Hi.

This is Aaron.

SAPI is fairly easy to break. Over the years I've run into problems with it a few times, and I've heard discussions of other people having similar horrors with it. Some voices will leave parts of themselves behind when you uninstall them, or perhaps they remove too much data, and now SAPi either doesn't work at all, or the voice is still there but doesn't work. My Yoga2pro laptop has several voices that don't work, for instance, even today.

This is when a customer contacts you to help them fix their SAPi since your game can't speak through it, even though the game itself is not the problem.

Additionally, many voices have an issue that causes a long delay after something is spoken before the voice reports that it's not speaking anymore. this results in long pauses in between strung together phrases. Microsoft's own SAPI5 version of the One Core voices like David are among the worst.

On my system, my game is silent for just about a full second and sometimes even longer before the game is able to continue, if it is waiting for SAPi to finish. I wind up just using the SAPI5 version of ESpeak on all of my computers because the rest are just terribly slow. I like Eloquence, but I haven't decided to buy the SAPI5 version of Eloquence yet.

Recently, just before releasing EKS, I was troubleshooting a problem where my game crashed after Windows 10 updated to version 1903. When SAPI would try to speak, game crash every time. I have code that tries to check if a voice is going to work, and if not, it spells the message using the recorded speech instead. This wasn't working. Come to find out, I forgot to account for the internal list of voices not containing as many voices as the number the currently saved voice was. This is called an index out of bound. I call the fact that I missed this crappy coding on my part, and the fix was relatively easy.

However, I got the crash to stop, and then I was experiencing a several second pause every time SAPi would be supposed to say something, but not actually hear anything but silence. Even in settings in the game, I could arrow across the voices and it would never say anything.

I wound up setting up alerts to show me when it was trying to say things just what it was saying. Then I backtracked my SAPi code and alerted the heck out of it too trying to find the issue. I had at one point the knowledge that for some reason, the voice was reporting at speaking immediately, even though it wasn't speaking. The fact that it reported as speaking caused the code then to not break out and revert to spelling. So how in the heck was this the case? Since when did SAPi report speaking when it really wasn't? I banged on this for about an hour and a half changing a lot of code in the final release version of EKS, something I really didn't want to mess with.

Come to find out, it sort of was not a bug in SAPI after all, but it kind of was as well. I have the VB Cable driver installed on this computer. Before the Win10 update, everything was fine. VB Cable lets you put the audio output to its own virtual sound card, then you can pick up what went into it on the other end by pointing the input/microphone selection of what ever apps you want to it. You can listen to VB Cable through the default sound output device using the native windows functionality for that, so you can continue to hear what ever it was even though it was not being piped to the regular sound card directly.

What I had done was used the advanced app settings in the Win10 settings app under sounds to force Eks to output through the VB cable, even though EKS doesn't support that capability on its own. I then used Team Talk to listen to VB cable on the microphone side and was able to stream something I had recently changed over Team Talk to show it off to a close friend to see what they thought. It worked great, they heard everything just fine.

After that Win10 update, somehow Eks was still outputting through the VB Cable, even though I know I changed it back to the default output when I had finished. Plus, in the past, any time you close out a program that has been forced to a certain output like that from the sounds section in pc settings, it reverts back to default anyway, a fact I found somewhat annoying when I closed the game out and opened it again to start fresh with a new game and such.

But, for some reason, SAPi  couldn't be heard, while the recorded speech was working perfectly. Beats the heck out of me, but I was pretty annoyed about that, to say the least.

So saying, it has always been my policy to keep these games as self contained as I can manage. Besides the fact that I've always liked a game that used recorded speech myself, I just wanted to be able to do that too. We've both mentioned this before, but fortunately or perhaps unfortunately, we're producing games we would want to play here, and for the most part, their interfaces and way of interacting with the human is what we personally prefer.

The main reason I added SAPi support in the first place was so we could say things that would otherwise have been spelled out. In fact, originally, spelling is exactly how strings we didn't have speech files for were handled. For instance, the Guidedog activation item in the main menu. When that would tell you who was logged in, it would spell your user name.

Now that we have it, we rely on it to keep the game Talking when I add new strings that haven't yet been recorded, and, as you all are certainly aware, for certain things that recording speech files for just isn't practical, like people's user names and long descriptions for 40 different planets.

I wish I knew how to use ESpeak's dll from the games. I would package that and either use that exclusively, or offer that choice. It sounds pretty terrible to me, but it even works reasonably well when small strings are split up to facilitate playing many different speech files, but none of them are made yet, or only some of them. It bounces between ESpeak and the recorded speech very fast and smoothly even when using the SAPi version of it, obviously it's a little odd hearing both voices help each other out like an old couple, but hey. Oi, don't call me old!

Regards:
Aaron

Hi.

Aaron again.

I've got code that can do this, but it doesn't work well with delaying. This is because the screen readers don't report through their various access methods this relies on when they finish speaking. There's a similar setup to how GMA's games handle it, where it goes by the number of words and you calibrate it by letting it say something and telling it when it stopped, then that creates a bit of an average that sort of works.

