telophase: (Default)
telophase ([personal profile] telophase) wrote2005-12-05 11:48 pm
Entry tags:

Mangatalk: A question...

--

ETA: GAH but I've got to re-adjust my new monitor! Those looked almost white on it. The brightness and contrast is obviously way the hell off. I'll fix that and redo the pages tonight, when I get home.

Much, much better now. Not perfect, but muuuuch better. You may need to force a reload if your browser's cached the original, awful, images.

--

ETA yet again: Just posted a followup-type entry, coming out of some comments down below.

--

I've been musing on various things - the differences between the current crop of OEL manga and Japanese manga, backgrounds and the lack of same, detail vs. no detail, and it hit me when I cracked open Fruits Basket 12 that maybe I should just post something and see what comes of talk about it.

This is page 18 from Fruits Basket 12 (no spoilers), once normally and once with as much of the text and text balloons cleaned out as I could do (and sketchily-fixed-up pictures, just to give a rough idea). This should be the page as it looks after inking and toning and before the text balloons are drawn in (well, ok, so I expect that the mangaka draws the balloons in the pencils, and inks the balloons at the same time the rest of the panels are inked, really).

Anyway. This page has many of the same characteristics as a lot of amateur manga out there on the Web that I see - lots of empty space with tone thrown in to fill space, almost no background whatsoever, figures drawn in profile (because it's easier than 3/4 view), very little variation in line width. The question, then, is...

WHY DOES IT WORK?

ETA 2: I'm specifically asking, in this case, why the empty space, use of tone, and lack of detail and line width variation works with this page and yet not with pages of amateur manga that have the same characteristics. Yeah, the answer is, technically, "because the mangaka is a pro," but I'm trying to get at what makes the difference between the pro Japanese work and the amateur* work and much of the OEL manga that's been published so far so marked.

* or, rather, 'inexperienced,' since there's some pro-level amateurs out there - we need better terminology.

I'll probably track down some examples of what I'm talking about tonight and post a new entry with them.


Read right to left, Japanese style.

Image hosted by Photobucket.com


Image hosted by Photobucket.com



I ahve a few ideas, which I'll put here in white text on a white background so as not to prejudice your thinking. Highlight to read:

-----------------

1. Very, very,very, very, very thin lines. Fat lines don't work with this - the linework stands out a lot when there's not much of it. Somethihng like Naruto can get away with thicker lines, because there's much mroe detail. Here the mangaka is using the barest minimum to define the shapes. If she needs to make an area darker, she uses several fine lines instead of one thicker one - the floorboards and baseboard in panel 4, Tohru's hair detail.

2. The profiles arne't straight - Tohru's tilted up taling to Kunimitsu and he's tilted down, talking to her. They connect across panels, and the panels relate to each other that way. In panel 4, Tohru's head is also tilted up at Kunimitsu, still talking to him.

3. Graphic design. The darks and lights are balanced. TOnes are never thrown in just to fill space; they're very carefully placed. And they don't fill a panel or a shape solidly, either.

4. What keeps the whole page from being static with the figures is the tilt in perspective in the last panel - if the mangaka had drawn her knee in straight profile like the other two profiles, it would be dead and uninteresting.

5. No figure is exactly in the center of a panel. They're just barely slightly to one side or another, which gives a gentle rhythm and flow trhough the page. Centering something would kill it dead. This also allows the speech balloons to ever-so-gently ripple back and forth, instead of reading straight down, which adds to the movement. And that movement drives your gaze directly through the figures in each panel where there are balloons and figures.


-----------------


Anyway, that's it. I'm going to bed and I'll see your responses in the morning. This may or may not turn into an essay, but not for a while.


(and next time I do something, maybe I'll damn well try shoujo JUST SO I CAN GET AWAY WITH THIS SORT OF THING!)

[identity profile] akemi-art.livejournal.com 2005-12-06 06:19 am (UTC)(link)
Simplicity is beautiful. All that negative space in the manga makes it easy to read and focuses the attention on what is important, the visual story.

Compared to North American comics that are packed with detail, heavy inking, fully fleshed figures and backgrounds, I prefer the fruits baskets style. NA comic style is a whole lot of art to cover up the lack of substance.

[identity profile] telophase.livejournal.com 2005-12-06 02:53 pm (UTC)(link)
Except that I am deliberately comparing it to amateur stuff and to the OEL manga that has less detail, in general, not American comics. Why does the less detail here work and the less detail there not work? That's the question I'm asking.

