Sorry about the title, I just wanted to show my emotions about the current trend of voiced dialogs. :banghead: :yuck:

This might not seem so, but this is the relatively brief version of what I want to say about dialogs and voice acting, and if anyone wants me to go into more detail about any part of the following text, please inform me, as I have thought about these matters for years now, and I was versatile enough to start afresh whenever needed.

The more I think about it, the more I am convinced that the root of a lot of problems in Oblivion was the decision to make it fully voice acted without the proper technology to implement it the right way, as it handicapped the quest builders a lot, and wasted a lot of time, and filled a lot of space, and used a lot of budget that could be used in a better way.

The voice actors needed to recite every line of text in the dialogs one or more times, and for each race the same line was to be voiced again with different voice actors, and if a quest builder decided to change the line, then the voice actors needed to repeat the procedure, and some crew would also be needed to check the results and decide to repeat it or correct it with sound editors, and so on...

Those voice files filled more than half of the DVD, so there was less space for models, textures and so on...

Those voice actors and sound editors and the like spent a lot of time on recording the voices, checking and editing sound files, and repeating the procedure when needed.

Those voice actors, sound editors and the celebrity people and the rest, needed a lot of budget to do their job, (the budget that could be used to hire more writers, quest builders, dungeon designers, artist and the like).

Those voiced dialogs made the life of the official quest designers a lot harder, and I do not want to talk about the modders, who do not have access to the official voice actors.

The end result was a shallower game that could have more wealth of story, lore, and quests, like its predecessor, and was quite hard for modders to add such content to the game.

OK, so how do we can overcome such a problem?

One solution is to revert to the old method of voiced greetings and special lines and silent bulk of the text. It's the easy way out of the problem.

Another option is developing a new specialized voice synthesis system that is focused on synthesizing in-game voices and dialogs, and performs better than the general voice synthesis software available.

So, instead of voice acting, voice actors and voiced lines, we would have voice building, voice donors, voice samples, voice sample sets, voice elements, text lines, text meanings, text http://www.nuspel.org/chart.gif, and emotion symbols (or codes), and so on...

We could still have celebrity voice actors and voiced lines, for important dialogs and key moments, but for the rest of the dialogs, we can let the system build the resulting voice output.

I will go into the detail about implementing such a system later, but for now, let's have a general look at what it can be and see what happens if we have such a system at our service:

First of all, All the text lines could be made of one or more different threads, so we can have the written text which is shown on the screen, a code thread that might contain phonetics, lip movements, emotions and other character behavior, which would help the system build the voiced line to play, and control the behavior of the speaker, voice thread which could store the voice acted line for special cases, and a meaning which is understood by NPC-AI, (I will go into more detail about that in another place).

Text: Written text shown on the screen.
Code: Phonetics, emotions, lip motions, and other behavior.
Voice: Voice acted line.
Meaning: For NPC-AI.

All those parts are optional, so if we have lines that would not to be voiced, or the default voice engine's output is enough, then there is no need for "Code" thread.

And if we have voiced lines that do not need to be shown on the screen then the "Text" part can be ignored, and if there is no need for any NPC to understand the meaning of the dialog line, then we can ignore the "Meaning" thread.

OK, I will write more about it later, but for now let's see what we will gain if we have a functional voice building system, and a related NPC-AI, but I do not want to go into more details about AI, events, and quests here, so I would be brief:

There can be an application that helps people gather voice samples and make voice sets. It can have some prebuilt voice sets for generic imperials, bretons, khajits, argonians, mers, red guards, and so on...

Bethesda can distribute that application and tell people that anybody who sends them a usable voice-set, would have his name in the credits as a voice set donor and would get the special add-on for the package when the game is gold, and if anyone can create a new distinct and usable sub-race pronunciation, he would be gifted a free copy of the final game with the add-on.

As for sub-races, we can have some variety of pronunciations in different areas of the land, so imperials of colovian highlands can have different pronunciation than imperials of the niben bay. People can try to make distinct new pronunciations, or mimic and add to a current available sub-race, with more chance of being accepted.

Those sub-race samples make a hierarchy of relations like this.

Men (Imperials, Bretons, Redguards...)
Mer (Altmers, Dunmers, Bosmers, Orsimers, Dwemers, Ayleids...)
Beast (Khajits, Argonians, Goblins...)
Daedra (Daedra Lords, Golden Saints, Seducers, Dremora, Xivilais, Scamps...)

Then each race can still have sub-races which would present the different pronunciations of different parts of the province.

