, if it looks that much better I might retrofit it.

I'll take this point first, because this is an important one from an art perspective.

Stop thinking like this, it doesn't work like this. At all. If you're going to be outlining and coloring your self, you can't think this. Or you will quite literally, be stuck at the same quality for ever. There is no perfect scale of improvement, no matter what you do it will always look out of place, more so at such a low level. Because you get exponential gains for such little improvement. Unless you want another artist to take this comic entirely off your hands and do all this work them selves; you're going to have to learn to live, or atleast deal with your comic changing over time. Until you get to a point where the gain for improvements starts to drop, and thing really become less noticeable. This is typically reserved for a much higher quality of art.

I think that needs a little more explanation (aside from the fact that I don't have the greyscale version).  Am I right in thinking that this is what you're suggesting?

1. Hollow out the grey line-art.
2. Manually brush in the base colour on a layer underneath the lines, painting it using a tablet or something.
3. Add a new layer for highlights and shade

If so, there are two pitfalls I can see.  For 1, it's going to mean a shitload of cleaning up the pieces of grey that shouldn't be in there.  Ren uses this technique nowadays and that seems to be the bottleneck for him.

The second one is the highlighting.  The only way I've been able to get it to look right is by replicating the base colour and dodging it.  This gives you an additive effect - through multiple passes - and a lot of control over the depth of the effect.
I have tried to do it using filters and just painting in the parts I want to highlight - the slight gloss on Abel's dark fur in this for example:
http://project-future.org/stuff/things/Amber2010.png

...but it's not really giving me the effect I want.  Of course the problem might instead be with what I consider to 'look right'...

I'm not sure you understood at all. If you create a transparent layer above the scan. You should be able to draw on it with.. Gasp, not having to change anything underneath at all ever. This is how the greater majority of artists work out there. They typically do not use the scan for colouring at all, but insted use it as a base image, improve over it on a new layer. And then dump the scan entirely.

This will leave you with a layer that is totally transparent except for your line work. Where you can draw under it. So the lines don't get messed or changed at all. This is the typical affair.

Top Layer: Highlights.
Middle Layer: Line art.
Bottom Layer: Base colour (One, possibly two at most)

As for the effects, you can use whatever you want I guess. But you're putting in a lot of effort a little bit of return when most of it is a few lines on the base colour.

And yes, I know it's ment to be seen at "300DPI" which is a silly term to use, because that's a printing specification. And not a resolution. But you arn't doing that. And to be honest it looks so glaringly bad at a higher resolution i'd almost say you should try putting gradients on everything, the down sizing in this case is actually making it look better.

By the way, on the topic of inking via tablet:
http://project-future.org/strips/comic/strip094a.png
...This was supposed to look shit, but somehow the lines ended up kind of crinkly, which I did not intend.

http://project-future.org/strips/comic/strip094.png
...This one was done because the earlier version looked a bit too good.

Yup, you suck at that. No doubt about it. But you're going to have to learn. I'm afraid using GIMP Only doesn't make this an easy process. Because it's not really designed for drawing in per say. But with a tablet, some proper sensitivity settings, you should be able to over come this. Even if you simply just trace the artwork at a higher resolution.

I hate to say it, but with out taking a personal and more hands on approach. There's not really any other way to approve, you may want to look into letting Ren do everything.

[Tapewolf]

I'll take this point first, because this is an important one from an art perspective.
Stop thinking like this, it doesn't work like this. At all. If you're going to be outlining and coloring your self, you can't think this. Or you will quite literally, be stuck at the same quality for ever.

Actually I was saying that I didn't want to do a sudden, massive change in the middle of a chapter.  A sudden change at the start of a new chapter, that I am open to.  I later clarified it when I realised that the current chapter has only just started anyway.

If you create a transparent layer above the scan. You should be able to draw on it with.. Gasp, not having to change anything underneath at all ever. This is how the greater majority of artists work out there. They typically do not use the scan for colouring at all, but insted use it as a base image, improve over it on a new layer. And then dump the scan entirely.

Right, now I see where you're coming from.
For what it's worth, I collect PSD files from different artists, hoping to learn new techniques from them.  In doing so I've seen lots of different approaches for the base colouring, including the one I've described, and the one you are describing here.
Your earlier one-line description matched several techniques I'm aware of, hence my confusion.

(And by '300dpi' I mean that the file is of a resolution intended for printing at 300dpi on an A4 page.  It was easier to say that than look up the resolution.)

