A lot of comics it is NOT obvious who is speaking until you read a few more panels, or see the action. Sometimes this is intentional and interesting use of the medium. Other times it isn't - which I think shows it's a hard thing to get right.
Second - I am assuming constant font, size and panel space that can potentially be used for text. This is true of single panel cartoons (DoctorFun?, TheFarSide?) but not of word bubble ones. It will do as an approximation though.
Anywhere there might be a word bubble, there might be some text. Text is just a string of characters. The presence or absence of text in any given place can be expressed as binary, 1 for there is text there, 0 for there is not. The numbers produced aren't distributed evenly - some text layouts are far more common - others look bad so are less used. That gives an information limit, and I suggest not a very large one. Then you have the text itself - and that obviously has a bandwidth.
Now, within this bandwidth, you have to convey many things.
Style (nice layouts etc.)
And anything else you want your comic to say.
Now, while obviously, these are not orthogonal vectors - neither do they compress prefectly.
Remember that this information channel in a comic is very low. A page may contain only ten or twenty words, and maybe a couple of deliberate silences (as opposed to ones used for placement)
Each combination of symbols only has one intended set of meanings.
Gah, I've gone all mathsy. But I think there really is a communications bandwidth limit in place.
I'm not entirely sure I understand your point. Surely in any form of communication, direct or indirect, there is a limit to the amount of EmotionalBandwidth (or any type of bandwidth). I think your definition leaves quite a lot to be desired though. In many manga (TokyoBabylon springs to mind) it is precisely the pages without dialogue/text that convey the most. You can be quite happily skimming through pages of dialogue only to come to a complete stop when you hit a double page spread with a single frame as you try to comprehend the implications of what is in front of you. --Kazuhiko
All I was trying to say is that a comic has less bandwidth than many other media, and has to carry just as much. So the emotional bandwidth is often sacrificed. This was in context of me being unable to identify most of the characters in PhoenixFeathers and trying to justify it as anything other than 'because I am dumb' As for silence conveying a lot - I agree. Especially when it is a big change from the preceding stuff. But InformationTheory covers that too - it's the unexpected symbols that convey the most information. (Plus, as you say, it makes you slow down and NOTICE the small details, which are often there in the complex scenes too) --Vitenka
DouglasReay thinks you may be missing a point. Which is that a single word can carry a great deal of information with the right context. The word "foo" can carry the meaning of the entire book "The Hobbit", if you also supply a dictionary which defines "foo" to mean "In a hole, in the ground, there lives a Hobbit....". In the case of graphic novels (aka cartoons or manga), the dictionary supplying the additional (and changing) context for the meanings supplied by the words are the drawings. The tilt of an eyebrow says whether a word is meant in jest or as a threat. Or whether it means "Do you still love me?".
I agree that this is the ideal - but generally by the time you get to know the characters well enough for this to work the comic has ended. The exception to this is extremely crude visual shorthand (such as anger lines) which do work - but don't add a huge amount of bandwidth. Anyway - my original point was that a comic carries less bandwidth than a movie, to which the above also applies (only, you know, better). --Vitenka
'By the time you get to know the characters well enough' depends on the skill with drawing. If the artist is good enough at drawing the characters consistently, you don't need all that many pictures to get to the stage where an unexpected expression can be meaningful. --Angoel
Take Neil Gaiman's Sandman Series, say #50 Ramadan, illustrated by P. Craig Russell. [this page, for example]. My guess is that, if you tried to put down in text alone, all the descriptions, the emotional subtext shown by the expressions, body postures, etc in sufficient detail that it had the same emotional impact as the original text+drawings, it would take up more space.
Oooh, dodgy seeming site. I think I agree though - for that example a visual medium is the right thing to use. But look at what has been sacrificed to get that lovely tone in. Plot character and action have all had to give way. --Vitenka