And each voice-set would be related to a sub-race, or the generic race.

It works like this: You start the application, select the gender, race, and sub-race(if available), (or you might want to create a new sub-race), and then copy the voice-set that is most related to your choice to a new name, (or create a new one from scratch).

If you have already been working on a voice-set, then you open that and all the settings would be already selected.

After your current voice-set is created/loaded, you are shown a line of text, and if you like, you can listen to the current(if available) or the nearest generic voice for that line, then you can pronounce that line as you like.

The line texts are selected in a way to produce more voice elements that are needed to build the future text voices. So the engine tries to trim your voiced line, and extract the voice elements from that line, and you can listen those elements to see if they need some tweaks, like moving the beginning and end of the selected section of the voiced line for that sample by a few milliseconds to or fro.

After that you can listen to a few lines that could be produced by those samples to see if that worked, and if you found a glitch, you could try to tweak the problematic samples until you are good.

This procedure can be continued for new text lines until all the required voice elements of the voice-set are created, then you can start with voice effects like, Pain, Attack, Power Attack, and the like.

The engine might require you to repeat the procedure for "Whisper" mode and "Shout" mode as well.

If any voice element, or voice effect is not supplied with the voice-set, then the ones from its ancestor voice set are used. And sometimes we can supply a combination of some voice elements or a whole word, or a whole sentence, for the more used ones, and whenever the engine finds a place that it can use those words or sentences, then it would be selected instead of those elements. And those words or whole sentences would be given a code that could be added to or changed in a descendant voice-set.

After a while, Bethesda can gather those voice sets as they start to be delivered, and select the more usable ones from them, tweak them if needed, and add them to the available pool for each race and sub-race.

In the construction set, those voice-sets can be assigned for any NPC with pitch modulation and other effects, so each voice set can be used for more than one NPC without being exactly the same.

When in editor a designer types the text for a dialog line (or the like), he could open the behavior window of the editor to define the behavior of the character while speaks the line, like his emotions, looking down, or shaking of the head, or examining of the item in his hands, or looking at a nearby item and so on...

All of those can be selected from a list of actions, check boxes, and radio buttons and so on... But in the end all of those are compiled into some phonetics and behavior codes, and might be interlaced with lip motion data as well, but the designer can look at the codes and change them manually if he liked.

Those behavior codes might change the pitch of a section of the resulting voice a bit when needed to simulate a change of emotion, or other effects can be achieved as well.

All of these might seem a bit complicated to implement, and this is why, it is more suitable to be a middleware tool set where an independent developer could put all of its manpower on the single project to make an efficient middleware tool set for other developers to use within their projects.

Those middleware developers could make the application to collect voice-sets and make the api to use them in the dialogs.

The end result would be an efficient system, that would let us have voiced dialog lines without the need for available and on call voice actors while developing dialogs and quests, they do their job at the beginning of the projects and produce voice-sets to be used later within the dialogs and quests, and after that you could have voiced lines for anything you like, just by typing the text and defining some behaviors.

After that we would be only limited by our imagination when creating quests and dialogs, and put our resources and manpower where it would give better result.

We could develop an event manager that could create semi-random events and quests on the fly and the lines of the texts for those events and quests could be voiced by any actors, without the need to know beforehand who those speakers are. Think of the possibilities.

I have just scrapped the surface of what is needed for the effect, and what can be the outcome of such an effort, as I wanted to post this as soon as possible, and empty the burden on my mind, but I could go deeper in any section of those ideas and details, if needed.

Yeah, there is already enough threads on this thoroughly pulverized dead horse.

Some interesting ideas here. As Stahlbrand said, none of it's exactly new, but at least we have it here in one place. It's all a bit over my head, though.

One of the potential problems of this system which struck me as I read is the fact that most people don't have access to quality recording equipment to do voice acting themselves. And even if they did, I don't know that Bethesda would want to have to sort through line after line of sub-par recording to find one that might work for the game. From a development standpoint, that's all very unnecessary when you can just hire a cast of voice actors to do lines for you. But if you could get professionals to do the sounds and whatever else necessary for such a system, then I don't see why something like this couldn't work, at least in principle. I'd have to see it in action, however, before I gave it any kind of stamp of approval.

I think your middleware solution is perhaps the most ingenious part of this whole thing. If this kind of thing became standard in the game industry for huge games like the Elder Scrolls series, then perhaps more developers would be interested in generating richer, deeper stories imparted through dialogue.

I'm pretty sure Oblivion was the first RPG in history with fully-voiced dialogue, which is probably why it was so clunky.

