have you tried sh4speech? if not, why not?

[minsc_tdp]

All too often it's easy to work on features that are interesting problems to solve in the code rather than what people actually want. So this poll will help me prioritize new features.

Here's my standard (short) responses to each issue:

1. Hard/time consuming setup - Setup has been overhauled! Now all you need is the MS Speech SDK (see readme), and no other install is required. Just unzip and run sh4speech to start!
2. Setup problems - Ask anyone who has had trouble, I spend a lot of time helping people personally until it works.
3. Don't have a mic/headset - they're cheap
4. Mic/headset is a pain to setup physically - not much I can say here
5. Not looking forward to training - A single 10 minute training session is all that is required. You just read the text on screen. And you can probably even get away with skipping this entirely, I haven't tried but I wouldn't be surprised due to the very small command list.
6. Skeptical / probably bad recognition - After one training session, it's nearly perfect. Nothing like Enigma.
7. Doesn't support my resolution - Coming soon, top priority issue. UPDATE: Fixed in 1.7b
8. No full torp / TDC control - Coming soon, #2 issue after resolution fixes
9. Too embarassing - try doing it while wearing a captain's hat :)
10. Just didn't like it - nothing I can say, explain

[irish1958]

No, because I am stupid and nerdly challanged.

[panthercules]

LOL - as a dedicated SH4Speech user, I was plowing through all the possible answers looking for the "I tried it and I love it" response to the poll, then I took another look at the beginning of the poll and realized it was just for folks who aren't using it. I hope you get some useful responses that help you address any concerns other folks might have so they'll be more likely to use it - as for me, once I got past some initial installation problems (which you were great about walking me through), I've had a real blast playing around with this thing. You've included a vast array of potential commands, but also made it easy to customize, and I've had a blast coming up with a bunch of custom commands for my own use (even watched Destination Tokyo again last night to come up with some more ways to give orders. :D )

I used Shoot with SH3 and really liked it, but this (especially with the latest addition of the course and depth setting commands) seems to have way more potential and I'm really looking forward to using this in game a lot (whenever my spousal unit isn't around to complain about me yelling orders to my crew :lol: ).

I just hope the long list of generally "negative" poll response choices doesn't scare people off from trying this - they'll be missing out on some great fun if it does. Great work :up:

[JScones]

I'm not sure if I am in your "target" group with this poll, but the reason I haven't tried it is simply because I've never enjoyed playing games with "voice"; even though I guess it'd be less frustrating from time-to-time than fudging around looking for, or remembering, the key that does <whatever needs to be done within 0.5 of a second>.

I guess it goes back to playing games late at night and the anger it would create within the household, ie "will you shutup!", LOL!

So in other words, my non-use has nothing to do with your creation, just a "childhood conditioning" (which oddly carried through to my wife on those late night TeamSpeak sessions I used to have a year or so ago :hmm:).

But you did ask why not, so there's my reason. :)

[AVGWarhawk]

I always wondered how well it worked.:hmm: I did not want to spend the time setting it up only to find it did not work well.

[ReallyDedPoet]

It sounds good :yep:, however I am not there yet as far as wanting to use it in-game.


RDP

[C DuDe]

I'm lazy so I haven't installed it...

Don't want to spend a lot of time getting it to work by having to adjust some key bind’s to make it work.. No offence!

[dean_acheson]

becuase my mike is attached to headphones, and with vista and creative soundcards, you can only get sound out of one speaker, likely for the same reason that you can only get 2.1 sound out of the damned sound card anyway.

thanks microsoft.

thanks creative.

[Rockin Robbins]

I'm following the progress closely, as I use Shoot with SH3 and love it. But I've been busy with a campaign and playing with different attack strategies. I will definitely give it a chance, but it's not front burner right now.

I followed from your first post on the subject, wondering why you wanted to reinvent the wheel, since Shoot works so well. As events have developed it's clear that you're doing things Shoot can only dream of, and this is truly something worthwhile.

The greatest thing about voice command is that keyboard layouts no longer matter. What's shift-n do in Trigger Maru? Who cares! I can just say "track nearest merchant" and be done with it. And the crewman says, "Yes, sir!" Now if I can only train my kids to do that...:stare:

[Hawk_345]

It looks complicated and time consuming to set up and i really dont have that good of a mic so im a little skeptical of installing all this stuff just to get it to maybe work. now maybe if all that stuff u listed that needs to be installed came with sh4speach, i might try it seeing i wont download and install the wrong thing. but only if i can get it off my system easily enough if it dont work.