Yup, you suck at that. No doubt about it. But you're going to have to learn. I'm afraid using GIMP Only doesn't make this an easy process. Because it's not really designed for drawing in per say. But with a tablet, some proper sensitivity settings, you should be able to over come this. Even if you simply just trace the artwork at a higher resolution.

I will try this, but there is a very real chance it will backslide from where we are now, and that is what I'm trying to avoid, rather than a gradual progression.

[Alteisentier]

It's going to back slide, you obviously can't draw at all. But you're going to have to start at the bottom and develop these skills; just looking at PDS's is about as good as looking out your window and going "Yup, grass is green"

Sadly grass is not entirely green when you look at it as an artist. It's Green, blue, yellow. Different shades of all of those, mixed with white's and grays. What you can do however is whenever you have free time, is to practice just drawing over your comics, it's going to look shit at first. But everyone started out shitty.

[Mao]

I'm going to say only one thing:  Nothing ventured, nothing gained.

Realistically, the worst that happens is that in your free time you try the suggestions given here and they don't work.  You don't necessarily have to apply them to the current comics and can wait until a time to show them that you're comfortable with.  You've been given a lot of good advice by talented peers and, while I know some of it can seem like a blow to your pride, don't think of it that way.  Even if, by some off chance it is meant as one, there are lessons to be learned here and things you can take away from this.

I would suggest though, not just trying them once and giving up if they don't quite work like magic.  Instead, try the suggestions and post up the results for all to see.  Iteratively improve them, and then when you've reached a comfortable point, start putting the fruits of your labor into the comic production process.  This will ensure the comfortable transition that you desire, and ensure that you continue to improve.

[RobbieThe1st]

I have grave doubts about vectorizing things like plasma, Gaussian blur effects or the additive lighting.  Frankly, the vector lines added by ComicLife look pretty decent when the thing is rendered to 300dpi, and IIRC it can output higher than that too.  It's the 72dpi output that's being a problem.

I can't say I disagree with you; I meant simply using vectors for "improving" the lineart quality, then working on it as a raster image like normal


Now, just throwing this off the wall here, but I've found the Paths tool in Gimp to be very useful when you can't draw a steady line(like me). Its kind of time-consuming, but it might be useful in some cases as it creates "perfect" lines(At least for speech boxes, panel-frames etc).
You may already be doing this, but I figured I'd point it out anyway.

Edit:
Also, I'm curious: What would your perfect speech-bubble tool have in terms of features?
Would you like something that automatically sizes the bubble to the text as you type? Or a set of static bubbles? Or...?

[Tapewolf]

I would suggest though, not just trying them once and giving up if they don't quite work like magic.  Instead, try the suggestions and post up the results for all to see.  Iteratively improve them, and then when you've reached a comfortable point, start putting the fruits of your labor into the comic production process.  This will ensure the comfortable transition that you desire, and ensure that you continue to improve.

That's pretty much what I'm going to do.  The colouring I do on the side, e.g. "Dorcan - Road Warrior" is the sort of place where I experiment with new techniques, which I will do, and fold these into the comic process as appropriate.

I can't say I disagree with you; I meant simply using vectors for "improving" the lineart quality, then working on it as a raster image like normal

Ah, right.

Now, just throwing this off the wall here, but I've found the Paths tool in Gimp to be very useful when you can't draw a steady line(like me). Its kind of time-consuming, but it might be useful in some cases as it creates "perfect" lines(At least for speech boxes, panel-frames etc).
You may already be doing this, but I figured I'd point it out anyway.

Paths are something I've never really used up to this point, but my interest is definitely piqued.

Also, I'm curious: What would your perfect speech-bubble tool have in terms of features?
Would you like something that automatically sizes the bubble to the text as you type? Or a set of static bubbles? Or...?

Okay, here's the wishlist.  Ideally it should run on something other than windows, because Windows can't see the network.  At the moment the Mac is doing all the edit-assembly work, so support for that would be preferred.  Something that runs under X11 would be fine.

Speech bubbles
It should flow the text around the speech bubble and expand it vertically as necessary. It should reflow the text as the bubble's size is altered.  It should be possible to expand a bubble into two linked ones (p132), or optionally merge bubbles together by overlapping them (though I rarely use this). You should be able to drag the tail end around, and preferably also be able to move the base of the tail as well, which is something Comiclife doesn't support.  Different tail styles are a must.
You should be able to set the fill, border and text colour and alpha level independently.
Speech bubbles should be able to escape from or be clipped by the panel boundary, controlled by a checkbox per-bubble.
It should be possible to import new bubbles as SVG files, which Comiclife does not support.
Square text boxes (e.g. "Elsewhere..." or the video display on page 129) should be able to magnetically click to the panel boundaries.  Implementing free text boxes as square speech bubbles with the tail turned off is of course one possibility.  It should be possible to rotate textboxes arbitrarily.
The ability to have a speech bubble containing an image is greatly desirable - this currently requires a bit of frigging around with ComicLife, or else doing the bubble manually in GIMP (as was done on page 24 using the path tool.  I wasn't too happy with that one).