Fallout 3 addressed the problem of linearity and we got branching quests again. However, the number of quests was so drastically reduced that it really goes to show what the sacrifice is.

Even if the technology becomes available to squeeze in lots of audio files, you still need the time and money to record each line in a studio.

The synthesizer you suggest has been recommended before. I've heard them used and they sound, to me, like a robot. So it wouldn't really add any more realism, in my humble opinion.

I'll say it again: the real problem is the fact that bethesda tries to create a game with the depth of an oldschool RPG and the visceral experience of an action game. They're trying to have their cake and eat it too. And with the way market trends work, and with the way publishers pressure developers, when a compromise needs to be made, visceral shallowness is always going to win out over depth.

Which means that, if it takes a lot of work and money to implement a new technology to give depth to fully-voiced dialogue, they're most likely going to opt for the quick solution and just keep dialogue shallow.

I still think the best solution would be to release two separate lines of games. A line of fast-paced action-RPGs and a line of slow-paced, strategic RPGs. Bethesda has proven that they can do both well and will produce better games if they start to focus on one at a time rather than try to churn out more half-assed hybrids.

So, that's my proposed solution. They should stop trying to compromise.

Oh look, another thread saying Oblivion's system was horrible! As if we haven't seen that before...
As Stahl said, this has been discussed already, stop beating a dead horse

Anytime anyone criticizes any aspect of Oblivion, they must be bitter and wrong because Oblivion was just perfection....

Anyway, back to reality....

Commenting on the voice acting doesn't really solve much in the scheme of things because:

Fully-voiced dialogue is actually just one of the more painful symptoms of a greater disease, which some call "dumbing-down" and others call "streamlining." The correct term, I'm pretty sure, is "RPG-cancer."

You see, any time a work of art starts to make money, the all-mighty business types will get their grubby paws on it and start "tweaking" it to broaden the audience and rake in even more cash. Because why have one yacht when you can have six and race them? Why fill your golden bidet with water when you can fill it with champagne?

You, sir, are a wordsmith.

And yes, RPGs are probably getting dumber to appeal to a larger crowd, but I doubt there's much to keep that from happening no matter what. The only realistic solution is a compromise. Obviously its not currently possible to do what's been proposed by Sphagne and expect it to be a valuable contribution to the game. The technology isn't there yet, but perhaps if it were seen as worthwhile we'd see investment and improvement.

Or we could find ways to tell a story that don't rely so heavily on dialogue (see http://www.escapistmagazine.com/videos/view/extra-credits/1887-Video-Games-Bad-Writing).

I only skimmed it, but I agree.

This is just plain wrong. I suppose 3D graphics were dumbing down too? What about unlimited save slots? Just because something is different doesn't mean its worse or catering to a lower demographic. Voiced dialog is meant to add immersion to the game, not to make the game more accessible or easier to play. It costs far less money to write dialog than to voice it, taking money out of the equation.

Don't like something? Fine, its your right. Blaming anything you don't like on greed or dumbing-down is however completely nonconstructive, unfair, and flame-y.

As you have noticed, I started with the idea of such a system, then in the middle of typing those ideas I decided that it was better if implemented as middleware, so that part is no longer valid. :whistling:


There are several alternatives:

• Revert to old method of less voice acting.
  
• Continue with the trend.
  
• Use synthesized voices.
  
• ...

The first method works great if we do not mind voiceless dialogs, but I do not think that is going to happen, so it is out.
The second method is the current trend wich results in shallower games, which leaves us no hope that it might get better in future.
On the other hand, the currently made synthesized voices are the result of currently available technology. And the fact that the nature of technology is that it's getting constantly better, would leave us high hope that it would get good enough to be a valid choice.

What I'm currently suggesting here is different from all the implementations that I have seen till now, and I think I know the direction that would lead us to a finally acceptable solution, and I can guess what would be a good implementation, but the details are still vague.

I do not suggest a fully synthesized voice, but one that is made of sample voice elements, that are put together to make a whole sentences, then the emotion codes start their works on the tone, pitch, magnitude, and other effects to try to emulate a real voiced sentence, then the resulting wave is sent to the output sound card.

All we have to do is to gradually fine-tune our voice elements, and the effects of the emotion codes to make an acceptable emulation of the real voiced sentences, and iteratively repeat the procedure until we are satisfied with the result.

It's like finding a solution for a complex differential equation via a numerical method, test a solution, and check the results, tweak a part and test the result again and so on...


