Conversation tutorial, feedback please

[RJFerret]

Please let me know what I missed or anything incorrect:

 

Presumably your AI are pathing along, minding their own business, when you want them to interact, say something specific, or glance at one another. Here's how to set that up, step by step...

 

First they have to run into a trigger to initiate a "conversation", either from a pathnode (once TDM 2.02 is available probably) "Trigger" spawnarg, another trigger entity, or a stim/response trigger, or equivalent.

 

The trigger targets an "atdm:target_startconversation" entity (found in targets folder). That entity gets the spawnarg "conversation" "whatever" (substitute "whatever" with a label to identify this conversation compared to future ones in the list).

 

Before you jump in, you will need to type in which soundshaders you will be using if you want them to talk. So if you haven't already picked your lines, pop in a temporary Speaker and get the names of the soundshaders you'll want to use.

 

For example, I have an AI greet a horse with "tdm_ai_wench_greet_civilian_to_civilian", to which the horse whinnies via "animal_horse_idle", then she commiserates to it from, "tdm_ai_wench_idle". If you desire a specific line, or custom, you'll have to create your own soundshader (a text file listing an audiofile).

 

Now enter the Conversation Editor (map menu), which will create an entity if you don't already have one, be sure it's not placed in the void!

 

Click +Add down at the bottom to create a new conversation, then edit it. You'll notice toward the top, "Actors must be within talk distance" and "always facing", these cause them to move toward each other, and/or turn when the conversation is triggered, not that the conversation waits until they are in range. +Add an actor (you may have just one, or multiple AI interacting), it's/their names go here to identify who does what/when in the next step.

 

Now in the bottom you may add commands, which run in sequence (generally). You may choose for one to not wait, in which case the next will occur perhaps before the current completes. This can be good to have AI talk more naturally overlapping, rather than a lengthy pause between each spoken line.

 

Typical commands are "Talk" to say something (paste in a soundshader name you noted earlier), look at something or someone (including player1), wait a few seconds, etc. For example, my wench looks at the horse, says her line, the horse whinnies in reply, then she delivers her final line. That's it, now check to see how it works out in game.

 

(Note: if you want to mute your AI generally, so they don't use their usual vocalizations the rest of the time (perhaps not making sense leading in or out of your conversation), change "def_vocal_set" to "atdm:ai_vocal_set_mute".)

 

PS: This is not meant to be all inclusive, as there is a wiki article listing all the spawnargs, so don't want to be redundant to that, just a straightforward overview of the process. Whatcha' think?

 

PPS: In wiki now: http://wiki.thedarkmod.com/index.php?title=Conversation_Tutorial

Edited March 8, 2014 by RJFerret

[Springheel]

I think it's a good idea, but you're covering a lot of ground already done here: http://wiki.thedarkmod.com/index.php?title=Cutscenes_Part_3:_Lighting,_Placing_the_Player,_and_Conversations#Conversations

 

Have you read that? It's not in the most obvious place.

[RJFerret]

I had actually, thanks, I don't remember why I was reading into the third page of cutscenes though, probably back when I was trying to learn everything to get an idea what capabilities existed. The first thing I learned is conversations can't be done without scripting knowledge and shelved the idea of using them--is incorrect.

 

That also doesn't help anyone searching for "conversation" on the wiki I'm afraid, nor even via Google (which I know few realize can be done to get more comprehensive results).

 

Obviously it's purpose is not be a tutorial, but to get that conversation implemented in that map, hence it doesn't flow in chronological order of the process, and leaves out information to get on with it's purpose.

 

This intent is for someone who has never made a conversation in their own map before, who doesn't know if scripting is required or not, where they can find sounds for their own map, how their AI go from pathing to conversing, and might have looked at the blank conversation editor mystified, poked all around it, then tried searching for info only to conclude conversations are way too complicated and they could better spend their time on other things in their map.

 

If you don't feel this is worth providing as a wiki article, or is redundant, perhaps that page should be linked from the conversations page? (It didn't make sense to me to try to add missing general info to that page, it's already so lengthy, and would distract from it's specific intent.)

[Springheel]

It is linked, I believe.

 

I wasn't saying yours was redundant, just that you could use parts from the existing page.

[RJFerret]

Gotcha', I'm not sure what to draw, I kind of ran into that problem considering trying to edit that page to simplify it, or add general asides to it. Adding general info would interrupt it's process and distract.

 

That is also so lengthy, I was trying to keep this short and sweet, although I know I can be wordy. ;-)

[SteveL]

+1 An easy-to-find walkthrough for making a first conversation will be useful. The current article is more of a reference. If your audience is people who haven't set one up before, perhaps name some specific shaders right there in the article and add a couple more example specific actions so a first-time user can simply follow your steps and have a working conversation that they can then play with and adapt. Tutorial-style, that is.

 

One more suggestion: I think you should mention the "Actors must be within talk distance" setting. It's the top one, and fundamental to the set-up, and I reckon it could easily be misunderstood. The wording suggests to me that the conversation will not be triggered unless the actors are close, but in fact it means the actors will be made to approach one another when the conversation is triggered.

[RJFerret]

Super insights, I have had talk distance off, so thanks for that.

 

I added a couple examples into the original per your suggestion, hopefully they read well!