I wouldn't say that manga, as a class, has less detail than American comics do. Just off the top of my head: Blame, Planetes, Blade of the Immortal, Qwan, Saiyuki, MPD Psycho - they're all detailed. Now, shoujo manga, in general, has less detail than American comics, but manga as a class, no.

(no subject)

[identity profile] akemi-art.livejournal.com - 2005-12-07 02:00 (UTC) - Expand

(no subject)

[identity profile] telophase.livejournal.com - 2005-12-07 02:04 (UTC) - Expand

(no subject)

[identity profile] akemi-art.livejournal.com - 2005-12-07 02:28 (UTC) - Expand

(no subject)

[identity profile] telophase.livejournal.com - 2005-12-07 02:34 (UTC) - Expand

[identity profile] telophase.livejournal.com 2005-12-06 03:45 pm (UTC)(link)
* Oh yeah - I do understand what you mean about the normal American comics style, with the heavy inking and stuff. :) I'm just trying to understand why it often doesn't work - or, perhaps "doesn't work the same way" is a better phrase - when inexperienced artists try it and in the OEL stuff I've seen that's borrowing elements of shoujo.
octopedingenue: (kyou's like what's up?)

Here's a cleaner page, for a better look at the white space

[personal profile] octopedingenue 2005-12-06 06:22 am (UTC)(link)
Fruits Basket volume 12 page 18 from scanlations

I'm thinking about it, but I haven't decided anything yet. (Other than that Natsuki Takaya owns my soul and could draw stick figures and I would love them as long as she kept the page layouts and the eyes, but that is nothing new.)

Re: Here's a cleaner page, for a better look at the white space

[identity profile] telophase.livejournal.com 2005-12-06 03:07 pm (UTC)(link)
Heh - I've discovered that my new monitor is still way the fuck off on the balance and contrast. I'll do what I can to fix that tonight, now that I've got an idea of what it looks like (arg).

I really need that second page with the text and the speech bubbles gone, for this question, arg.

[identity profile] maiteoida.livejournal.com 2005-12-06 06:29 am (UTC)(link)
Hmm, that's interesting. I've been wondering myself how shoujo can get away with being so damn minimalistic. ^_^;; (Even though I guess the closest thing I could classify my stuff as is "shoujo" I get so self conscious when doing stuff like that...it would be really interesting to see just how the mechanics of it work).

I would agree that the use of tone isn't just to fill in spaces- the danger in doing pages like that is that if one uses tone to "fill in", the pages tend to look too grey (or too busy). I think a lot of it just boils down to making things as simple as possible without compromising the overall composition. And despite how simple it makes pages seem...it's awfully complicated. ^^;

[identity profile] telophase.livejournal.com 2005-12-06 03:09 pm (UTC)(link)
Yup. :) Taking the speech bubbles out of that page really shows that - the layout itself is very simple, with a simple rhythm. And then when the bubbles go in, they add a slight, but noticeable, movement.

[identity profile] moonlit-page.livejournal.com 2005-12-06 06:35 am (UTC)(link)
i'm a layman when it comes to artwork, but the first thing i noticed was balance. rule of thirds. three horizontal panels of equal size. one vertical, perpendicular to the bottom (anchoring) box. the distance of the 'lens' from the subjects in the panel is also balanced. so are the grey tones -- one, two, three. balanced. not symmetrical, just, in harmony. this page works because, if you were scrapbooking with your photo album, it would appear clean and balanced and fresh, despite the simplicity of the drawings inside. it simply clicks on an unconscious level.

[identity profile] telophase.livejournal.com 2005-12-06 03:12 pm (UTC)(link)
I'll have to clean up a few more of the pages, some more complicated as well, from the manga to see if that holds through the rest of the book, or if the basic composition varies. Er, I'm not expressing myself clearly, but I ahvne't had caffeine yet this morning.

[identity profile] prettyism.livejournal.com 2005-12-06 03:21 pm (UTC)(link)
I've completely forgotten my Asian aesthetics class, but the little I remember about traditional Japanese aesthetics. Simplicity, juxtaposition, asymmetry etc are all part of a long tradition which may come from zen budhism(ok I'm guessing here)but the elements can be found in other areas. I've kind of given on using shoujo layouts, I decided to use the golden mean since I'm more familiar with it:D

[identity profile] telophase.livejournal.com 2005-12-06 03:50 pm (UTC)(link)
:) I think what I want to do is to winkle out the guidelines behind the simplicity, juxtaposition, and asymmetry, though. What makes this simplicity work right and yet an equally simple page not work? Basic rules of composition, I'm sure, but I want to make them explicit and understandable.