There's also the issue that if you want to interrupt the speech you can't do so with control unless the screen reader supports speakInterrupt with a blank string. I think all of them do. Note that if you don't need the keyboard hooking function in the game to be on then control will work anyway.

The main issue is that delay estimator thing isn't always accurate. You'll have the game continuing on either too soon or too late in situations where it was supposed to wait until speech finished.

A future game we have planned will need a lot of long winded speech output that won't lend itself to recorded speech, so we'll likely explore this and hopefully other options while we're at it at that point.

Another reason not to do this though is that players can slide in a translation for the game, or their own preferred speech folder even if the same language, then all the recorded speech might be in say, Spanish even though the game is really English. It's never perfect, but it's sometimes good enough to get by, where translating all strings that would go to the screen reader or SAPI or which ever is not a simple proposition for us or who ever wants to handle the translation.

We are working on generating a text file with the file names and the strings each file has which may help translators or other players who just want to mod the game by providing their own pack instead. The bulk of the work here will be on whom ever wants to make the mod, but it should be a possibility. I have to code a new entry in settings though to allow for this, because if you put modified speech files in the speech folder the next time the game updates it may accidentally delete them if they're in the default speech folder. So this will involve pointing the game at a completely separate folder.

Well just to tell you guys I have the game now as of 2 days ago.
Not played it yet. but still

Hi.

On our end the speech is not unclear. There are quite a few people playing it without issue. Two things come to mind.

1. The speech is pretty loud. Try turning it down in settings or just by pressing f5 after the game is running. Even if it doesn't sound loud, it might be possible that it is distorting the compression on your sound card. I actually have a sound card that does that, and I have to turn affects off in windows itself. I have issues with everything sounding silly though when this is on, and I quite honestly wonder why pc manufacturers even bother. With this Lenovo if I wear headphones and use my screen reader or play any game for very long at all my ears start ringing and I get dizzy and my head hurts, until I turn Dolby off.

I also have a little tablet AsusTeK T100TA-Am that distorts if the speech is at anything louder than say, -10 on speakers. Works fine with headphones though. The speakers distort easily on that tablet anyway so this is not a surprise.

2. The speech files are all jumbled together. It may take practice to understand the speech this way. The default separation is what we think is a good compromise of intelligibility versus efficiency, but it may help to try setting speech file queue before end delay all the way down to 0 to bring this more in line with how most games using recorded speech work. This item is also in settings.

3. Can't promise anything yet because this is in the early stages and we may hit an unforeseen road block, but we might possibly be making it easy to switch to a different speech pack.

Aaron

So I'm going to echo the sentiment of others.
I can not get in to these games due to the recorded speech.
I don't want to be rude because I know what it's like to put a lot of time and effort in to a project, but the speech detracts immensely from the overall experience.
if you are dead set on using human speech, I would be willing to sit down and record professional narrations for your games if you guys want. Get in touch if you'd like to work out something. I'm fair and will work for a decent price.

No moderation or anything, but there is no need for that sort of comment. The author of the game is profoundly deaf so I think that has a lot to do with his voice being the way it is. Either way. Why not be constructive instead?

Liam,

Thanks for the offer, and for the comment in post 88.

Re: voice discussion.

I can only assume that people are either choosing to ignore, or haven't read, the thread just on this page of the discussion.  Aaron has mentioned twice in this page that he's working on a solution, that while it might not meet with any of the critics' approval, is a solution that they can take advantage of.

As to whether our intransigence is hurting the game, or whether people's stated discontent should be a reason to change it, the final decision is ours.  We listen, and we have talked about this issue with each release.  Given that a number of people on this forum have complained about the name of Guidedog as being too blind specific, i'm surprised that so many people think that someone who has "a speech impediment" or is "profoundly handicapped" should not use their voice or perform useful work in society.

You can't have it both ways.  Either you want things a certain way, or you should encourage entrepreneurship within the disabled community, no matter whether it's a blind person, deaf, mentally challenged, or whatever you wish to pick.

That said, as I pointed out earlier, Aaron's already made two announcements of a change in this policy, so it seems even worse to continue to beat on a dead issue.

Frastlin,

What would you like to see as an alternative to the current way ships are discuss on planets?  If you fleet them up, you can see them in a menu one at a time or on scrollers.  If you choose not to, the way we did it seems to us the most elegant and least spammy way of handling it, but we're open to other suggestions.

Take care,

Jeremy

@90 It would still have been an insensitive remark even not knowing that. Try attacking the problem and not the person. I think we're all too quick to forget that there are real people behind the keyboard at times.

@89 your ideals do not line up with reality, unfortunately. You claim we quote, want it both ways, end quote; that is simply not the case. Seeing how the main method of feedback in this game is recorded human speech, it is not unreasonable to assert that the voice in question should be clear, easy to understand and at least somewhat pleasant. You seem to be forgetting the fact that we do have a fair few international community members who do understand English; however, with it being their second or even third language, a clear voice would really help them out.