:swear: :angel:


Why so sure about that? There is always hope in any situation, unless you are already dead.

If we finally find a voice building solution that satisfies both developers and audience, then we are ready to finally ignore the problem and start to put our time and effort were it would count, and make more decent games.


I hope so, and I just know that my solution if not a classy one, is a safe one and would finally reach to a level of quality that would become acceptable.

And I declare it here, I have thought about such matters for years now and might be able to help whoever decides to start such a project even to the smallest details, like how to make an easy UI for the application that would build the sample voice sets, but as my real life situation prevents me from joining any such group, I can only participate as a distant adviser.

^

But the full voicing mimics other best-selling games. It's called following market-trends. As in, trying to compete with the other big names by mimicking them, rather than sticking with a system that was doing well and not broken to begin with.

And how much immersion does it really add? Less available quests and less guilds to join means a less interactive world.

And you miss my point because it's immersive in a more visceral way, which isn't what RPGs are about, as I argued above. I don't get why people hate the idea of catering to everyone by trying multiple genres, instead of focusing on the one.

Correlation does not imply causation. Just because Oblivion has voice acting and fewer guilds/quests does not mean that voice acting prevents having lots of guilds/quests. Just because some popular games have voice acting does not mean that voice acting makes games popular or that Bethesda added voice acting solely to compete with other popular games.

Voice acting, when done right, is immersive because it mimics real life in a way that text can not. For example, there is no way to show emotion, accent, pacing, or pitch in speech using only written dialog. "Luke, I am your father" doesn't actually express any emotion. If spoken, it can be done dramatically to show that its a revelation, or it can be done in such a way as to say that its an obvious thing, or the emphasis can be placed on the "I" to show that he is the father as opposed to someone else, or it can imply that Luke should listen to parental authority. I'd say details like these are very important to immersion and to world building.

Furthermore, I'd argue that the number of quests or guilds has nothing to do with immersion. It adds to interactivity and provides more content, but it does not help one feel like the world is a real, living, breathing place. Daggerfall, for example, had endlessly generated misc. quests that provided endless gameplay. However, because they did not reveal anything about the world they did not do anything for immersion. On the other hand, Morrowind had a much more limited number of misc. quests, but they often (though far from always) revealed something about history, culture, religion, race interactions, etc.

a game thats rpg to its roots but has the combat system of a good action game? its a nice concept that should work. oblivion did a good job of it. An rpg doesnt have to have bad combat, just saying.

Again, I guess it depends on what you are looking for, then. Because, again, that's all immersive in a visceral way.

Whereas a world in which more is going on, not on the outside as in animations and noise, but under the surface, as in more rumors and conflicts and history and sidestories, is a whole different kind of immersive. It's more engaging on a different level. And games like this really do make the world feel alive for me. Because it's not game you just beat, it's a game you delve into. So, if there isn't much going on underneath the visual presentation, the world will seem pretty static and dead after a while. Which is what happened to me in Oblivion, when all my new characters were pretty much re-living the lives of my older ones, due to the same quests with the same outcomes over and over.

I think a compromise between the two ultimately hurts the latter more.

When I want a more visceral experience, I play an action game. When I want a more mentally engaging experience I play a strategy game or an RPG. I'm not talking down to any side because I like both.

And I want to keep playing both.

And by continually compromising one for the other... the other is going to be gone completely. Which is why, for the third time, I think they should do both, instead of combine two into one. Then it doesn't matter that we're looking for opposite things in an RPG, because both will be available.

I personally don't see voiced dialogue as a problem with Oblivion. If it wasn't for the LGNPC mods for Morrowind, the dialogue would be just as repetitive and immersion breaking as Oblivion. Add in that it was over-wordy and I actually favored Oblivion's style of "getting to the point." I say keep the voice acting, but bring in more actors. It's not as if BGS can't afford the actors or an extra DVD to hold it all (or a Blu-Ray release too!)

Voice synthesis...

I'm not sure I like the idea. I'm also not sure where I stand on the argument above. I like the fully-voiced dialogue, but I also wouldn't mind going back to text. However, That would seem like taking a step backwards for most players and reviewers, even if it makes more space, time and money for other features.

I say we leave this to our beloved developers, it's their choice.


Huh? :blink:

I just realized how good the game could be without voice acting... Well... actually I just learned that voice acting isn't effected much by modding. You can still make NPCs that are unique and with a lot of personality without a voice.

:D I made a waitress just now. And you can make such interesting dialogue trees! In fact, I might as well make a Dialogue Orchard with some characters!

