Desktop Tuesday: Tuning Conversations!

I liked the latest, the conversation series on DT and its extended explanation on adding features to Stonehearth. :merry:

I appreciated the survey link in this DT. I hope it helps you all.

And does that mean that 23 will be a crafting update and 24/25 the water/building engine updates?


Wow, I wasn’t sure what I wanted to see as a wrap to this conversations deep-dive, but now I know… it was that!

It’s so great to see how this feature came together, especially how our feedback over the last few weeks has shaped recent patches to help tune conversations. There’s a sort of satisfaction and closure to see those earlier queries and concerns being addressed in this video, as well as the team’s internal QA process, to show how you guys handle that tuning process.

Looking forward to future DT’s, I’d love to see a dive into the building improvements, with a particular focus on any functions which have been added or changed recently. Building has always been a “split” subject – for some of us it’s natural and easy to stack up masterpieces, while to others among us it’s an arcane art and liable to make something explode if looked at from the wrong angle hahah! So, a deep-dive into the whys and hows of the ‘new’ building editor (by which I mean any changes done since the last major focus on the editor) would probably be useful.

Of course, when the water improvements are finished you’ll have to spend a few weeks on that!

I’m personally interested in the subject of technical debt, but I imagine that might be a bit dry for those players who are more interested in gaming and less interested in game dev… depending on how complicated a subject that presents, it may not be suitable for a quicker series; and a long series may bore the audience, but it’s one to think about at least :merry:


The explanation is really great and detailed. And I’d really love to see such series talking about water system, just cannot help wondering how the cubemitters are going to simulate the flow of fluid (like in a waterfall) :grinning:.

btw, is Candyland mod still updating till today? :joy:


I will start paying more attention to my hearthlings conversations after watching the example with the doors in this DT. Up until now, i have not seen their chatter as anything but anoying, abrupting work in the settlement and delaying everything even more than it already is.
And still…Might just be me…I dont get the same information out of the image of a door that you aparently do. If you had not said in the video what they where talking about, i would not have registred it.
Maybe if you play the game just to observe your hearthlings? I am usually to busy planing out new structures and other tasks to just sit and watch it as a fishtank…

Well i hope that this system will evolve in the future and i will go: aaaahhh that was what they meant! :jubilant:
Have a great week and thanks for all the great work you do, so we as players can get some entertainment.


I think 90% of that example is inferred from the animations used and the speech bubbles – the conversations aren’t so specific, it’s just a random subject pulled from a list and then a chain of a few reactions, followed by a dice roll over whether the hearthlings agree or disagree (actually I think it’s a little more complicated, but I know at some stage there’s a discrete “yes, I agree with that last statement” or “no, I disagree with that”), and then a summary animation to denote whether agreement (and its associated happiness boost) has been reached or not.

The system is rigged so it should look like they’re discussing a topic and having a back-and-forth, but it’s still up to the player to bring a story to that conversation. Just like how it’s up to us to bring a story to the rest of the town. It’s not required, some people are just into that kind of thing so the option is there for those who enjoy it :merry:


Originally I had stated that a few weeks ago that I understood why this conversation system was needed. Today, I’m still that way, but after hearing what @sdee was claiming was going on, I feel like we’ve only gotten half of the pie.

In most of the way, this system still tells us what’s going on in our town, but as stated by @YetiChow, it’s also leaving A LOT to makeup, that really we shouldn’t be trying to make up because then we lose that touch of what’s going on with our Hearthlings.

So with that stated, is the conversation system really done, and thus being left feeling incomplete and borderline useless, or is it on hold till the next part of the puzzle is made?


To answer the question for myself of what I like to see; I actually like the way yall are going about it. As I like seeing the depth to what is being implemented and generally what is going on with certain features to come. Development itself is something I find fascinating. I like seeing what makes changes change and decisions of reasons what they are throughout the project. After all it is growing more and more entertaining with each tinkering, addition, and change.