Further, you seem to be under the impression that people are complaining about a perfectly pleasant voice, and that's just not the case. I don't want to harp on about it because of the circumstances, it's not exactly a nice thing to do, but perhaps the decision to do this in house with someone was a bit ill-conceived. As to your comments on whether a person with a speech impediment should use their voice professionally, i would say most likely not. I don't use my eyes professionally, and most totals wouldn't either. VI people might, depending on the severity of their condition, but even then, it's not affecting anyone else, and they're able to do what needs doing.

I understand how this is a touchy topic. I don't exactly enjoy pointing out someone's disability or making that the focal point of the discussion. In this case though, I will say that the voice in question is not suitable for professional work. I don't have any malice or ill will towards the owner of that voice, and indeed wish them well and the company itself well in its pursuit of making more games.

Yes, I agree that the decision is yours, but pointing out that intransigence and close-mindedness will hurt the game and the company overall is simply criticism. Take it or leave it as you will. The best games strive for a balance between the developers' vision and community feedback ((suggestions, criticisms, etc.).

Ironcross,

I would agree with you if there hadn't been numerous points made earlier that we were looking at changing this setting.

As to your opinion of the voice, that's fine, but the reason we did so are twofold:

1. Of the two of us, at the time, Aaron had the better microphone and recording setup to handle the voice duties. 

2.  Paying for professional voice acting is frankly unrealistic.  It would either be prohibitive and basically cut any profit we make to 0, or it would cause us to price the games far above where we wanted.

We could have others do so on a volunteer basis, but that then puts our production schedule at the mercy of numerous others whom we have to wait on, possibly not use depending on recording quality, and then coordinate.

For the first few games those considerations decided the matter.

With this game, the fact is like it or not, the voice is part of what sets our games apart from others. I accept your point about international players, but if the game is in English, no matter the voice, it's possible to run into the issues no matter what device is delivering the text. 

Working with the community has been something we have done extensively in the past, so I find this line of argument fececious at best.  We released updates to all of our games except Interceptor within a month or so of their release to address issues brought up by the community--other than the sself-voicing.  The only reason we have not offered an update to Interceptor yet, is that we are planning to do a complete overhaul of the game to make it a better experience, not just apply some patches to a game that we felt, in retrospect,  was not as good as it could be.

We have addressed this voicing issue numerous times, just in this thread, not to mention in previous threads.  We have noted three times now before this that we're exploring use of another synthesizer.  We're not going to address this further.  At this point, continuing to argue about the voice is frankly, an attack.  If someone tells you they are changing something, then continuing to criticize the thing being changed is not constructive nor criticism. 

Take care,

Jeremy

Well now you can do that as well because of 87's offer

Precisely. You have a small, niche market that will try anything, and if it is free of charge, they'll play the heck out of it even if it is mediocre at best. I think your game probably is beyond mediocre, and probably pretty good from what I've heard of it. However, this small, niche market also has a bit of an issue, which is finances. When you sell to a group of people with limited cash to spend on frivolities and luxury items like games, they are going to want to make smart purchases. To go along with that, they will want to feel as though the thing they're buying is worth it. When you have such an integral part of the game as the voice be less than optimal, it really does detract from the experience.

Your attitudes aren't helping matters either. And while you say change is at hand, I didn't see much commitment to that end, just that things would be looked into. However, maybe I missed something. But hey, maybe your sales numbers will speak to you in a way that you will be receptive to, as you have been minimally receptive toward us.

That should help those who don't like the voice, and that sounds like it is what you all are asking for. It also allows for the game to be translated into different languages.

Still, I hate ESpeak, but I would play a game that used it because of the game, just like I do with microsoft Anna. The problem here is that while this happens on mainstream forums, it is not the main source of advertising for the game. Not to mention the much wider audience. When four or five people there don't like something it probably gets lost in the noise, whereas here, that is a decent percentage and can dominate a discussion so much that the advertising of the game is overshadowed.

yes, but since it's a smaller audience, shouldn't they take those four or five people more seriously because it is a larger percentage?

Keep in mind too, the developers are blind, and might be in the same boat with limited resources to buy fancy things like voice actors or sound libraries. It's not like the amount of sales they are going to make, even if it becomes a top game is enough to live on. Manamon sold what, 300ish copies? That's $12,000.00 at $40 per copy. If EKS did the same, it would net them $6,000 to split between two of them, and that's not counting after credit card fees. I think the fact they are working on another way, even if the onus is then on the user to make recordings, is a fair compromise.

Still too, not liking someone's voice for voiceover work, is like saying blind people can't do web development because they can't see fonts or work with image placements. If you actually listened to my demo all the way through that I posted in post 74, or tried the game, you would get used to it. Sighted people don't understand or like our synthesizers either, but we are used to them, and while we might prefer eloquence over sapi, or dektalk over eloquence, we muddle through and eventually are okay with it.

I am very disappointed that there have been fewer discussions of the game itself rather than the cosmetics because it really is a pretty awesome game, and the update should incorporate some of the suggestions brought up by people here once the bugs get worked out.