Text to speech idiosyncrasy

[Wildeblood]

When selecting a prime-able equipment item called "Cross-hairs selector", the text to speech engine announces "Credits oss-hair selector". I assume any other word beginning "Cr" would suffer similarly.

[another_commander]

Actually the problem is that the " Cross-hair selector" string seems to have a space at its start. The speech engine will substitute "Cr" with "Credits" only when that space at start is present and the C in Cr is a capital one. If you try using the debug console to do

Code: Select all

P.consoleMessage("Cross-hair selector")

after having set Spoken Messages to On, you will find that the message is spoken correctly.

[Wildeblood]

I was about to apologize for raising a false alarm, and making such a noob mistake, then I thought I'd better check first.
This is the equipment.plist, and I don't see a leading space:-

Code: Select all

(
    (
        0, 0, "Cross-hairs Selector", "EQ_SELECTABLE_CROSSHAIRS", "",
        {
		"available_to_all" = "yes";
		"damage_probability" = 0.0;
		script = "selectable_crosshairs.js";
		visible = "no";
        }
    )
)

[another_commander]

Interesting. Maybe the leading space is added during processing of the string inside the script that handles the equipment? I only checked the case briefly using the debug console and the only way to have Cr spelled as Credits as to add that space.

[Wildeblood]

I've just uploaded Select-o-matic Crosshairs, which is now available using the OXZ manager. I hear the same mis-pronunciation in Oolite 1.80 and the nightly build.

[another_commander]

The extra space is added by the game, when setting up the string: "Equipment primed: Cross-hairs Selector". You can avoid the Cr-Credits substitution by adding a tab character at the beginning of the equipment's name in equipment.plist, which is invisible and therefore you bypass the problem. Like this:

Code: Select all

(
    (
        0, 0, "\tCross-hairs Selector", "EQ_SELECTABLE_CROSSHAIRS", "",
        {
		"available_to_all" = "yes";
		"damage_probability" = 0.0;
		script = "selectable_crosshairs.js";
		visible = "no";
        }
    )
)

[Neelix]

That seems a bit hacky...

Could the speech code possibly be modified to only do the credits substitution if the r in Cr isn't followed by another letter?

- Neelix

[Zireael]

That seems a bit hacky...

Could the speech code possibly be modified to only do the credits substitution if the r in Cr isn't followed by another letter?

- Neelix

Good point. IIRC there's a few planets whose names start with "Cr".

[cim]

Could the speech code possibly be modified to only do the credits substitution if the r in Cr isn't followed by another letter?

No, at least not straightforwardly. There's the speech_pronunciation_guide.plist file, so you could put a proper pronunciation for "Cross" into it - but because of the way it gets merged, the one for " Cr" will get picked up first.

It seems slightly odd that in that file " cr." is "credits", but " Cr" (no trailing .) is also "credits" - but there are enough OXPs out there that use " Cr " for credits that it's probably got to stay that way.

We could add "Cross" to the core file - it's a common enough word or word-component that it's worth getting right. That would at least fix this case if not others. Looking through descriptions.plist, though, there are a lot of others.

I think the correct fix - and looking at existing OXPs this shouldn't break anything - is to change the " Cr" in the core file to be " Cr " and " Cr." as two separate entries.

IIRC there's a few planets whose names start with "Cr".

There are none, at least without OXPs - the name generation rules make that substring impossible.

[bananaboat]

Hi everybody, I've recently upgraded my Oolite to 1.80 from 1.75something (brilliant job, guys!) and it's my first time posting here.

Anyway, I think there is a fairly straightforward way to fix this problem.

I encountered this after installing the Random Ship Names OXP (which is very good, by the way) - this generates ships with names like "Gecko: The Crow and Langoustine II", which the speech synthesizer renders as "Gecko: The Creditsow and Langoustine Roman Two".

I can just about live with "Roman Two" for "II", but "Creditsow" for "Crow"? I found the problem in the speech_pronunciation_guide.plist, and I think I've fixed it by creating another speech_pronunciation_guide.plist in my Add Ons folder - basically it allows Cr on its own to be pronounced "Credits", but Cr immediately followed by another letter gets pronounced correctly. I'll post the file here if I can figure out how to do so.

[Cody]

I'll post the file here if I can figure out how to do so.

Copy and paste the .plist into a post, highlight it, then hit the 'Code' button - and welcome aboard, Commander!
Alternatively, host it somewhere (Box or Dropbox, for example) and post a link.

[Wildeblood]

This one has already been fixed in trunk (1.81), BTW.

[bananaboat]

Here goes:

Code: Select all

(
// Format:
// (
//   [required] Original text,
//   [required] replacement text (Apple),
//   [optional] replacement text (espeak) or "_" (original text is unchanged)
// )
	(
		" credits",
		" Cr"
	),
//main game sets "Cr" to "credits" - we need to re-set it

	(	"Cra", "cra" ),
	(	"Cre", "cre" ),	
	(	"Cri", "cri" ),
	(	"Cro", "cro" ),	
	(	"Cru", "cru" ),
	(	"Cry", "cry" ),	
//correct Cr if followed by another letter (vowels and y)

	(	"Va", "va" ),	
	(	"Ve", "ve" ),	
	(	"Vi", "vi" ),
	(	"Vo", "vo" ),	
	(	"Vu", "vu" ),
	(	"Vy", "vy" ),		
	(	"Vl", "vl" ),	
//roman numerals - if V is followed by another letter then not 5

	(	" IV", "Four" ),
	(	" VIII", "eight" ),
	(	" VII", "seven" ),
	(	" VI", "Six" ),
	(	" V",	"Five" ),
	(	" III", "Three" ),
	(	" II", "Two" ),				
	(	" IX", "nine" ),

//roman numerals - order is important

(	"S.S. ", "s s " ),

// otherwise says "s dot s"

	(	"RRS", "R  R S" ),

// tweak for RRS OXP

(
		" Cr",
		" credits"
	),

// re-set "Cr" as per main game

)

Copy this lot into a file named speech_pronunciation_guide.plist and put it in your Add Ons folder. You don't need to do anything with the original speech_pronunciation_guide.plist in the Resources folder (I'm on Windows 32 bit - hope Mac etc. is the same)

This also includes roman numerals from 2 to 9 and a couple of other tweaks.

[fronclynne]

Aren't roman numerals normally used as ordinals in English? eg "second", "tertius", wha-have-you.