Going to be interesting what my hearthlings say from my old save Sorenia, when it comes to stable. :slight_smile:


This system is incomplete, but it’s the base layer of a much stronger system to come.


I will agree with this a little, i think the main reason is that the body’s motions are clear and emotional, but the problem is that the two conflict, if you know about delivering information (not that i’m an expert, but i heard it somewhere) there’s a point that Text and Visuals compete, and having bolth at the same time (without naration) will actually make the delivery more ineffective
This also is a thing with visual design, having text makes the text a very gravitating part of the image for the eyes

I think SH has the same issue here, they’re competing, the hearthling’s animations are great, they give context to the simple ‘like/hate/meh’ icons above, but the problem is that the player is looking at the conversation bubble and not the Hearthling’s motion, the bubble has Reeealy high contrast with the background, so it pops, but the Hearthling has comparatively weak contrast, and even if the motion is clear in testing, in the actual game, the silhouette is not strong enough to compete with the icons and bubbles above, and the fact that the animations are quite slow don’t help either, because motion attracts the eye, and those are not fast enough to be interesting to the eyes

And i guess this is a task for @malley (also, sorry to have thees feedback after the release and all the work, but it just stuck me now)

I think there are two ways to fix/improve this, and first is to have the animations not shorter, but snappier,
Here, this might help

At about 2:00 they show a clip from Steven Universe, and i think it should look more like that in motion, to put it the othet way, imagine that same clip, but all poses were smoothly connecting to each other, in other words, the animation was always moving smoothly.
It’s definitely more ‘realistic’ but it would be much less readable because the viewer dosn’t have the time to take in the emotions of the pose, and the constant movement will distract the viewer from the overal pose

A good example is the emote for the conversation where the Hearthling kind of points here and there and then tosses the conversation back to the other Hearthling, it’s good, but because the hand never really stops during the pointing, i personally felt like he/she was just moving/waving their hands rather then pointing in any sense, and it was quite confusing because i almost lost track of the motion and didn’t read any poses beside the ending where they point to their own hand and tosses it back because i also had to look at the speach bubble (it has a ‘…’ but it still competes)

I feel like Smooth animations work for most actions because it doesn’t have to compete, but here where it’s competing with the speach bubble, it has to be ‘Dumbed Down’ so the pose/silhouette and emotion has time to sink in and deliver it’s part

The second way would unsurprisingly enough be the face, i understand Scale in animations have been pushed a bit back due to tech reasons and that’s fine, but this is where it really mattered, because the part of a Harthling that has the Highest contrast is their eyes because they’re pure black (not so much for darker skins, but it is still quite visible for them) and as i mentioned earlier, the body-language is weak because of the lak of contrast, but the Eyes are currently one of the few things that can compete with the speach bubble, aaaaaand they’re still snd emotionless, again, hindering the delivery of context by the animation because it is the single most visible part of the Hearthling that has high contrast and is high enough to be close to the speach bubble within the player’s view

Another small way would be the voices, if you were to add some mumbling of some sort that effectively deliveres the tone and emotions, that would greatly help the context be delivered and aid the animations to be read because the player doesn’t have to pay as much attention to the speach bubble anymore

Honestly? If we were going for just making the Hearthlings feel more alive, and not care about commemorating your actions in the town, i’d just remove the Speach Bubbles, as for why? Well look at how Starbound’s NPC’s interact with each other, sure they have tags or speaches later to hint at what happened, but at the very moment of the conversation? They deliver everything perfectly with simple poses, faces, and motion (not even sound)

But i understand that SH has a goal of commemorating your actions, so more complications yay

I hope that all made sense and i didn’t contradict myself or something, but that is hoe i feel looking at the delivery is in honest and critical light


I wouldn’t describe the system as incomplete. What it’s intended to do is create a means to add flavor to your town and provide a light reflection on what is going on, while also providing a mechanism for hearthlings to interact directly with each other in a structured way. Now that we have a structured interaction there are ways we can build on it, such as with adding a social system, but that doesn’t mean that we feel the current implementation is incomplete.