[CaptainCox]

I tried this back when the mod/app came out for SH3...never got it to work properly :( .But the biggest reason is probably my wife...talking to my PC would just make her case so much more believable, in fact 100% convincing. As to what her case is?...well go figure :p

I might give it a go, but has to be on a weekend when shes out or something ;)

[mountainmanUK]

I've tried using Speech Recognition programmes before (Shoot and VAC) and they both seemed OK. Most seem to be set up for users with US American keyboards though, and I've always messed things up while reconfiguring for a UK keyboard.

This programme actually works VERY well on my computer, as far as I've tried so far, apart from the same old Keyboard configuration thing!!

If a UK-friendly version could be done, I'd be over the moon!
(Yes I know that I just need to reassign keys, but I'd much rather just fire it up and GO!)

[Digital_Trucker]

But the biggest reason is probably my wife...talking to my PC would just make her case so much more believable, in fact 100% convincing. As to what her case is?...well go figure :p

I might give it a go, but has to be on a weekend when shes out or something ;)

After a while she'll get used to it. I have to knock on my wife's head now with my knuckles to let her know I'm talking to her:lol: And she's stopped dedicating "They're coming to take me away, HaHa" to me on the local radio station.

[CaptainCox]

:rotfl: :rotfl: :rotfl:
GOD ONE! I should try that ;)

[PepsiCan]

I haven't tried for two reasons:

1) I don't have a need for it. Keyboard works fine.
2) Problems with my SH4 install at the moment, so I need to reinstall first.

[minsc_tdp]

Good responses. Some I can't do much about. The only one that really bothers me though is concern over the time and complexity of setup.

The setup has gotten significantly better since the early versions. Initially, I was installing 7 components plus sh4speech itself, now there's only 4 and it's a lot more reliable. There are links in the readme for the Perl/Python components, and the rest is handled by the install.bat.

If I showed a video of zilch to up and running in less than 3 minutes, on a fresh system with nothing but SH4 installed, would you try it? I'm going to strip my system bare tonight and film this for y'all.

As far as the recognition quality, it is magnificent on my system. Generally if the mic gain is at a "just right" level, for this simple app, it will recognize just fine. Most speech recognition sucks because it's trying to match against the entire english dictionary instead of a reduced command list. For example, if I'm trying to figure out if you just said "torpedo" or "command room", and those are the ONLY two things you might have said, I could probably understand you even if you were gagged and wearing full scuba gear. That's the idea behind better recognition for a short list of commands versus the entire english dictionary and complex grammar/sentence structures.

BTW, whoever said "Tried, but still having setup problem" please PM me and I'll help. I wasn't aware there was anyone left that was still having trouble getting it running.

[Hawk_345]

well ok i may try it, i just hope it works with my poor mic.

[minsc_tdp]

BTW, DigitalTrucker, I've been wanting to say this for a while but I keep holding back... I can no longer...

the sub in your sig is....FABULOUSSSSsssss *snap* *snap* lol

[Ducimus]

I haven't because id feel really stupid sitting there talking to myself. I'd be teased by my wife, and my neighbors would look at me funny if i left the window open for some fresh air.

[skwasjer]

minsc_tdp, why don't you provide a proper installer package, like a setup.exe or msi-file (which I prefer). I strongly recommend a single and easy "next > next > finish" installer to take away any concerns of users regarding setup. You may think 3 minutes isn't much, but it's a show stopper for many people.

If you don't have the tools or experience, I could offer some help, although I'm very tied up with my own tool, work and family... There are many good free tools available for you to try/use.

As far as using sh4speech, I personally don't like python, perl, java and the like, and the command line window approach. It wasn't on the poll choice list though ;)
My vote doesn't really matter though. I don't have any time anyway to play, haven't played in months, just because of my own tool. Actually only the first month after release, when my collectors edition was delivered :)

Nevertheless, I think the tool can be (is) very useful and fun once you've got the remaining problems worked out.

PS: I think many people on the MOD forum are too occupied with modding/testing their own projects so I think your main target audience is more the other forums where the regular users/players hang out.

[Digital_Trucker]

BTW, DigitalTrucker, I've been wanting to say this for a while but I keep holding back... I can no longer...

the sub in your sig is....FABULOUSSSSsssss *snap* *snap* lol

Don't know how I missed this earlier, but thanks. One of my hobbies is 3D and I just love makin' stuff!

[DirtyHarry3033]

My reason: Other

I've been using Shoot 1.6.4 for all my VR needs for years now, including SH3 (I'm using a Shoot SH3 profile that I spent a lot of time modifying for SH4, and it's working perfect!)