(no subject)

[identity profile] telophase.livejournal.com - 2005-12-06 16:28 (UTC) - Expand

(no subject)

[personal profile] octopedingenue - 2005-12-06 17:58 (UTC) - Expand

(no subject)

[identity profile] telophase.livejournal.com - 2005-12-06 19:03 (UTC) - Expand

(no subject)

[identity profile] prettyism.livejournal.com - 2005-12-06 20:47 (UTC) - Expand

(no subject)

[identity profile] telophase.livejournal.com - 2005-12-07 00:38 (UTC) - Expand

(no subject)

[identity profile] prettyism.livejournal.com - 2005-12-07 03:24 (UTC) - Expand
ext_3743: (Haru sweet (Flamika))

[identity profile] umadoshi.livejournal.com 2005-12-06 03:23 pm (UTC)(link)
I'm not a very visual person, so my automatic assumption has usually been that Takaya (and other gifted manga-ka) simply have an intuition about how to make their work flow. It's really interesting to see it broken down into its more scientific components and be reminded that it's a craft as well as an art. ^_^

Oh, and Tohru's talking to Kunimitsu (Kazuma's assistant) in that scene.

Ysabet, somewhat new to reading your journal

[identity profile] telophase.livejournal.com 2005-12-06 03:49 pm (UTC)(link)
:D Yeah, it *is* an intuition often, but it's the sort of intuition that's carefully cultivated through lots of graphic design classes and familiarizing yourself with graphic arts and composition - bang all that deeply into your head and practice it a lot, and *then* it becomes almost unconscious.

Thanks for the ID - I've edited it appropriately. :)

(*waves* Welcome!)
kate_nepveu: Kenren, three-quarters profile, cigarette (Saiyuki Gaiden (Kenren))

[personal profile] kate_nepveu 2005-12-06 04:28 pm (UTC)(link)
I noticed when reading the FMA manga that it was less detailed than _Saiyuki_ and also had thinner, more delicate lines, so I think you may be on to a useful guideline there.

[identity profile] telophase.livejournal.com 2005-12-06 04:51 pm (UTC)(link)
This all sort of started right after I was looking at a page submitted for the PB anthology and decided I was going to tell the artist to use more line weight variation - and had made a copy of the file and added some, which made it really pop. And then picked up Fruits Basket and the first thing I noticed was the utter lack of line width variation - and I'm trying to figure out why the hell it works when it didn't work in the page I was looking at. Part of it is that the lines are so delicate, but part of it has to be the general page composition. Hrm.

"Vary your line weights" is often a good bit of advice, but sometimes it isn't and I want to figure out the difference.

[identity profile] telophase.livejournal.com 2005-12-06 04:55 pm (UTC)(link)
I think I can forumlate a guideline here, in that the less detailed your artwork is, the less variation in line weight you need. Hm. Makes perfect sence the other way - if you've got a complex piece, you need the weight variation to make sense of the image and direct people's eyes.

(no subject)

[personal profile] kate_nepveu - 2005-12-06 17:08 (UTC) - Expand

(no subject)

[personal profile] kate_nepveu - 2005-12-06 17:09 (UTC) - Expand

(no subject)

[identity profile] telophase.livejournal.com - 2005-12-06 17:30 (UTC) - Expand

(no subject)

[identity profile] tammylee.livejournal.com - 2005-12-06 18:42 (UTC) - Expand
ext_6428: (Default)

[identity profile] coffeeandink.livejournal.com 2005-12-06 05:24 pm (UTC)(link)
This probably isn't what you're looking for, but I notice that Takaya has very carefully posed the characters for maximum emotional impact. The postures and cut-offs aren't random; they've all got a clear effect. In panel 1, Tohru is clearly in a posture of supplication: looking up, eyes wider than usual, mouth open to ask, hands clutching the phone. You don't see her for a few panels because this remains the emotional impact ... and then the final panel of her emphasizes this, having her kneeling as if before a shrine, looking up to the phone and to Hatori, with the length of the phone stand emphasizing how small and powerless she is. The other person (Hatori? Shigure?) has a closed mouth, is looking slightly down: is withdrawn and reserved. (More likely Hatori than Shigure, then.)
octopedingenue: (iruka : daddy is a verb)

