Not every game can afford voice actors for every NPC, so they present their dialogues differently. I'm teetering between their dialogue appearing above their heads if it's isometric with choices presented as such and in first person a transparent overlay with their dialogue and options present that still communicate you are talking to someone but are invested in the writing as opposed to their character model making poses. I never liked cinematic cameras and posing if there's no VA. Always seemed weirdly stilted because you're expecting a voice and there is none. I will however take large portraits that change expression according to dialogue.

My only concern is they make it clear who's talking. When text appears, I want to know which of the four people present is saying it. If it's just a line or two, speech bubbles are usually enough. But bubbles quickly become annoying when a lot of text is being said.

I think colored text or speech bubbles does a great job of tying text to a person. This guy uses gold text, or that guy uses green speech bubbles, etc.

If there's no spoken dialogue, I prefer the frame with an image of the speaker and the text beside it. It makes it clear who's speaking and gives a certain weight to the script. Speech bubbles are ok for random ingame chatter, but it gets annoying for heavier dialogue.

For immersion, onscreen text is better.
Text on the bottom is the easy way out, but it takes the eye focus away from the game.
Speech bubbles or - even better - text over the actors heads with a text shadow allows more dynamic, you can also play animations while displaying them.
Also give the speakers different text colors, which makes the distinction more intuitive and allows to have several people speak at once. Up to a certain degree it also gives an feeling of direction, by making the player focus on a certain part of the screen you can give him a impression from where the "sound" is coming from. Of course this won't work on everyone, some lack the immagination.
Use it as a tool to direct the players focus, his eyes will be where the action is happening.

For reference play Indy 4. They did a great job there. You never ever have to think half a second about who is talking. Colors and positioning help to let multiple people talk at once.

Indy 3 had the text on top of the screen, which also works, but not as good, it's by far less dynamic, there is no direction and you don't look at the main screen while reading. Still ... if you check the scene right at the beginning when Indy enters the classroom or when Elsa adresses Dr. Jones, the different colors help a lot.

This even works in a VN with only two closely zoomed characters.

When you use text in panels with the speakers portrait displayed, you basically have to pause the game. The dialogue breaks the dynamic, stops the gameplay. This is good if you have some sort of action or tactics game, with this you make clear that the player can relax for a second and read, this moment the story is carried forward. Those who are not interested in the dialogue in any way will prefer this method when they want to skip through the "cutscene" and go on playing, because also the end of this scene is marked quite clearly by making the texts and portraits disappear.
You can also mix here. Important dialogues get panels, background chatter is onscreen overhead.


Just play a bunch of different adventure games and see what works best for you, what gives you the best immersion.

ps: People only expect a voice if by now they used to. Most adventure players grew up with games without voicing and also many modern indie games don't have voices.

consistent, apart from that it does not really matter.

That may not work for colorblind players. Having color being the only distinction between the speakers can lead to confusion if the player can't tell them apart.

Undertale did two interesting things here:
* Each major character has a different font.
* Each major character also has a different text sound.

I note that Earthbound does the different font thing at least once, and I remember Banjo-Kazooie doing different sounds, as well.

I guess I'm fine with whatever is on offer.
I don't think I have any preferences in that regard.

The only real necessity: it needs to be clear who's talking, else a conversation feels "wrong"(?) - though, usually, that's also not a problem, since you (again: usually) address one person in particular.

And I think it's actually pretty rare that a group of NPCs is talking in a jumble.

Color blind players will have problems either way, no matter if you have distinctive colors or not. But sure, no one stops any dev from using additional techniques.

Thing is, if there's text that needs to be read by the player, then either:
* The sequence with the text is taking the focus away from the game, anyway.
* The text *is* the game, and is the thing the player needs to interact with and be concerned with at the moment.

By the way, worth noting that:
* Dragon Quest games, and SNES-era (and NES and PSX) Final Fantasy games just use a text box (though in FF games often the box is at the *top* of the screen).
* Romancing SaGa games, on the other hand, use speech bubbles, as do the SaGa Frontier games. In fact, the Romancing SaGa games go even as far as using speech bubbles for things like the names of the skills party members use, and even post-battle stat gains. (The Game Boy SaGas do not, as the hardware is not as powerful, and the series didn't have its conventions established at that point.) This sort of pause is also good if the dialog is something the player has to interact with, like in WRPGs. (This occasionally happens in JRPGs, but is far less prominent there; I'd consider this to be, perhaps, one of the big differences between JRPGs and WRPGs. Of course, this is excluding things like shop dialog, which does happen in JRPGs and is sometimes on a separate screen.) I seem to remember this situation arising pretty regularly in the Final Fantasy series, starting with Final Fantasy 4.