Thanks for the critiques : ).

Just to get one quick thing out of the way - the reason there is no audio (or no new audio in general) is because we don’t currently have someone who makes audio at the studio : /. I believe we are looking to hire, but its a low priority right now.

As for what we’re going for, the speech bubbles and animations are supposed to be in cahoots - not competing with one another. The icons are supposed to convey the subject matter and the general structure of what is being said, while the animations are supposed to convey emotional impact of the delivery - and generally just add some visual interest. In a sense, the animations are purely fluff… but in another sense, they are almost the entire point of the conversation. Overall, the purpose of having conversations is to make the hearthlings feel more alive and sentient, instead of like little automatons. This does mean that they work slower (sorry), but that’s… more real too. I mean, think about it, do you get up every day, eat breakfast, go to school/work, eat lunch, go home, eat dinner, do your homework, then go to sleep? Of course you do, but that describes pretty much everyone and doesn’t make you a “sentient being”. That comes from how you do those things:
Are you quick out of bed, or do you hit snooze 5 times?
Do you eat eggs and bacon, cereal, or grab a quick granola bar and eat on the bus?
Do you go to highschool or college, or do you work as a doctor or banker or football superstar?

I think you get my point - and while we didn’t try to address these kinds of things, there is a similar connection to how you speak when in a conversation. Its not just a matter of saying “yes” or “no” or “maybe”, its how you say that. So yeah, the icons are supposed to be what you are saying, while the animations are how you are saying it.

Hopefully that makes sense? That’s interesting that you guys see them as competing bits of information though, if this continues to be a concern for you after playing for a bit longer, please let us know - we could be missing something.

On to actual animation critiques. I fully agree with your critique of the talk_generic, it feels very floaty and never really rests in a single pose. I will put it into my list of things to do : ).

I don’t think we should be pushing towards realism though, I would rather us go more towards a cartoony style.

I mean, check out the conversation poses in this example from Emperor’s New Groove:

Or Robinhood:

Or check out these great examples in Lady and the Tramp:

There are huge limitations we face though (most of which aren’t unique to our game): like we can’t control where the viewer is watching the scene from, we can’t make unique character faces and body types, we often can’t control the environment around the conversation - we aren’t constructing a scene like in a movie, etc. That said, I would still like to strive towards a very stylized cartoony animation style like you see in old (or new) disney/pixar films.

I am not an animator by trade, but am actually a generalist - which is not an excuse, its a reason ; ). I am getting better overtime, but I greatly value your critiques like the above (they help me to get better!). I’m sorry to say it will be a while before I have time to actually update the animations, but any feedback like this will be reviewed and jotted down to fix later if we agree : ).


Ok…but now we need hearthlings to pick up the critters and have conversations with them! Please make that happen…please! :jubilant:


Sorry. Couldn’t help myself.
Original by olegti.


Speaking about future DT topics, I’d like to see more about how you conduct your community work. I guess we all know the ways you make yourself more… public (trello, twitch, Steam news updates, DTs). But I wonder how exactly does the feedback work. What are the priorities, any favourite ways to collect info, who and how often watches the threads, how is it incorporated into making changes in the game, etc.


I definitely like the conversation feature and I haven’t tried it out just yet but I hope it is just there to add more character to the game and it wouldn’t interupt the player. In addition to this feature I would love to see the Harthlings create relationships with others including animals(that the shepherd brouight in) maybe. It could cause them to maybe be happy when their harthling foe gets injured or unhappy when a friend does. Maybe even wars between groups of people in your Stonehearth village.

1 Like

Hehe, conversations with pets - coming soon. Picking up pets… prolly not coming soon : /. I actually talked about the complications with picking up pets on a stream a while back (not sure which number, sorry), but basically… its hard and may be a bit before it becomes a priority that we need to sort out.