So I've already got something that does just what I want, and I really hate the thought of creating a whole new profile for sh4speech, and having to install Perl, Python and whatever else to run it, along with the extra load they would put on my system.

That said, if I did NOT have a VR prog installed and working the way I want, I wouldn't hesitate to try it!

My only issue would be the mouse commands, there is no way for your prog to "see" the screen and know how the compass, telegraph and depth gauge are currently configured. So how does it know where/what to click, to do what you want? Sounds like problems waiting to happen to me. Sounds like you'd have to reconfigure it every time you changed depth gauge from shallow to deep or vice-versa, or changed from compass to rudder, or from std telegraph to kts.

Just my thoughts...

DH

[panthercules]

My only issue would be the mouse commands, there is no way for your prog to "see" the screen and know how the compass, telegraph and depth gauge are currently configured. So how does it know where/what to click, to do what you want? Sounds like problems waiting to happen to me. Sounds like you'd have to reconfigure it every time you changed depth gauge from shallow to deep or vice-versa, or changed from compass to rudder, or from std telegraph to kts.

Just my thoughts...

DH

If I understand what he's done correctly (haven't had time myself to test it in game but it seems to work well in test mode so far), he seems to have this licked at least for speed, depth and course/rudder dials. Given that you can make all the changes you desire to speed, depth and course by voice, you'll never need to be clicking/changing these dials during the game, so he's got it set up where you just have to issue 2 or 3 variable-setting voice commands once at the very beginning when you load the game (to reset the dials and tubes to their beginning state), and then you're done.

[minsc_tdp]

My only issue would be the mouse commands, there is no way (I beg to differ!) for your prog to "see" the screen and know how the compass, telegraph and depth gauge are currently configured. So how does it know where/what to click, to do what you want? Sounds like problems waiting to happen to me. Sounds like you'd have to reconfigure it every time you changed depth gauge from shallow to deep or vice-versa, or changed from compass to rudder, or from std telegraph to kts.
Well the short answer is, as long as you use voice for controlling these dials every time and avoid the mouse, it really does handle the various dial modes and works just fine.

The long answer is:

1. The location of each of these buttons and dials had to be painstakenly entered for 1024x768 mode. From those, depending on your actual resolution, I can determine the location as SH4 uses a fairly predictable and simple algorithm for arranging the controls. For example, if a dial is at 1000x700 pixels in 1024x768 mode, and you're running 1920x1080i, then the X is at 1000+(1920-1024)=1896. There's four slight variations of this algorithm depending on which quadrant of the screen the control is located in, but it's not really that hard. I'm not completely finished testing this enhancement for all resolutions, but 1024x768 works perfectly, and I'm so confident it will work that I've deleted the 1920x1080i data that I had painstakenly added in.

2. Yes, the bottom three dials have two modes each. Depending on what you say, it has to be sure it is in the right mode. That's where I say you need to commit to using voice for these always, because that allows me to keep track of the status of those dials and click the mode switch only when necessary. At the start of a game, you just make sure the dials on on their defaults (Telegraph/Compass/Shallow) and say "reset dials" and it works perfectly from there.

3. In fact, if you want 162 degrees heading, the voice thing will probably get closer to clicking the exact spot than you can with the mouse, since it uses basic trig/geometry to figure out which pixel to click based on the angle you say.

I know you have Shoot working, but if you decide you want dial control, then you'll have to consider this program. Creating the "profile" is virtually a non-issue - you didn't have to create one at all for Enigma, nor for this - you can just learn the voice commands it's pre-set with (and comes with a handy printable reference card (http://knepfler.com/sh4speech/sh4speech_refcard.htm)) and use those (hell I don't think you could even customize them in Enigma.) Making a new voice command "alias" for a command is as easy as opening a text file, copying and pasting a line, and typing in whatever voice command you want to use.

[minsc_tdp]

Thanks for the feedback everyone. I think the resolution compatibility and installer are the hot-button issues.

I'm thinking of compiling these scripts into EXEs which would completely eliminate the need for installing ActivePerl/ActivePython. It would make the download significantly larger but sounds like it might be worth doing. Maybe I'll keep the "small/advanced version" like it is now and also provide the "big/easy to install" version with EXEs as a dual-release. I already have a working perl-to-exe compiler but I've never done it with Python. I gave it a quick try and the compiled perl is fine and ready to go, but the compiled python exe doesn't work due to using win32com. Any python wizards out there want to help me compile hear.py, I'd much appreciate.

[minsc_tdp]

along with the extra load they would put on my system.

