As a user of a virtual world it's more than likely that you can take a snapshot of that world. This will be either using a built in function of the virtual world client, browsers, or viewer for that world, using your own screen capture program, or your operating system's native utility. If your comic has only one main character (your avatar) and perhaps the avatars you meet in that world, then the images for your comic can be done as you like when you're logged in to that world.
Other circumstances, where there's more than one main or regular character (including where all those characters are only outfit sets of your only avatar) require more organisation. In those cases you need to pose all the characters in a scene so that you can take a snapshot. Using live avatars (those belonging to other users) requires some advance messaging and agreement on when and where the avatars will turn up, what characters they are playing, and any extra requirements (e.g. specific poses) needed for the scene. Even if you are only using your own avatar and alts/NPCs, you still need to be clear about what images you need from the shoot, or you won't get what you need.
Before you do any shoots, you need a SCRIPT, so that you know which characters are needed where, wearing what. The following is an example of a formal comic script, from Comics Experience, which has an archive of comic script formats for people to use:
Note that in the above example, the script tells us which page and panel it refers to. The description of the scene was on a previous page. The panel descriptions tell us which characters are in those panels and the actions taken. The SFX and speech are more relevant to post-production, but if a character's talking or growing, their expressions and poses should match. Now you don't have to create a script as formal as this, but you need something that will give you enough information to organise your shoots. You might like to check out How to Format a Comic Script for a more formal tutorial on comic scripts.
A scene is what happens at a particular location or set, before the story moves to another location / set. Some scenes will be short with only a single or few panels needed whereas others might continue for pages and require many panels. From the script you should make a list of scenes and for each scene note the the location/set used, and which characters appear in each scene. I call this "blocking a scene", after film an equivalent film term. Once you've blocked the scenes you will know which characters are to appear in each. The next step is assigning an "actor" for each character. There are three possible situations that affect determine this:
In an acting group, it's important to keep track of which avatar or alt can play which character. It really doesn't matter if the actor is an avatar or an alt, in both cases the owner needs to source any character outfits in the inventory. Figure # shows part of a spreadsheet used to keep track of the Seconds Crew. Each row is a either a different character or variation of a character. The letters in the columns represent a member of the Seconds Crew: E=Elsie Broek, L=Laurel Galli, M=Marble Baar, O=Optimum Actor, and D=Dummy Quicksand (yes, those are their names!):
As you can see, some characters, like Marble (Human) can be played by Laurel, Marble and Dummy. They can all play Pod People, but only Elsie can play the Harimori character (who was originally played by a friend's avatar but later was unavailable). In figure # below we see a similar spreadsheet used to determine which member of the group plays which character in a scene:
Scene 7 takes place in the D'Olby Base kitchen area and features four characters: Egbert and Elizabeth Quacken (both older versions of the character), Octobriana, and Jenny Quacken (a.k.a Jenny Everywhere). Both Laurel and Dummy can play Jenny, but only Laurel can play Elizabeth, so Dummy will play Jenny in the scene.
Scene 8 takes place in the Conference room, and has six characters but there are only five players.This will require two sets of snapshots, the first with four of the characters in (Marble as Trevor, Optimum as Dr Optimum, Elsie as Harimori, and Dummy as Jenny), and a second with the two remaining characters (Optimum as Egbert, and Laurel as Elizabeth), who'll be composited into the images of the first set in post-production.
For each panel you need at least one or more images to choose from. Like a photographer or a film director, by having more than one version of an image for a panel, you give yourself a extra choice later when deciding on the best shot. If you only take one shot for each panel, you will forced to use it, unless you do a re-shoot.
Someone has to be the director who tells others what to do, and someone has to be a camera person who takes snapshots/screenshots of each set of poses to create images for the panels of the scene. In a one person outfit that will be the same person - you! In a larger production crew that might be different people.
In either case, whichever avatar is the camera person (and they may also be an actors in a scene) is responsible for consistent sets of images. For a an image to be consistent it needs to be taken from the same viewpoint, with the same client/browser/viewer. Before taking any shots it's preferable to set the environment of a region manually, otherwise the sun's position will change over time and produce different lighting effects. For most of the Return To Second LIfe storyline, the Seconds Crew used a commercial HUD called ClubCamera HUD Mk III Box (or the ClubCamera HUD Mk III Maxi Box). Both allowed its user to save and manage 9 or 12 viewpoints from within the current region. The avatar's viewpoint could be set to any of these, and if used later in the same region, you could go back to the same viewpoints again. This made post-production compositing much easier to do as multiple images could easily be superimposed on one another.
There are other HUDS available which you can use in both Second Life Open Simulator:
Assuming you have scripts or HUDs to record positions, then What happens next depends on what sort of group you have:
After shooting is complete, some compositing may be necessary, to either create a image that never existed in the first place, or adding characters in images in situations where not all characters in that panel's image could be shot together. The following two examples show how this is done. Both examples were done using Adobe Fireworks (which has since been discontinued) but you could easy use Photoshop, GIMP, or any other editor that supports layers, blending and transparency.
This examples shows how a panel on the S-Bend page was created. As shown below, Laura_S is meeting Elsie on a bridge between worlds. In this case he bridge was shot in the virtual world of Cloud Party.
However, Cloud Party ceased operating before that scene was shot, so it was no longer possible to shoot in that world. The work around was to use the existing images and the image below was chosen for that panel.
The real life Laura S (yes, she needs to lose weight) was shot with a digital camera before a plain coloured wall and then imported to the graphics editor where the background was masked out or deleted (the right hand side was deleted by accident, but that made no difference). This is pasted and positioned in a layer above the previous image.
The Elsie character is shot separately in-world against a plain background and like the Laura_S image, is pasted and positioned on a layer above both previous layers.
The result is the image below. Further adjustments could have been made - shadows and extra highlights added, but that was skipped because of time issues.
This example is taken from a panel on the No harm done page, as shown below. There were six characters in the shot and only five avatars available, so at least one character would have to be added after the initial shot.
Marble was the camera person for this panel and his position doesn't change. The first image shot also shows a version of Optimum (played by Optimum) talking to the Jane Lane character (played by Laurel).
In the next image Optimum is gone and Daria Morgendorffer (played by Elsie) is in the background. This image was pasted into a layer above the previous image. That layer's then reduced temporarily to 70% opacity and aligned with the one below it. Then a portion of the layer is deleted/masked out so that certain bits of of the layer below show through. In this case we get to see Optimum and Jane's head (which was in a sleeping position on this layer). Then the opacity is restored to 100%.
The third image includes Elsie and Laurel (both played by themselves) in the foreground. We paste this into the layer above, and follow the same procedure as outlined for the previous image. In this case we can delete/mask everything to the left and above of Laurel and Elsie.
The final result is shown below. You can see that there are slight differences in brightness by the line of contrast above Elsie. This could have been fixed, but in the final version there's a speech balloon that obscures that.
Sometimes, no matter what you do, it all goes wrong, and you have to shoot one or more panels again. This is a good reason to save viewpoints from earlier shoots. You'll just have to go back and do it again.