Ah, i think my delivery was bad here :sweat_smile: what i meant by ‘Compete’ is that they fight for the viewers attention and not that the content in them fight each other, i mean that because the Speach bubble is such a strong attraction to the eyes the rest of the body kinda fuzzes out (maybe ask @Allie ? I bet she will be able to understand what i mean, being the artist and all)

Just to test this, in the DT, where they’re talking about doors, try muting the audio - forget of what it was about for a bit, and pull your eyes to the speach bubble and hold it there, focus on it, can you see the pose clearly and get context from it? Because it’s out of focus it’s difficult to read the small details in hand movement, the most you get is some motion and the head tilts

The thought goes(for me)

  • what is in that bubble?
  • oh, it’s a door
  • so they’re talking about a door they went through in that house
  • what about the door?
  • (aaand the animation is already almost over)
    I mean, it’s a bit exaggerated, but you get what i mean, they need to be able to see them at the same time

I was trying to say that the animations of the body might be better off holding a pose for a relatively long time to get a stronger impression and convey how they talk. After-all the emotion and gist of the conversation by icons isn’t enough to show what they’re saying, the rhythm the animations give allow the player to go and follow and speak over it

For example the pointing thing again:
When i look at that animation in get the impression ot "So there’s [A], and considering [B] how do you feel about [A]?"
And that’s what i get from the Rythm of the motion, it gives the small halts and emphasis that you give when speaking, and we base our own dialogue based on it

As i mentioned above, the rhythm and context doesn’t convey, not because the animation is not correct, but the viewer doesn’t see it, they’re too focused and distracted by the speach bubble that they don’t see the small details in the animation, giving them more snappy pauses and held poses i think will allow the viewer to read that pose while it’s playing (they won’t be focused on it but the held silhouette will be enough to read)

As for faces, yeah it’s tech, i was trying to say that like above, they compete with the speach bubble, and the eyes are one of the few things that can hold a fight against it, but them being not animated hurts the delivery quite a bit

The film animations i think are so successful because just like above, they don’t compete with Text, it’s like those old Disney and Charlie Zepplin films that have no real speaking sound, but with having icons pop up instead in Typography animation style with no audio and with equal emphasis on bolth the icons and the animation, sure you get a bit more information, but your eyes are too slow to see all of them and you’ll need bolth to really get what they’re saying(and it’s worse sinse you can’t really see the motion of the character, because it’s more important than the icons)

I definitely like the more exaggerated and expressive nature of old Disney and Pixar films (and new ones) i just don’t theink they perfectly match with the circumstances, at-least without some touching up to make it fit a bit more


Just as an example i found this

You can see that the character/mascot becomes clearer and clearer More comparable in strength with the text, eventually getting rid of the text but the third i think is a good balance for what we have here (maybe still a bit heavy in the text)

1 Like

To put it another way the human mind can only really hold so many things in focus I think it’s like 7 or 8 so your animations are literally fighting for one of those slots and the speech bubbles are fighting for one also really most folks are going to view one or the other and not both as a single thing. Maybe lessen the brightness of the bubbles and let the hearthlings pop a bit more from their backdrop. Make them feel line one cohesive focus instead of two fighting enities

1 Like

I’m pretty sure I get what you’re saying, but I with the animations being roughly 5 seconds or so (some are fairly short ~2 seconds, but most are closer to 5), there should be more than enough time for a player to glance at the icon, then glance back down and watch the animation. Plus overtime the icons seem less interesting to me - as they’re just a call-out of information - while the animation seems to mean more and more over time.

Please give it a little time and think about it when a conversation catches your eye. I think you will find that the icons are almost always the same and stop being important once you know what they’re talking about - then your eye pretty much only stays on the animations, which give you a better sense of what the hearthlings are ‘feeling’.