I'd like to also address this. Perl only uses about 67 MB of disk space, Python uses about 40 MB. For comparison, SH4 weighs in at over 5000 MB. They have no resident programs running after install and are only invoked when you run sh4speech or another perl/python script. sh4speech itself is 0.15 MB of disk space unzipped.

As far as CPU/Memory load while running, Perl uses ~5MB and Python around ~9 MB of RAM. In Task Manager, when I run sh4speech, memory usage only increases by 26 MB (the MS speech engine is probably using the other 12 MB). SH4 probably uses over 1000 MB.

[minsc_tdp]

Well, 1.7 BETA is now here, and you now have support for all resolutions, and full control over all the torpedo settings, PK, and TDC (aob and target speed). See this post (http://www.subsim.com/radioroom/showthread.php?t=119430&page=7)if you want to try it out before it is released for real. I only put the BETA tag on it since the changes were so significant, but my testing so far shows it's fine, I doubt anything will change between this and 1.7 Final.

At this point, the installation process and localization are about the only things that need any work. See my sig..

[C DuDe]

As far as CPU/Memory load while running, Perl uses ~5MB and Python around ~9 MB of RAM. In Task Manager, when I run sh4speech, memory usage only increases by 26 MB (the MS speech engine is probably using the other 12 MB). SH4 probably uses over 1000 MB.

Myn runs at 820Mb.. checked it last night (forgot to suspend NAV)

[minsc_tdp]

Myn runs at 820Mb.. checked it last night (forgot to suspend NAV)

I assume you mean SH4 not sh4speech right? Cuz if I'm using that much, something is horribly wrong

[minsc_tdp]

I just posted some videos at:

(file:///C:/sh4speech/voice_commands.csv)http://knepfler.com/sh4speech/videos

The videos show the entire installation process, as well as a demo video at the end of me sinking a sitting duck almost entirely by voice. Control over these functions is brand spanking new and I haven't practiced any of the commands.

So, in that vid, you might notice one or two missed recognitions - angle on bow is especially tricky for some reason, but it's the only one I've seen behave like that. "toggle PK" was just my throat needing to be cleared. Also, when using calculate speed, it didn't auto-send to TDC, but that's already fixed and posted.

UPDATE: Having lowered my Pronunciation Sensitivity a little, I get much better recognition on the Angle on Bow commands, as expected. You have to tune that once in a while, as the ambient conditions change, as well as microphone distance and gain (mine is on a stand on the table) so you want to use the test mode once in a while and make sure it's recognizing this command and a few others well, as well as not doing a false match when you say something nonsensical into the microphone.

UPDATE: Updated 1.7b (no version change) with a minor fix that speeds up the dial twisting action by using larger steps, instead of hitting every degree on the dial between 0 and the requested setting, which could be kind of slow (watch the AOB turn in my video to see what I mean.)

I'm also considering changing the voice commands from "torpedo gyro left 5 degrees" and "increase angle on bow 90 degrees" to simpler phrases, such as "gyro right 5 degrees" or "spread right 5 degrees", and "increase target angle 90 degrees" (since "bow" seems hard for it to recognize for some strange reason.)

I'll probably release 1.7 Final tonight so I hope to get some feedback on all these new changes before then!

[minsc_tdp]

2.0b is done!

No more install of ActivePerl or ActivePython. Compiled EXEs! No install whatsoever, other than the one prerequisite of MS Speech SDK (link in readme.)

Like 1.7, support for all resolutions.

Also with 2.0b, minor fixes to dial spin speed.

[chopped50ford]

I always wondered how well it worked.:hmm: I did not want to spend the time setting it up only to find it did not work well.

My reason exactly...Im sure I could make it work, just dont want to cross the "path of issues." :)

[minsc_tdp]

I've only gotten a little feedback from a few people who have tried it, but everyone so far seems to have the same experience as I do - excellent recognition right away. Now that the install is a lot easier and there's a reference card, you should give it a shot, I think you'll be surprised.

[WilhelmSchulz.]

Never heard of it? :-?

[minsc_tdp]

My reason exactly...Im sure I could make it work, just dont want to cross the "path of issues." :)
I always try to think of a response that will encourage people to try it out, but this one was kind of hard. I assume you're mostly talking about the recognition quality, as well as perhaps some of the more exotic tube and dial commands.

I would say this: How do you know there will be any issues? I know everything has issues, but there's been a lot of testing and revisions to fix so many things, it's very polished right now. Nobody has ever reported anything wrong with the dial commands, they work perfectly at all resolutions. Trigger Maru is now 100% supported. There are virtually zero outstanding bugs right now, and there have been virtually no bugs that weren't fixed within a day of being reported.