Compositing
If the software also handles the compositing side, e.g. panel layout, it should be possible to drag the panel art into the panel and pan or rescale it as necessary.  Multiple page support is a must.  Panel art needs to be included in the data file, as opposed to a file link.
It needs to be able to export as PNG, or a similar lossless 24-bit format at a selectable resolution (or at a pinch, 1:1).  Ability to export PDF in various resolutions is highly desirable.

[Turnsky]

y'know, you should be able to make a speech bubble in any vector program using just the simplest of tools..

http://www.ariannia.com/images/stuff/speechbubble.svg

you can thank me later

edit: i'll elaborate inna bit, however.

edit2, the elaborating:
in illustrator
Basically all i did was make a simple circle, and with the pen tool i added a few extra anchors, the middle one being the 'tail' so to speak. tailored to shape, and with illustrator being illustrator, i saved it as a shape for use later on.

Photoshop has the same functionality in this regard, also. including saving it for later use.
Now i'm fairly certain that the other editors can use vectors like this, so there you go.

[Alteisentier]

y'know, you should be able to make a speech bubble in any vector program using just the simplest of tools..

http://www.ariannia.com/images/stuff/speechbubble.svg

you can thank me later

edit: i'll elaborate inna bit, however.

edit2, the elaborating:
in illustrator
Basically all i did was make a simple circle, and with the pen tool i added a few extra anchors, the middle one being the 'tail' so to speak. tailored to shape, and with illustrator being illustrator, i saved it as a shape for use later on.

Photoshop has the same functionality in this regard, also. including saving it for later use.
Now i'm fairly certain that the other editors can use vectors like this, so there you go.

There's also the fact Gimp does vectors and paths its self. With rather easily modified pathing size, so you can resize the things on the fly. Has SVG support obviously and you can probably mangle it in some way to resize text.

Really, the stuff you want is so horribly select and hard to appease that you may as well just learn to use your tool of choice; which is Gimp. Which can do all of these things you ask; it just requires you doing a bit of learning.

[llearch n'n'daCorna]

I suspect you already know this, Tapewolf, but in case anyone else was wondering...

Under linux, Debian specific, you can get optipng in the, well, optipng package. You can also get pnmtopng in netpbm; interestingly, I can't immediately find pngtopnm...

[Tapewolf]

Under linux, Debian specific, you can get optipng in the, well, optipng package. You can also get pnmtopng in netpbm; interestingly, I can't immediately find pngtopnm...

The mac was the awkward one, actually.  That one had to be compiled from source.  There is a win32 binary on optipng.sourceforge.net and as you say, it's available in the Debian and Ubuntu software repositories.

[Turnsky]

you know, thinking upon this for one reason or another, i've come up with an idea.

an artistic challenge if you will. You want to improve, yes? but you don't want to make improvements to the comic so as to not appear jarring to your audience, yes?
(which i personally find silly as a webcomic should have artistic progression, especially early on)

However, the idea is to take a piece that you do normally, be it a past comic page or whatever, and just experiment with it, or really pull out the stops when it comes to a cover page.

Doing a webcomic isn't just about setting a style, it's also constantly challenging yourself and driving yourself forward, do not fall into a "style" rut. Honestly it's sometimes good to take a look at what you've done, swallow your pride, and try something new.

you don't have to make changes overnight, but one should be able to at least display the ability to change, instead of stalwartly defending what one is currently capable of.

an artist's greatest resource isn't just the ability to see and draw, it's also the ability to listen and acknowledge as well. I've seen folks disregard what might be helpful advice, to just fob it off and continue on their merry way. It's generally not a pretty sight, especially if those same folks come around asking for your opinion on something, only to ignore it when it doesn't agree with what they want to hear.

[Tapewolf]

an artistic challenge if you will. You want to improve, yes? but you don't want to make improvements to the comic so as to not appear jarring to your audience, yes?
(which i personally find silly as a webcomic should have artistic progression, especially early on)

When I was a kid, I was idly watching 'South Pacific', which happened to be on.  At one point they either changed film stock or switched to a print kept in different conditions.  This happens in the middle of a scene, with the result that everything suddenly turns yellow.  If that had happened during a cut to an interior scene or something, it wouldn't have been so noticeable, even when they went back outside since it gives a chance to transition.

