Weekly 2018-04-02

Hey Patrons! A period of easter holidays started for me here; so I traveled to Poland for visiting the family of my wife. As I was traveling on Sunday, I'm writing only now the weekly Sunday blog entry. 

The main task on my radar last week was to end my rendering researches. I'll try to explain it even if it's not easy. The artwork I posted "The rendezvous" (on top) was the conclusion of this research period and also a mini laboratory full of experimentations. The landscape part was done from scratch during the Wednesday live-streaming (replay here). I commented in details all the steps during 3h. You can find source, full resolution and wallpapers derivate of this artwork in many format on the blog as well. 

( comparing Inking+colorization results 3h30 VS Painting results 2h30 )

I ended my research after painting the characters two time from scratch and comparing the two options. It's something I never did on the same artwork before. I compared; "inking the sketch and then use colorize mask to get flat area, shade it, detail it" (on left on picture above) VS "Painting-over sketch" (on right). 

( the base sketch, a scan from my sketchbook; HB pencil )

( Inking step in Krita over the sketch turned into bright blue )

( While digital inking on Cintiq13HD and Krita 4.0 )

Conclusion: line-art colored artworks simply doesn't work with me. I always was frustrated by the rendering of all episodes I did this way... Now I understand why. It's because of the painting experience to produce this artworks and the final rendering, no matter how good or stylized is the inking, no matter the technique I use (digital/traditional/pencil): the result is always too rigid for my feelings and the coloring always feel like a pass to decorate or support the inking. In this situation, coloring become a burden (a sub-process), because the artwork feels already "done" after inking. 

( gif animation: Krita 4.0 colorize-mask in action )

Colorize-mask and other GMIC tools were designed to reduce the pain and they are really nice (especially for professional colorist). These inking lines are like hard limits, boundaries happening too early in the process (especially if I control myself the full production stack). Inking prevents the creativity or the possibility to fix things during the coloring process. That's sad because the volumes, the pose, the expression of a picture can be improved during the many hours spent at colorizing it. It just need to be done in a cumulative way; starting from large brushes to thin brushes. But I understand why this inking workflow is widely used in the industry: if I had to work with a colorist less skilled than me, I'm sure I'll favor this workflow to keep control over the design, expressions and details before someone who come later in the process can change it!  I also understand better now why animation, comic studios use that too. You can validate a pre-result earlier, and then outsource finishing the piece for cheaper later... But does it make sens if i'm working alone? I feel I just monkey reproduced what the industry wildly does because it feels safe to be part of the group; and it is hard to contradict something everybody does. It was easier to feel I was "doing comic the wrong way". Now I know I only developed a very customized approach thanks to my year as a painter, illustrator and concept-artist. 

( Inking+colored workflow: from left to right, sketch => inking => flatting => shading )

( painting workflow: rough flatting direct over sketch => rough shading => paintover )

A positive surprise in this comparison was revealed by my timer : Inking+color wasn't more productive in itself. 3h30min for a colored inking artwork VS 2h30 for painting-over the sketch... That's why I'm pretty convinced I had a false idea in mind at considering all the Inking+color workflow more productive by default. I was 100% sure about it and being convinced about this idea was what triggered all this long research about inking. Because I just wanted to be faster, now I know it's obsolete to look into this direction; in my case.

( close up: Painting result : more brush work, better volume and more texture )

So, I'm back at using my old illustration/concept-art technique: painting-over my sketch. For sure this technique still has many pros and cons. It feels longer because the "end product" appears later in time than with inking (in inking, facial expression appears quickly and are definitive) ; but I'm faster at completing the full process. I also like better the way  the rendering is more organic with it. It feels less industrial and more the output of a human craft. this is very important for me. So, in order to improve this workflow I need to work on the economy of my stroke to avoid painting over and over the same area. Over smoothing things, making lines weaker, over detailing a part are the danger of this process. This process keeps me more awake during the long painting hours because every stroke matters to build the result and it's never too late to fix something (the position of an arm, an eye, etc). 

So I have my answers now, I'll be able focus on perfecting this process along the way of future episodes and illustrations. 

GNU/linux distro issues

(Apport in Ubuntu reporting a crash of Gnome-shell )

Last week on monday, my GNU/Linux system broke (a Ubuntu GNOME 17.10): it was a mix of slow-down caused by a newer Nvidia driver not really doing a good mix with my graphic card, cintiq, dual monitor, colorimeter and Gnome itself. It was affecting the latency while painting, so all my stroke were super slow and it was not funny for painting. I hate when this type of things happen, but it's not the first time and probably not the last time! 

So, I spent time to investigate because of course, it takes time to test and discover the culprit, it's not obvious ; the slow-down could be from Krita, the wacom driver, the tablet hardware itself, a crazy process looping in background, the GNOME compositor or the graphic driver... 

( my Kubuntu desktop 17.10 ;  I really like the capacity of the Dolphin file manager to display large thumbnails, it helps at managing different version of artworks. )

I started to get answer after trying another D.E (desktop environment) on the same machine and see none of the bug I experienced existed on them. The problems with those tests and researches ; they are destrutive. Installing other D.E. side by side, testing new graphic driver often results in a big mess of the installation. So, even if I was a bit tired to have to reinstall *again* my computer; I decided it was necessary to do it for working and a couple of live ISO tested later; I switched to Kubuntu 17.10 with the backport ppa to get Plasma5.12LTS. 

So far, I enjoy the move to Plasma5 (I was a KDE user for years around 2011, so its easy to readapt). It's far to be bug-free and it doesn't have all the tools for graphism ready ( nothing for colorimeter or Wacom tablet by default, as in GNOME ) but I could adapt by adding a bit of command-line magick, shell script and porting my nautilus script as dolphin services... I hope Kubuntu 18.04 LTS (the next release, planned in end of April) will be polished ; because Kubuntu 17.10 life support will end in three month and I'll have to upgrade to that.  I'll test it to see if it could be a good candidate distribution for refactoring my "Linux for Painter" guide on my blog  :)

That's all for this week, 

Thank you for your support!

-David