[personal profile] octopedingenue 2005-12-06 06:12 pm (UTC)(link)
It's neither Hatori or Shigure; it's That Guy Whose Name I Can Never Remember who's Kyou's adoptive father Kazuma's assistant at the dojo, as Tohru is calling to make an appointment with Kazuma. And I like the placement of the cutoff of his face, nearly cutting off his eyes but not quite. The complete cutting off/disappearance of eyes is always a very significant indicator of powerful and unreadable and/or negative emotions. Here we've got just a fraction of That Guy's eyes widened eye, indicating that he's surprised by the unusual nature of Tohru's call and request for anonymity. If his eyes were cut off entirely, it would make him look at lot more suspicious and involved than he is here, where he's just acting as intermediary between Tohru and Kazuma.

[identity profile] jou.livejournal.com 2005-12-06 05:51 pm (UTC)(link)
I think it's got a lot to do with spatial relationships and the framing/cropping of an image, that lends sparse shoujou pages their flair. I notice that paneling and placment in shoujo manga draws a lot of parallels to swiss/high design. ... So really, I think that talented shoujo artists are talented designers, as well. They really know how to work the negative space to their advantage.

...That said, I'm not so sure that this page is the best example. Shoujo design is every bit as evident in the shape and placement of word balloons as in any other part of the illustration. To remove the word balloons doesn't make sense to me, as they're an important part of the overall layout. I don't think Takaya designed her page without knowing that there would be balloons full of text taking up a lot of the "white space" in her panels.

[identity profile] telophase.livejournal.com 2005-12-06 07:00 pm (UTC)(link)
That was more a tool so *I* could see the design of the panels alone, otherwise I get distracted by the lines of visual flow, which is what the balloons do. The layout itself is still very nicely done and holds together without the balloons, but it's a lot more static - they provide the structural framework within which the words flow.

(no subject)

[identity profile] jou.livejournal.com - 2005-12-06 19:18 (UTC) - Expand

(no subject)

[identity profile] telophase.livejournal.com - 2005-12-06 19:44 (UTC) - Expand

(no subject)

[identity profile] jou.livejournal.com - 2005-12-06 20:38 (UTC) - Expand

(no subject)

[identity profile] telophase.livejournal.com - 2005-12-06 20:50 (UTC) - Expand

(no subject)

[identity profile] jou.livejournal.com - 2005-12-06 21:17 (UTC) - Expand

(no subject)

[identity profile] telophase.livejournal.com - 2005-12-07 00:27 (UTC) - Expand

(no subject)

[identity profile] telophase.livejournal.com - 2005-12-07 00:34 (UTC) - Expand

(no subject)

[identity profile] telophase.livejournal.com - 2005-12-07 00:35 (UTC) - Expand

(no subject)

[identity profile] telophase.livejournal.com - 2005-12-06 19:46 (UTC) - Expand

[identity profile] tammylee.livejournal.com 2005-12-06 07:03 pm (UTC)(link)
I was going to comment on the post then read your comment which was pretty much what I was going to say anyways!

Whether by training or by natural instinct a good designer can feel the 'energy of placement'; which is why they fuss and fiddle with placing elements just so.

In my mind a sequential artist is half artist and half designer. Just being an artist is good for pinups and the actual line work but you need the designer half to figure out where and how to place elements for maximum energy/impact.

I'm specifically asking, in this case, why the empty space, use of tone, and lack of detail and line width variation works with this page and yet not with pages of amateur manga that have the same characteristics.

The way I look at it is like this; you can see a design show on television, go out and purchase exactly the same furniture you saw on the show and throw it in your living room and you will have a living room filled with nice furniture. Unless you have the designer's skills and sense of aesthetic you will not be able to reproduce what you saw on television.

(no subject)

[identity profile] telophase.livejournal.com - 2005-12-06 19:37 (UTC) - Expand

(no subject)

[identity profile] tammylee.livejournal.com - 2005-12-06 19:54 (UTC) - Expand

(no subject)

[identity profile] telophase.livejournal.com - 2005-12-07 00:38 (UTC) - Expand

[identity profile] aratama.livejournal.com 2005-12-06 09:25 pm (UTC)(link)
(well, ok, so I expect that the mangaka draws the balloons in the pencils, and inks the balloons at the same time the rest of the panels are inked, really).