That's one of the things I'm trying to avoid.  On the other hand, as you say, staying still won't help.

However, the idea is to take a piece that you do normally, be it a past comic page or whatever, and just experiment with it, or really pull out the stops when it comes to a cover page.

Yes, I think that is the way forward.

Doing a webcomic isn't just about setting a style, it's also constantly challenging yourself and driving yourself forward, do not fall into a "style" rut. Honestly it's sometimes good to take a look at what you've done, swallow your pride, and try something new.
you don't have to make changes overnight, but one should be able to at least display the ability to change, instead of stalwartly defending what one is currently capable of.

If I didn't show some adaptability I'd still be working on everything in one layer 
I'm not completely close-minded, the techniques I'm using currently are not the same as the ones I used when I started.  I have been adding new ones to them and discarding ones that have become obsolete.  This is a process that I expect will continue.

an artist's greatest resource isn't just the ability to see and draw, it's also the ability to listen and acknowledge as well. I've seen folks disregard what might be helpful advice, to just fob it off and continue on their merry way. It's generally not a pretty sight, especially if those same folks come around asking for your opinion on something, only to ignore it when it doesn't agree with what they want to hear.

Yes, and I'm aware that is sometimes a bit of a problem for me.  But I'm still listening and still planning how I can slipstream some of the things that have been discussed here into what I'm doing.

Part of the issue we have here is that several entirely different problems have been conflated together.  Rather than focus on any one of them, the trends seems to have been on fixing them all, simultaneously.  Or at least that's how I've been interpreting it.

I may have failed Project Management at university, but I'm well aware that the best way to break something is to try and change too much of it at once.

[Turnsky]

an artist's greatest resource isn't just the ability to see and draw, it's also the ability to listen and acknowledge as well. I've seen folks disregard what might be helpful advice, to just fob it off and continue on their merry way. It's generally not a pretty sight, especially if those same folks come around asking for your opinion on something, only to ignore it when it doesn't agree with what they want to hear.

Yes, and I'm aware that is sometimes a bit of a problem for me.  But I'm still listening and still planning how I can slipstream some of the things that have been discussed here into what I'm doing.

Part of the issue we have here is that several entirely different problems have been conflated together.  Rather than focus on any one of them, the trends seems to have been on fixing them all, simultaneously.  Or at least that's how I've been interpreting it.

I may have failed Project Management at university, but I'm well aware that the best way to break something is to try and change too much of it at once.

i Wasn't just referring to you, but rather past experiences whilst a regular on the comicgen forums. but yeah, change too much and it can be jarring, however not changing at all can be worse, imagine what DMFA would've been like had Amber not moved forward or taken that break?

however, i get what you mean, in your case (no offense here, just an analogy) it'd be like the wizard of oz, where it starts in sepia tones, and when it gets to oz, it's in blazing technicolor. it's a heavy contrast for sure.

Anyways, incremental changes is what i'm advocating here, what would be a start would be the refinement in the lineart that we've talked about previously. To be fair you're also working with what Ren gives you, i can throw some suggestions his way, if he hasn't heard them already that is.  

photo blue pencil leads, archival ink pens (like pigma microns, zig millenniums, that kinda thing) and how to set up a scanner for clean lineart.
for the record, this remains an excellent reference to providing clean lineart, it's how i started, it might help you the same
http://machall.comicgenesis.com/info/art.html

(small edits made for grammar's sake)

[RobbieThe1st]

Tapewolf, I have to say: You are something of an inspiration.
After reading your post about what you'd want in a speech bubble tool, I started thinking about how one would go about that.
After realizing that Gimp outputs all of its paths in terms of cubic bezier curves, even what are supposed to be straight lines, I realized I had to find a way to deal with them.
I finally found an intersection function for Javascript, and realized that most if not all of what you want can be done directly with SVG and Javascript code.
As such, I made this tech-demo: http://robbiethe1st.afraid.org/files/svg/test/sample.html
In theory, this chunk of code should be able to take -any- bubble thats a bezier-curve-based SVG path and append a straight tail on it from any angle/distance.

Now, whether or not I actually end up with a tool as good as what you want... its really going to depend on finding some way of taking a final set of SVG-code and making a raster graphic out of it that looks nice. Firefox's own SVG implementation looks excellent, but I don't know a good way of using that to create a raster other than using a relatively-low-resolution(cropped) screenshot.