By the way, one other convention that I've seen (Game Boy SaGa, for example, and I believe Dragon Quest series as well) is to start the dialog with the character's name, followed by a colon (:), then the actual spoken words. This, I'm pretty sure, is the exact same approach that plays, when written out, have used for far longer than video games have been around.

Btw, if the game is played from the players view and his text is not to be shown on screen, then more important than the presentation is the way how the others react to it.

"You say that you are not interested in this item I have to sell and rather have me stick it where the sun never shines?" ... is the most stupid way to do it. This is sadly something that you find in many japanese games. I find this an insult to the intelligence of the reader.
The NPCs should react to the sentence, not repeat it word by word. Leave something to the players immagination.

For small screens and tiny characters or games that have no talking animation, that's probably best. Once you have animated chars of a certain size however it's not really necessary anymore, especially if the text is right above the character. And yes, of course reading takes the players focus away from the screen (no matter how it's presented), but he would still see the animation if the text was not positioned too far from the character.

Try immagining LEC adventures where the name was alyways shown.

A lot of good ground already covered in here. Everyone's said a lot of good things.

Make it clear who is talking.

I prefer boxes rather than bubbles for normal dialog (bubbles are OK for incidental chatter). But speech bubbles can work (Final Fantasy Tactics had a big cast per scene and they worked there, because they were sufficiently big they didn't really count as just bubbles anymore.)

Play with font, sounds effects, etc, for speech. Keep contrast sufficiently high between background and text.

I'm contrary to most and I like the "voiceover" "blahblahblah"1 or "ahh", "hmm", "oh!", "yes?" type 'voiceovers' too -- small bits that are much easier and cheaper to get. The blahblahblah types don't even have to be real language, just distinct. The "ah", "hmm", etc, types are often used in conjunction with character portraits that show an emotion.

I'm also a fan when color/font/whatever is reserved for key words, phrases, etc. So long as it's used consistently throughout the game. Especially when there's a good in-game reference for them.

A few issues: NEVER mess with the timing. Dialog should NEVER be auto-reveal; always character prompt. I prefer player-controlled scrolling dialog boxes (up and down), but I understand the "push button to continue" boxes too, since it can add to the medium with the pacing of the boxes. (But please, include the "push button for entire text box so far revealed" so the player can go back to). I've lost track of the lines of dialog I've irrecoverably missed because auto-reveal was too fast. And there are innumerable people who refuse to replay Dragon Quest Builders 2 [a great game] because of its horrible torturingly slow auto-reveals.

For font, even in a pixel-game, please use a fully-featured font. Pixel fonts can be notoriously hard to read, and any 'jarring' or 'does not match' one might experience between pixel fonts and the game art is far more easily overcome and forgotten by the player than "I can't read that". (Undertale always looks like an unreadable mess when anyone tries to convince me to play it.)

A game that does important dialog in a bad way: Hyrule Warriors. It's normal in "Koei Warriors" games to have the dialog just happen during battle with boxes appearing in the screen, and often with spoken lines too. The things that progress the plot of the mission and also inform the player about the status of the battle (i.e., "Nobunaga is getting his ass kicked, go rescue him else you'll lose the mission."). But Hyrule Warriors was really bad at it: so much of it, not well alerted to the players, etc. YT videos of in-mission will easily show how missable it is while you're focusing on playing the levels. Odd that they broke it after doing dozens of games that did it right.2

And please don't use comic book conventions. Comic books are notoriously hard to read: Don't use allcaps, or smallcaps. Or weirdly drawn/shaped (or animated) text boxes or speech bubbles.

1: Example of blahblah: Famously, Simglish in The Sims. Here's a clip from Band of Bugs (which was DRM-free on PC direct from the developer at some point): https://youtu.be/6igdFJOn4HQ?t=1388

2: Example video of Hyrules Warriors doing it wrong: https://youtu.be/pP424NSatiA?t=294 Watch for a minute and notice how you can't easily play and read allllllll the dialog at the bottom left.

Like in Banjo-Kazooie? (Also Donkey Kong 64 and Banjo-Tooie.)

The problem I have with comic books is that it's not clear in which order the text is meant to be read.

Many column formats have this same issue.

Banjo-Kazooie is another great example of the blahblah style. I dig.

Maybe I'm used to reading visual novels that I'm fine with having NPC dialogues being presented at the bottom of the screen, as long as there is a clear indication as to who's speaking. It could be just a text showing the name of the NPC, or having the camera focus on the body of the NPC whenever they're the one speaking. Or if the game developer decided to draw a portrait for every single of its NPCs, then one way to tell the player that an NPC is talking apart from showing its portrait is to first reduce the portrait's brightness and increase it when the NPC is talking. That is especially useful when the game is currently presenting you with more than one NPCs talking with one another.

And I'm also fine with having the dialogue be presented inside a box next to the portrait of the NPC, like Stardew Valley for example.