As far as the recognition quality, several people have tried it without doing any tuning of the speech stuff - just install and run - and had very good experiences, made only slightly better later on by dialing in the speech control panel and minor tweaks to the voice commands.

I think AVGWarhawk said it best, "it really feels like you are skippering the boat!" When I saw that, I thought to myself, "hey, another person finally gets it." It's definitely a weird feeling to speak the commands and hear the instantaneous "yes sir!". It truly changes the game. It may not be for everyone, but you won't know until you try!

[kikn79]

minsc_tdp, what type of headphone/microphone do you recommend?

Thanks,
Chuck

[minsc_tdp]

minsc_tdp, what type of headphone/microphone do you recommend?
I've tried it with a crappy laptop built-in mic while traveling and it worked amazingly well. Most cheap desktop or headset mics are better than this so just about anything will work.

Keeping that in mind, the following is a more nitpicky and high-end mic discussion.

Here's the pros and cons of desktop vs headset in my opinion:

Desktop Mics:
** Most are very, very cheap and crappy. Most of the innovation being done in the under $50 range is in headsets, not a lot of desktop mic options, and most I've seen are all the same - low end junk. Perfectly fine for starting out though. Can be had for as little as $1.
** But, you can buy high-end Hypercardio mics at Guitar Center for $85+
** Might think you ordered a voice command when it's actually something that blasted out of your 5.1 speakers, but my volume is blasting and it doesn't have this problem
** Have to be conscious of the distance from your mouth, as that tends to vary, though this is less of an issue with high-end mics
** No comfort issues - set it up and play all day long

Headsets:
** No chance of game sounds or people in the room interfering with voice commands
** Consistent distance from mouth means less tuning/tweaking and better recognition than a $10 desktop mic
** Uncomfortable to wear for too long
** Few options in the high-end quality range (> $100. In contrast you can spend almost unlimited amounts of cash on a real mic. I saw several at Guitar Center over $5000.)
** Most don't have the air dampeners (the spongey thing) or a cheap one that barely helps.

I haven't bought a headset/mic combo in forever so I don't know much about them. The last one I had was a $50 Plantronics and it sucked, which surprised and annoyed me. The mic had no air dampening so every whoosh was picked up like it was in a wind tunnel, and I had to move it really far from my mouth, which then sounded like crap.

So I decided to go high end and stop screwing around, since I also had other recording needs.

A good choice in the $100 range are the Shure SM microphones. These are serious vocal mics and musicians will always recommend them.
Shure SM (http://www.guitarcenter.com/shop/product/buy_shure_sm58_mic?full_sku=100038284)

I have a very similar Audix brand, which saved me about $10-15 bucks.
Audix OM-2 (http://www.guitarcenter.com/shop/product/buy_audix_om2_microphone?full_sku=100371957)

When I first plugged that thing in and recorded myself and played it back, I was blown away. Awesome quality, and that helps recognition.

Note: Most people who shop at guitar center are cheap bastards, and Guitar Center has survived by learning how to negotiate and accept lower offers. Which means damn near everything is overpriced. They can be bartered down on nearly everything. Act like a stoned, jobless starving musician.

Both of those mics are extremely rugged. Don't go too crazy on these high-end mics as many expensive ones more fragile, designed to be mounted permanently in a studio and left alone. The above two mics are extremely rugged, karaoke-type mics that people abuse all day long. I would warn that even these require some level tuning BECAUSE they're so good. Fact is all mics will require level and gain tuning to get the best quality. A high quality sound card also helps a little, primarily one with a real gain adjustment knob, like the Creative X-Fi Platinum.

Note that any high-end mic from Guitar Center will use an XLR cable (I've never even seen those before) and require a special adapter/transformer to work with standard stereo inputs that most sound cards use.

All of this aside, you can do very, very well with a cheap mic by getting the gain and distance from your mouth optimized, and do the training functions of the speech control panel! That makes even more of a difference than a $100 mic upgrade. That's just the icing on the cake (I'm just a sucker for expensive electronics)

UPDATE: I should add an important fter doing some experiments with my expensive mic using an XLR to Stereo connection. I switched to the mini stereo connector so I could plug it into the back where people wouldn't step on it. Problem is, that connector doesn't have a gain control to heavily amplify the signal. Only on the front panel of my X-Fi card on the alternate, large stereo input do I have that, and these mics seem to need it. They're very good quality and sensitive, but they're desktop mics on a stand so it's about 2 feet from my mouth, so the gain is really important. Don't go spending $100 on a mic if you don't have a gain dial on your card.