Yes, they do. They ink from foreground to background, so the baloons are usually inked first.

figures drawn in profile (because it's easier than 3/4 view)

No, it's not. I think this is you pushing your personal preference onto Takaya. As I see it the profile of the girl (whatshername) is intended to direct our eyes up again and onto the next bubbles and the next face, and also to make it more obvious that the two characters are indeed speaking to each other. It's a phone-conversation variation of how manga (and movies) usually cut back and forth between two (or more) people who are talking to each other: matching their lines of sight. (Aaaaaand having highlighted your text, I can see you agree. I still disagree with that throw-off comment of yours, though.)

That having been said, any visual style (simple or not) works because it fits with the necessities of storytelling, I guess. Not having much background details (I mean actual backgrounds, not mood backgrounds) works on this page because 1) they're not moving anywhere and thus there is no necessity to reinforce where they are in every single panel (though a reminder here or there is convenient), 2) the story is not reliant on where they are physically, anyway. Examples of where point 1 might be relevant is sequences with lots of action, mostly seen in shonen manga but also quite common in shojo manga of the adventure/sci-fi/etc. genre. If there is no background, the readers will be confused as to where the characters have moved to; it's a necessity. Examples of point 2 could be stories set in foreign places (America, fantasy worlds), where there is a need to describe the world as well as the people. Or maybe it could have some relevance what time of the day it is, or what the weather is like. Neither is the case here. Additionally, the point of this scene is to show the interaction of the two characters. Too much background would maybe have distracted from this.

What I'm trying to say is ... only people looking to analyze this page will see it as "simple". Anyone who is just reading and enjoying it won't even notice, because unless something necessary is missing, there's nothing to stumble over and complain about. I personally don't see it as simple, if I have to be honest: it has what it needs, and a few things more. Also, your other points in the white text ... um, which I agree that some of those things make the art pretty, I don't agree that those are the things that make it work. This is manga, after all, and not a piece of still art. Had Takaya been less of and artist and drawn her knee at a less interesting angle, or left out things from that panel which weren't relevant to what is happening in the story (I mean everything but the fact that she's writing something down), fewer people may have called her a great artist but I seriously doubt fewer people would have enjoyed the manga.

But uh, seeing as you edited out the text bubbles, looking at this page as a page of manga perhaps wasn't what you were interested in. In that case, my comment is pointless. Ooops!

[identity profile] telophase.livejournal.com 2005-12-07 01:12 am (UTC)(link)
figures drawn in profile (because it's easier than 3/4 view)
No, it's not. I think this is you pushing your personal preference onto
Takaya


I was referring to amateur artists, not Takaya. I see a huge number of amateur manga online where the artists draw the characters almost exclusively from a front view or a profile view, both of which are usually easier to do when you're inexperienced.

Fruits Basket, however, has very few backgrounds anyway. Takaya reduces them to the fewest lines possible, and often does away with them altogether. You never get a good idea of what the interior of the house looks like other than "traditional Japanese" - you can't be quite sure where the doors in each room are, or how the bedrooms are arranged upstairs. Compare it to its polar opposite, Death Note, and you realize that Obata could tell you the location of every cushion and pillow in the hotel room.

When it comes to the question of backgrounds, what I'm interested in knowing is why does the lack of them work best with the Furuba story, and why does the hyper-detailed setting of Death Note work best for it? I've got a lot of DN doujinshi where the artists draw it with shoujo sensibility, and they really fall apart because you can't place the characters in space.

"Simple" does not mean "easy," and I am not under the impression that it does. Simple graphic design is often much harder than complex. But inexperienced people don't know this.

We have different definitions of "work". You are using "People love this" as "It works". I am using "This manga is better art than other manga". And I'm interested in knowing why.

(no subject)

[identity profile] telophase.livejournal.com - 2005-12-07 01:31 (UTC) - Expand

[identity profile] fusakugyoku.livejournal.com 2005-12-06 11:29 pm (UTC)(link)
Well, yes, Japanese emphasizes negative space, and I suppose that shoujo is more "Japanese" than shounen, simply because it's generally got more Japanese imagery, and shounen is closer to Western comics in style and storytelling. (not too strong an argument)

Shoujo's also more expressive, and Scott McCloud notes that simplicity=universality, so shoujo is more universal/relatable-to than shounen (and manga in comparison to Western comics)

It's also kinda ironic--since shoujo manga tends to come out in more monthly anthos and shounen in more weekly anthos (the majority of each), one should think that shoujo manga would have more time for details and backgrounds and stuff than shounen.

Just some thoughts. I'm interested to see this essay!

