Здравейте,Препращам ви една кореспонденция между мен и Jonathan Duddington относно няколко бъга в eSpeak за които му писах и които мисля че сме обсъждали и тук. Кореспонденцията е на английски и ако това ви затруднява - кажете и ще преведа поне основно за какво става дума.
Поздрави, Коста -------- Оригинално писмо -------- Тема: Re: Bugs in eSpeak Дата: Sun, 15 Jul 2012 09:03:50 +0000 (GMT) От: Jonathan Duddington <jonsd@xxxxxxxxxxxx> До: Kostadin Kolev <kostadin.kolev1985@xxxxxxxxx> On 11 Jul, Kostadin Kolev <kostadin.kolev1985@xxxxxxxxx> wrote:
I have a few questions to discuss with you about eSpeak:
1. Bug: This is related to NVDA as well, but anyway. In some cases (specially when a text string begins with 2 or more upper chars), NVDA may give an error, when reading with eSpeak in NVDA's "say all" mode. In similar conditions, it may occur that some parameters of eSpeak will get modified (e.g. speed, pitch, etc.) as well.
I am not aware of this problem. It has not been reported to me by the NVDA developers. I do not get this problem when I use eSpeak by itself. So I need to know exactly how NVDA calls eSpeak to produce this effect.
2. Abbreviations: the syntax used for abbreviations in eSpeak should be case sensitive and should be able to accept dots at the end of the abbreviated strings. I don't know if at present it is possible, but it is not applied for the bulgarian language and that causes a lot of misinterpretations of abbreviations.
Yes, eSpeak data can specify that an abbreviation: 1. Can be followed by as dot. Attribute $dot 2. Must be followed by a dot. Attribute $hasdot 3. Must be all-capitals. Attribute $allcaps 4. Must start with a capital letter. Attribute $capitals These options can be set for each abbreviation which is listed in the eSpeak Bulgarian data (file dictsource/list). Please give some examples of problems.
3. Bug: When in a word with Cyrillic characters by mistake there is a letter from the latin alphabet, eSpeak can't read the word correctly and reads something like this: "leter for two three, letter six five one, ..." etc., as if it tries to read unknown characters.
When using the Bulgarian voice, if eSpeak finds a word which contains Latin characters, it speaks the word using the English voice. But the English voice doesn't know Cyrillic characters. So what should eSpeak do if it finds a word which contains both Cyrillic and Latin characters? poluchavate tozi e-mail, zashtoto ste chast ot poshtenskia spisak eSpeak-bg. Za da izpratite pismo do tozi spisak izpratete e-mail na: espeak-bg@xxxxxxxxxxxxx za poveche informacia za tozi spisak, za da se zapishete ili otpishete, posetete: //www.freelists.org/list/espeak-bg