[panthercules]

I think AVGWarhawk said it best, "it really feels like you are skippering the boat!" When I saw that, I thought to myself, "hey, another person finally gets it." It's definitely a weird feeling to speak the commands and hear the instantaneous "yes sir!". It truly changes the game. It may not be for everyone, but you won't know until you try!

That's probably one of the coolest things about it, and it has really gotten me jazzed about the work Leo et al. are doing to get more of the spoken sounds activated in the game. Having the crew confirm/repeat back your orders when you've spoken them is just way cooler than hearing them say something when you've clicked something with a mouse, and really does add immensely to the immersion experience.

BTW - I'm using a cheap (<$50) plantronics headset/microphone combo and getting great results from it, so maybe that says their gear is good but their QA sucks? Anyway, you clearly don't need to spend big bucks to enjoy this. Well done:up:

[vindex]

But the biggest reason is probably my wife...talking to my PC would just make her case so much more believable, in fact 100% convincing. As to what her case is?...well go figure :p

I might give it a go, but has to be on a weekend when shes out or something ;)


Captain Cox is dead on, this is my situation exactly. I couldn't have put it better myself.

[Nokia]

Tryed it but having problems getting it to work correctly for example i say "Crash Dive!" and i go to the Nav map. But I ain't gonna stop trying before i get it to work!

[tomoose]

I would second skwjaser's comment about creating a 'setup.exe' or auto-install/uninstall.

If I see even one thread about installation problems/issues that will be enough for me to hold off getting a mod. This is not a reflection on the modder or the mod but rather a reflection of my poor troubleshooting skills should something go wrong (i.e. visions of complete meltdown due to an error on my part and reinstalling SHIV from scratch, LOL).

Numpties like me thank god for Jonesoft!! LOL.:doh:

[minsc_tdp]

I would second skwjaser's comment about creating a 'setup.exe' or auto-install/uninstall.

If I see even one thread about installation problems/issues that will be enough for me to hold off getting a mod. This is not a reflection on the modder or the mod but rather a reflection of my poor troubleshooting skills should something go wrong (i.e. visions of complete meltdown due to an error on my part and reinstalling SHIV from scratch, LOL).

Numpties like me thank god for Jonesoft!! LOL.:doh:
Well doing an install at this point wouldn't make any sense. The program is a free-standing EXE that just runs. You unzip it and run it. Installing it would just copy it somewhere, which you've already done by unzipping it. There's nothing to configure or set up within my stuff.

The separate prerequisite of MS Speech SDK has to be downloaded and installed separately, but that takes a few minutes and the install is a "next next next" thing, no choices to be made at all. An install would add no value at this point. All of the install/setup discussions you might have read in the thread are extremely obsolete now that I'm using compiled EXEs instead of scripts that require perl/python.

If I made it JGSME compatible, would that give you a cause for a second look? I recently had some ideas on how to do this, it wouldn't be that hard.

[minsc_tdp]

Tryed it but having problems getting it to work correctly for example i say "Crash Dive!" and i go to the Nav map. But I ain't gonna stop trying before i get it to work!

Usually a recognition this bad is due to very poor audio quality. Start > Run > sndrec32, record yourself and play it back. Listen for staticy bursts or see if it's too quiet. You usually have to crank your speakers way up to hear yourself, this is pretty normal, but it should sound clear.

Another likely cause is bad config under Control Panel > Speech > Settings. The top slider should be about 3/4 to 4/5ths of the way to the right, the bottom one all the way to the right.

Of course, Crash Dive! just presses the C key... is that still the right key for your game config? Did you assign Map to C? :) I know unlikely but I have to ask :)

[Nokia]

Tryed it but having problems getting it to work correctly for example i say "Crash Dive!" and i go to the Nav map. But I ain't gonna stop trying before i get it to work!

Usually a recognition this bad is due to very poor audio quality. Start > Run > sndrec32, record yourself and play it back. Listen for staticy bursts or see if it's too quiet. You usually have to crank your speakers way up to hear yourself, this is pretty normal, but it should sound clear.

Another likely cause is bad config under Control Panel > Speech > Settings. The top slider should be about 3/4 to 4/5ths of the way to the right, the bottom one all the way to the right.

Of course, Crash Dive! just presses the C key... is that still the right key for your game config? Did you assign Map to C? :) I know unlikely but I have to ask :)

Well found out that the problem was that the settings i had where all messed up and now it runs like it's supposed to and it's great!:yep:

[skwasjer]

Well doing an install at this point wouldn't make any sense. The program is a free-standing EXE that just runs. You unzip it and run it. Installing it would just copy it somewhere, which you've already done by unzipping it. There's nothing to configure or set up within my stuff.