I could tolerate the game without voice acting, but I'd prefer it if it was kept. It doesn't wrecking modding like people say it does... at least for me.

If it costs far less money to write dialog then voice it,then how does that take money out of the equation? i think what he was saying is they added voice acting because it was popular so they could get more costumers hence more money.



WOW this guy can speak to me i am sooo immersed. or rather in the real world, i can be evil, no wait i can be good, no wait i can do whatever the hell i want because they spent time and space on the game instead of [censored] voice acting.



I know exactly what you mean, its very hard to play again because its the same thing over and over and if voice acting really takes half the disk im sure as hell going to blame it.

And those are my thoughts on voice acting, aka screw it give me a good game over voice acting.

Voice acting does indeed affect the general game design, directly and indirectly:

Directly: Lets say that bethesda could somehow develop a quest management system that could generate(select, reshape and refit) fully interactive quest-lines and pick any nearby character fitting the role templates of the quests, and start the quests over the available resources and characters, what effect do you thing a fully and fixedly voice acted game would have on such a system, do you really think that it would be possible to make something that flexible in such a game?

You know, I know that if we loose ourselves from the shackles of voice acted dialog lines, some really innovative and ground breaking solutions would start to appear that would revolutionize the quest building concepts and character AI, and IMHO the current trend of fully voice acted dialogs acts just like shackles. It does not kill quests, but fully limits what we could do with them, or with character AI.

Indirectly: If we think of the time, resource, and funds that are going into the fully voice acted dialogs, and consider that game developers do not have unlimited time, resource and funds at hand in order to develop all the aspects of their games, it would become quite visible how it all affects the entire project.

The consumed time, resource, and funds could be spent on more writers, quest builders, dungeon designers, engine programmers and so on...

Solution: look at the next part.


For starters, let compare synthesized voices with graphics:

Computer games started without any voices, and graphics, but eventually some really awful(compared to nowadays) graphics and beeps began to show in games, those graphics were their attempt to synthesize the life forms in computers, but they did have little in common with a real world shapes.

We coped with bad graphics, but tried to improve them, so we started to develop our ways to synthesize the real-world shapes into computer graphics, and we have reached the stages that sometimes one would look at a scene with awe and say: "Is this really computer generated?"

And suppose we had decided that those graphics were too bad and hopeless, so we had better start to capture real-life pictures and somehow use them in the games, were we would be now? We would be watching some restricted and guided interactive movies in the computer. And we would not have the freedom to do the things that we do in games now, because any free form interaction with players would bring unpredictable results and the captured life shapes could not handle the results.

So captured real life actor voices is exactly like captured real life actor shapes, and if we stick to this form of game development, we would loose the unimaginable brain-blowing future advancements in game development.

Could the initial creators of "Packman" and "Space invaders" possibly imagine what we are playing now? Their pathetic efforts to create a synthesized life form on the computer screen were the initial steps to the place that we are standing now, and where would be now if the original founders of the games had told them that they did not like the results and lets start showing only real life pictures?

I say lets put all of our manpower and resources into developing good voice synthesizing engines and bear with initial results so that they can start to grow, and we would surely reach places that we can not even begin to imagine now.

That's all.

Edit: some typos.

Your part about the voice synthesizers and graphics is very inspiring.

Bethesda just needs to stop hiring big-name voice actors and throwing them off the face of the earth in the main story. Will save a lot of money. Morrowind's voice actors were fine, and I even prefered the Dark Elves' voice in that game.

Wow, I never really thought of it that way. I was always in favor of working on voice synths, but you really expressed that well.

Well, I have a lot of great ideas about innovative and new styles of quest design and NPC-AI, and more, but however I tried to implement those idea (in my mind), i crashed full throttle into the concrete wall that we call voice acted dialogs.

This darned new trend feels like a slime pool that I am sinking into, and I am can not express my feelings about that, and when I saw that people blindly veto any other idea and do not see the true potential of the alternatives, I got mad and decided to try to show them the true nature of the beast.

I don't know what happened after that, but somehow in that heated moment, the graphics example offered itself to me, and you know the rest. :blush:

I know that the tone of my previous post looks a bit hot, but that was my feelings at that moment, because I've got a lot of ideas for almost every aspect of game design as I have thought about them for years now, but this urge to use voice actors for every line of the dialogs has blocked the progression of a lot of my ideas regarding quest design and AI, hence this thread.

I personally would like to see voice synths in other games first. So that BGS is not experimenting. I just do not want TES V to be worse than it could be.