Is there a global Currency variable?

fesquivel · **Posted:** Tue Oct 05, 2004 4:00 pm

I live in a country where Spanish is the language being spoken, so I used the SetLanguage step to define Spanish. However, when using the Say step to vocalize an available amount, let's say "1450.55", it spells: "Mil cuatrocientos cincuenta DOLARES periodo cincuenta y cinco"... Where does DOLARES ("dollars" spoken) come from? In fact, the given amount is in the local currency! I already take care of speaking the correct WAV file for dollars or the local currency ("colones" in this case), but the Say step still dictates "dollars" after the integer amount and before speaking out the decimal part... How can I avoid that (the currency dictation) or at least rectify it (so as the correct currency voice is used instead of just dollars)?

Tim · **Posted:** Wed Oct 06, 2004 8:43 am

The Say step has a "format to speak" value. It is a radio button. I suspect you have it set to "Money", which instructs TeleFlow to say the number as currency. The setting you want is "Number". The words used by the Say step are in individual files on disk. You can replace any or all of the files with your own recordings, so you could replace DOLARES.WAV with the appropriate recording.

fesquivel · **Posted:** Thu Oct 07, 2004 8:41 am

I tried that in English and Spanish. The results are the same: Oct 07 09:40:18.05: [22] Say Oct 07 09:40:18.06: Speaking '1450.55' Oct 07 09:40:18.06: Playing: 'E:\32\TeleFlow\Languages\English\01.wav' Oct 07 09:40:18.78: Playing: 'E:\32\TeleFlow\Languages\English\thousand.wav' Oct 07 09:40:19.61: Playing: 'E:\32\TeleFlow\Languages\English\04.wav' Oct 07 09:40:20.28: Playing: 'E:\32\TeleFlow\Languages\English\hundred.wav' Oct 07 09:40:21.09: Playing: 'E:\32\TeleFlow\Languages\English\50.wav' Oct 07 09:40:21.83: Playing: 'E:\32\TeleFlow\Languages\English\period.wav' Oct 07 09:40:22.74: Playing: 'E:\32\TeleFlow\Languages\English\01.wav' Oct 07 09:40:23.43: Playing: 'E:\32\TeleFlow\Languages\English\04.wav' Oct 07 09:40:24.09: Playing: 'E:\32\TeleFlow\Languages\English\05.wav' Oct 07 09:40:24.74: Playing: 'E:\32\TeleFlow\Languages\English\00.wav' Oct 07 09:40:25.55: Playing: 'E:\32\TeleFlow\Languages\English\05.wav' Oct 07 09:40:26.19: Playing: 'E:\32\TeleFlow\Languages\English\05.wav' That's just weird! What's going on?

Chris · **Posted:** Thu Oct 07, 2004 9:52 am

Please send a complete TeleFlow log of this to support@engenic.com. One log of Spanish, and another of English would be helpful. Also, if you could screen capture the step properties of the say step only, paste it into a word doc, and send that as well, that would be a big help. We will follow-up in this thread. Thank you.

fesquivel · **Posted:** Thu Oct 07, 2004 2:17 pm

Done. I also attached a screenshot of the test project I'm using for this, so you can see the workflow. Is there a way to alter the Say step so it speaks the correct currency voice? Maybe by previously defining a global variable with the path to the WAV/VOX file that should be used instead of assuming it's alwasy "dollars"...

Chris · **Posted:** Thu Oct 07, 2004 5:25 pm

There are a combination of issues here, I believe. Looking at the English log, something is currently not working with saying numbers. We will be looking into this further, and will update here. Having said that, I think even when this issue is fixed, you find that you aren't playing quite what you want. If you say "1450.55 dollars", which is what your application is doing right now, it isn't likely to sound natural (or right). As a workaround, I suggest you "chop" your number variable into two variables: @AVAIL_DOLLARS and @AVAIL_CENTS. Then, play @AVAIL_DOLLARS as a number, followed by your own "dollars" playback, then play @AVAIL_CENTS as a number, followed by your own "cents" playback. If you would rather replace all of the files that say numbers ("dollars", etc) with your own, you must record all of the standard prompts/words, and put them in the appropriate directories. If you don't specify otherwise in the Set Language step, TeleFlow gets all of these from C:\Program Files\TeleFlow\Languages. If you replace all of the files in the English and Spanish directories with your own versions (you should back up the existing files first, to be safe), you can probably set the system back to saying Money instead of Number, and that should work. (It will use the correct currency voice...) These recordings are for all the numbers, dates, etc that the say step might need. (It says "One thousand" using 1.wav and thousand.wav, for example)

fesquivel · **Posted:** Fri Oct 08, 2004 11:16 am

Assuming I replace all the voice prompts and words under the Spanish directory (our customers will surely prefer a girl with a local accent for this), how will the Say step use the correct currency voice in Money mode? As I see it, Say will always dictate the integer part first, then whatever sound is recorded under "dollars.wav" (probably "dólares") and then will dictate the cents. I appreciate that fact you're admitting there's something not working quite right when saying numbers, in the current release... While looking at it and eventually fixing it (and releasing an update), would it be possible to add a parameter to the Say step, so the currency voice can be set programmatically? Or at least define a new global system variable @CURRENCY that the Say step uses to determine which voice to use when dictating Money data? Thanks in advance.

Chris · **Posted:** Thu Oct 14, 2004 10:45 am

Actually, what I was suggesting for a workaround was that you play the dollars as a Number (not Currency), then play your own "dollars" recording, then play the cents as a Number, and then play your own "cents" recording. You can control where the files come from when playing anything with the Say step. When you use the Set Language step to set the current language, you can also control where the Say step finds the prompt recordings. (Set this in the "Location of prompt recordings:" field)

fesquivel · **Posted:** Thu Oct 14, 2004 11:58 am

I understand your workaround perfectly and I will probably use it... once the Say step is able to dictate Numbers as it is supposed to do... (did you see my post on October 7th?).

Chris · **Posted:** Thu Oct 14, 2004 12:11 pm

It does say numbers properly. You must change the Say step's "Format to speak" radio button to "Number", instead of "Money". Then, when it goes to speak "1442", for example, it says only "one thousand four hundred fourty two". It is the "Money" setting, specifically, that has the problem you encountered.

fesquivel · **Posted:** Thu Oct 14, 2004 2:36 pm

You did not read my posts... The "Number" setting has another problem too: For "1450.55" (the example run above), it says "One thousand fifty period one four five zero five five"... That's what I'm talking about! See my post of October 7th for an extract of the log file. I already sent the entire logs (as asked) to the support e-mail account.

Chris · **Posted:** Thu Oct 14, 2004 2:47 pm

Sorry, I didn't realize that "Number" had the same issue. However, if you do things as I suggested, you can still avoid the problem: Play only the dollar portion using Say first. (ie. "1450") Then play "dollars", then play "and" Then, play only the cents portion using Say. (ie. "55") Then, play "cents". If you play 1450.55 altogether, it will have the error as you described. You must chop up your variable into two variables, one with dollars, and one with cents, and then play them separately using separate say steps. That is the only way the workaround can work.