The separate prerequisite of MS Speech SDK has to be downloaded and installed separately, but that takes a few minutes and the install is a "next next next" thing, no choices to be made at all. An install would add no value at this point. All of the install/setup discussions you might have read in the thread are extremely obsolete now that I'm using compiled EXEs instead of scripts that require perl/python.

If I made it JGSME compatible, would that give you a cause for a second look? I recently had some ideas on how to do this, it wouldn't be that hard.
An installer still makes sense to me. I'm working professionally in the field and my team is constantly receiving feedback about our products. And trust me, it's the simplest things that hold users back.

While my S3D installer isn't perfect either yet (it does not install DirectX for users, amongst other details), I strongly believe an installer must take away all the steps involved of setting an application up. It's why installers are invented in the first place... This includes installation to a default folder (but changable per user pref), creating shortcuts (start menu, optionally desktop) to the main executables and help documentation and all kinds of other things (perform system check/requirements, uninstallation, etc.). Your solution of a zipfile, with an runtime exe still requires the user to manually navigate to a location, create folder, drag the files to it, create shortcuts, etc.

In any way, I don't really mind what you decide or do, just thought I'd add my 2 cents...

[minsc_tdp]

I strongly believe an installer must take away all the steps involved of setting an application up
All good points. I think I take it for granted that everyone knows how to unzip, make shortcuts, and that they'll always unzip it the exact same way to the same location, etc, so I am leaving a few steps to the user.

One thing I'm concerned about is the CSVs though. Some users customize these, so they have to manually merge in any new changes with an updated release into their existing files. The steps to do this vary by user depending on their changes so the installer would have to deposit these files into a holding area if it detected the files were different. But that really wouldn't help all that much, other than providing a warning that "hey there's some differences here you need to take care of."

In the end I just see an installer as a giant bloated thing that does a few basic things that everyone can already do and adds virtually no value, and then starts developing problems of it's own.

I appreciate the ideas and I'll be constantly weighing the value of an installer versus the complexity it creates.

One of the inspirations I used was uTorrent, which is undoubtedly the most popular free Torrent program out there. It was recently bought by BitTorrent.com to become the official client. Amazingly, it was a single EXE, no other files whatsoever, no install, and a tiny memory footprint. Just runs in place, and yet like any other app, always remembers it's settings and has a million features. The uninstall is the delete key. The shortcut is unnecessary. The install is done when your browser is done downloading it. I wish my stuff could be this simple but it's pretty close!

EDIT: lol I just tried the latest uTorrent and funny enough, when you run the EXE it says "Do you want to install?" with desktop shortcuts and whatnot. But they've stayed true to the original design - if you say "no" it just runs anyway, and subsequent launches don't bother you about it. :)

[skwasjer]

A solution to the csv's is to split user settings apart from application (predefined) settings. Have the csv's the user can change sit in another location (My documents, or a subfolder of your app), and let settings in these files override any application defined settings. If you (the developer) have made changes that break these user csv's, you could notify the user with subsequent updates.

In fact, providing a visual UI to edit the settings would be even better. Allows less advanced users to change commands and bindings, while not screwing up everything.

I agree uTorrent is a good and powerful tool. And yes, they've also taken note of the common problems many user encounter, see my point? Not everyone is a guru, especially your target audience, keep that in mind...

[minsc_tdp]

Reading additional, seperate user CSVs is a really good idea, and easy. Wish I'd thought of it earlier! I could implement that and let people migrate any existing mods from the stock CSVs into the user ones at their own pace, and that would make upgrading really easy.

Another thing I do plan to do from here on out is less frequent updates, and when updating, carefully explain what is necessary to preserve customizations. As more time goes by and updates are less frequent, the number of modded CSVs will increase so that will become more important.

[dr1zzt05]

I've tried it and don't like it because I still use my mouse for things while in game. I like to use an out side view and watch my torps go at the target and if while doing so I give a command like dive or stop my screen view goes in all different directions. Same with while i'm using the deck gun (i shoot better than my crew) I like to be able to call out all stop or right full rudder (if I notice a destroyer is about to hit me) without the gun being yanked violently to the bottom right. But to be honest I think it is a really awsome Idea and program. If I could get past the need to still use my mouse and rely solely on voice it would be a program I would use everytime I play. Ive used the speech thing for sh3 all the time I don't mind the looks I get from my wife or the crap my older brother gives me for talking to my game lol. Hey but the game talks back.