[identity profile] telophase.livejournal.com 2005-12-07 01:35 am (UTC)(link)
Eh, it's not that weak an argument - you might be able to pull together an argument that Western comic artists, even when they borrow shoujo techniques, are more steeped in Western art and art history than Japanese artists are, so it feeds into their artwork and renders it less "Japanese," for want of a better description.

I'm not entirely sure that shoujo's more relatabe to - the conventions of the visual language, I think, require more learning than does shounen, at least for people used to reading Western comics. I wish I bookmarked it, but I ran across a short online review of Fruits Basket by someone unfamiliar with shoujo tropes a few months ago, and I remember her crying out "Hasn't anyone ever told this woman about panels??"

(no subject)

[identity profile] cerusee.livejournal.com - 2005-12-07 02:21 (UTC) - Expand

(no subject)

[identity profile] telophase.livejournal.com - 2005-12-07 02:43 (UTC) - Expand

(no subject)

[identity profile] telophase.livejournal.com - 2005-12-07 02:44 (UTC) - Expand

[identity profile] redplasticglass.livejournal.com 2005-12-07 04:16 am (UTC)(link)
IMO, I think a lot of it has to do with positive/negative space balance on the page _as a whole_.

In most comics/manga how to books, and even when creators talk about their work themselves-- there's a lot of focus on how panels pace from each part, and how to move the reader's eye from one part of the page to the next. There's very little discussion (from what I've read) about how a page looks by itself? The main fault I find with most of the amateur work out there is that complete lack of design of their pages.

Their darks and lights, level of clutter vs empty spaces-- It doesn't matter which direction you chose to go art-wise if none of it balances out properly on the page. The less lines there are, the harder it is to find that proper balance.

That's why, I've always been very vocal about how anyone that wants to deal with sequential art really should take some basic classes in design and typography. (well, it's not necessary, but I think it'd be damned helpful) It helps the page balance as a whole, and it's really really important, imo.

[identity profile] telophase.livejournal.com 2005-12-07 04:21 am (UTC)(link)
Yup, good point. I think a lot of amateurs focus on the panel level of the comic instead of the page level.


(speaking of design - I've been scanning in 2-page spreads from several differen shoujo-type books that I own, and I did 2 pages from CLAMP's Clover, which is, out of the books that I own, the best example of overt graphic design.)

[identity profile] semiramis.livejournal.com 2006-08-02 10:59 am (UTC)(link)
Hi, late to the party by like seven or eight months :D

Apologies if this doesn't make as much sense as it should; I'm not quite as eloquent as you are with abstract ideas like this (I'm a terrible math tutor, haha). Also if it's redundant due to someone else already coming up with it in a different post.

Anyway. My two cents: I don't think this is really about positive/negative space at all. It's about basic shapes.

The very simplicity of the style and presentation is itself the reason why this works the way it does, I think--it makes it easier to ignore what details there are. And if you ignore the details completely to just focus on the basic shapes represented, what you're left with in each individual panel (with the exception of the second panel) is a single focal object surrounded by a bunch of some kind of text device.

The foci are themselves identifiable, primarily, by the way they're structured with the same basic shape (http://img.photobucket.com/albums/v329/shamiram/FB1.jpg) as the text devices that dominate the composition in the same panel, leading the brain to subconsciously lump them together. They are then further incorporated in the way they fit snugly into the back-and-forth motion (http://img.photobucket.com/albums/v329/shamiram/FB2.jpg) of the text devices--in fact, the overall flow of the image is, I feel, rather negatively impacted (http://img.photobucket.com/albums/v329/shamiram/FB3.jpg) if you take the focal points away.

Therefore, the visual focus in the individual, isolated panels is generated first by pattern recognition, and then by linear direction, almost as an afterthought--and most of that linear direction comes from the text devices of the page as a whole, not the imagery in the panels themselves. This is why, in the balloon-less version of panel four, the eye is drawn not to Tohru's head--the original focus--but upwards through two panels into absolutely nothing. The presence of speech bubbles, however, says, "Hi, triangles and rectangles and all those other losers not important. Circles pwn all. Look at us." Tohru's head is similar enough that that draws attention to it, but different enough that it doesn't fade in with the other circles, so your vision is more drawn to it. Too much noise in the image would compromise that sort of lopsided symmetry, because subtle visual motifs rather fade into the background unless they aren't being crowded out.

...

Or. Uh. I've never really taken a graphics design class or anything, so it's quite possible I've just pulled all this out of absolutely nowhere.