[panthercules]

I've tried it and don't like it because I still use my mouse for things while in game. I like to use an out side view and watch my torps go at the target and if while doing so I give a command like dive or stop my screen view goes in all different directions. Same with while i'm using the deck gun (i shoot better than my crew) I like to be able to call out all stop or right full rudder (if I notice a destroyer is about to hit me) without the gun being yanked violently to the bottom right. But to be honest I think it is a really awsome Idea and program. If I could get past the need to still use my mouse and rely solely on voice it would be a program I would use everytime I play. Ive used the speech thing for sh3 all the time I don't mind the looks I get from my wife or the crap my older brother gives me for talking to my game lol. Hey but the game talks back.

Well, unfortunately the really cool feature/advancement over earlier voice apps, the ability to use voice commands to do mouse-driven actions, can have the downside you have experienced. I still use my mouse for many things in the game, and it sometimes also happens to me, if I've forgotten to click out of mouse-view/pan or whatever it's called before I speak a mouse-based command (it's become second nature to me for the most part, but I still slip up occasionally).

There are easy workarounds to many of the problems you're seeing, in that many of these common commands that you mention (e.g., the speed commands, and rudder commands (at least full left/right and amidships), and dive, etc.) all have key-stroke equivalents and do not need to be mouse-driven - therefore, you can simply set those commands up in SH4Speech to use the keys and not the mouse method, which would let you issue those commands regardless of whether you are using the deck gun or external view or whatever. Some of the cool things he's done, for example making fine depth or course change adjustments by voice, do have to use the mouse so you would still have this problem with those commands, but they aren't really the sort of thing you're likely to need to do in a panic/hurry while at the gun station or whatever.

I think it's just a matter of learning what you like about it and what you don't, and then disabling or not using the parts you don't like while still being able to use the rest of it. That's basically what I've done with my set-up - while I think what he did with the TDC, etc. is a very impressive accomplishment, it doesn't really suit my style of play so, after trying it a bit, I just cut it out of my .csv files and added in a few things I liked better, till I had something tailored to the way I wanted to play. He's done a great job of making it accessible for folks like me with no real programming skills (but some basic familiarity with Excel) to customize it to their liking.

I think you should give it another shot before you just chuck it - there's really no need to throw the baby out with the bath water, so to speak, and it really does bring the game "alive" when you bark out your orders and your crew answers you back. :rock:

[Notewire]

I am one of those guys having a problem - just put Vista on my computer and it seems like there is a compatability problem perhaps?

The issue is that when I turn my mike on in Vista - it automatically uses the voice recognition to affect things in my computer - and if I turn my mike off in Vista - the voice commands don't get through to SH4. Please let me know if you have fought this one before.

Thanks in advance for the help,
Notewire.

[panthercules]

I am one of those guys having a problem - just put Vista on my computer and it seems like there is a compatability problem perhaps?

The issue is that when I turn my mike on in Vista - it automatically uses the voice recognition to affect things in my computer - and if I turn my mike off in Vista - the voice commands don't get through to SH4. Please let me know if you have fought this one before.

Thanks in advance for the help,
Notewire.

Hopefully minsc_tdp or someone else can help you out here - I'm still avoiding Vista like the plague and will continue to do so until hardware or DirectX problems force me to take the plunge, by which time hopefully MS will have worked the bugs out with a service pack or two (but then, I'm an OS Luddite - I'd still be running Win 98SE if I could still get driver support for it for the newer hardware). At any rate, since I don't use Vista I can't address your mike problems with it - hope you get it sorted though, 'cause this speech thing is very cool when it works.

[Notewire]

Ok, I got it working, it can work fine with Vista. Only problem I have encountered is that some of the commands aren't recognized (my computer is very voice trained) - is there a seperate profile needed for RFB? Or maybe I am just confused.

[panthercules]

Ok, I got it working, it can work fine with Vista. Only problem I have encountered is that some of the commands aren't recognized (my computer is very voice trained) - is there a seperate profile needed for RFB? Or maybe I am just confused.

Glad you got the Vista problem sorted out. It's been a while since I actually played around with setting up my .csv files for this (the equivalent of a "profile", I suppose), so I'm going from probably faulty memory here. Also, I manually added a bunch of new key commands to my own setup in order to use them with this speech program, so my "profile" is a bit non-standard at this point. However, IIRC I think it is possible that RFB changed some of the key commands from the regular stock setup, which might result in some of the commands in SH4Speech not being recognized if they're based on stock.

If you can post which particular commands seem to be giving you problems, it might be possible to track down why that is and come up with a fix/workaround.