Busstop announcement tool

While I'm working on my own map, I noticed that I really want (need) busstop announcements. To be honest, my voice isn't particularly sexy. The voice of my girlfriend is better, but her English pronouncements are not fluent.

So as a programmer I looked for some technical solution to create some more professional sounding announcements. Not only do I want to announce the next stop, I also want to play a jingle before and after each announcement. Today I sat down at the kitchen table and started defining some requirements. Now I know that a lot of map makers also want announcements, so I thought maybe I can make other happy too! The tool is uses the Microsoft Speech API 5 (SAPI) and I want to use the HOF file as a starting point for setting up the announcements.

I haven't worked out everything, but this is what I have defined so far:
- Busstop / terminus prefix jingle
- Busstop prefix announcement

- Busstop announcement

- Busstop post announcement
- Busstop / terminus post jingle


Every stop is composed of 5 'events'. First there is an option to play a jingle (like a chime or ting (link)). The tool makes a distinction between a 'regular' busstop and a terminus busstop. It allows you to play a different jingle when you reach the terminus. Then there's the 'fixed' announcement (busstop prefix announcement) that is played before every announcement. Here you can define something like "Next stop' so you don't have to include that in every announcement.

Then we have a announcement of the busstop itself. By default the tool uses line 3 and 4 from the busstop definitions for the initial announcements. After the announcement of the stop there's the post busstop announcement and chime that work in the same way as the prefix ones, but only in reverse order.

Every announcement part is composed of (multiple) text and pause segments. The pause between between these 'events' can be changed. The tool will combine the 5 events (in the order describe above) and record the result to a wav file and is named exactly as the busstop including any diacritic symbols.

So is anyone interested in such a tool and if so, do you have any remarks or additional wishes?

Looks interesting... Looking forward for this release.

Yes, i am interested .






WinXP

Sounds good!

Very interesting. I hope that you are not using the windows XP API as it is heavily incompatible with later versions of windows

Nice field of speech syntesis. There are some languages for which it's quite hard to create nice speech. But good luck!

Amy Rose schrieb:

Very interesting. I hope that you are not using the windows XP API as it is heavily incompatible with later versions of windows

The API I'm using is provided by the Microsoft .NET framework and in the past (2007) I created a RSS reader that would 'read' you the article. That application worked both on XP and Vista, so I guess Microsoft framework already took care of that.

However I noticed that not all SAPI 5 voices are compatible (at least not the ones that were converted from XP to vista). Some of the voice manufactures do offer some sort of guarantee. However this window is usually very short (less than 2 hours) and you are required to mention the application where the voice was going to be used. If it's a rare application, they might ask you to upload it to their FTP server. To prevent problems most manufactures only have 'full' versions of their voices. Some provide a few low quality version often with added noise so U can't use it in an real application.

nemeza schrieb:

Nice field of speech syntesis. There are some languages for which it's quite hard to create nice speech. But good luck!

Well each voice is created for a specific language. In the past I created a RSS reader with support for English, Dutch, German, French, Spanish and Italian. I can only speak Dutch and English fluently and I do know some German words. In the second grade of collage I had to transfer schools near Christmas. My old school only started with German in the second year while the new school started in the first year so I attended only the last 6 months. In my third and fourth year German wasn't part of my classes. Although I can't speak it properly, I do understand most of the words when they're spoken as German and Dutch are quite similar.

I do find it petty know, because since I moved near Eindhoven I'm only 20 minutes from the German border. Occasional we visit Frankfurt, Dordmunt and Oberhausen. The last one only where there's a rave..
Being able to read German would also helped with OMSI ;-). But Google translate does get the job done. Hooray!

I have no idea if the tool would create proper announcements for Cyrillic, Arabic or Asian languages. Within a few weeks I will also release the tool as open source on Github so other can make contributions to it to support more languages. For now it resides in my private TFS repository.

It would be nice to have a automatic announcement system, based on GPS coordinates so you don`t have to press a key.

Its possible, but the trigger has to be wide enough, and it wouldn't be by gps coordinates

(Wrong section, moved it )

All this sounds very interesting, I'd have an additional suggestion:

Sometimes people like to record their own announcements or use other tools (like the voice of Google Translate).
You could implement tho modes: One like you mentioned, and with the other one you select a directory with self-recorded announcements and you can plus select e.g. a terminus sound or a "next stop" sound.
Your software could then stick these parts together so that in the end you get complete announcements. Just a suggestion, but I think a two-modus-version could be a good idea

Janniboy schrieb:

(Wrong section, moved it )

All this sounds very interesting, I'd have an additional suggestion:

Sometimes people like to record their own announcements or use other tools (like the voice of Google Translate).
You could implement tho modes: One like you mentioned, and with the other one you select a directory with self-recorded announcements and you can plus select e.g. a terminus sound or a "next stop" sound.
Your software could then stick these parts together so that in the end you get complete announcements. Just a suggestion, but I think a two-modus-version could be a good idea

I have thought of that before (although you also have to record the 'next stop' sample) otherwise the announcement would sound like an automatic voicebox. Joining several samples together to one 'recording' isn't that hard (as long all samples have the same bitrate and quality). I have the generator module already designed as a two-stage process. When you provide you're own samples it's not so hard to skip the first step where the tool generates the samples and only merge the samples.

However I don't think it's wise to use Google Translate (or a similar) tool to make translations for announcements. Google Translate is great for translating a document so you understand what it says. However these translation are far from perfect. Naturally as a map designer this is there choice.

CityBus schrieb:

It would be nice to have a automatic announcement system, based on GPS coordinates so you don`t have to press a key.

Except for parsing a HOF file for some initial announcements, the tools doesn't have any knowledge about Omsi. This tool is meant for map creators so they can easy add announcements to their map.

However even in real life these information systems can't rely on GPS only. There are always situations like detours where GPS coordinates simply aren't enough. They easiest way to trigger the next announcement would be using the front door trigger with a delay. However this requires a mod for IBIS. Looking at the 'IBIS' from the Karosa bus, I even doubt if you can get away with a generic mod. My tool is for creating the announcements, not to trigger them while playing Omsi.

Just about Google Translate: I meant e.g. the translation from German to German, so I'm interested in the audio output, not in the translation itself

EDIT: I noticed I forgot to move this topic when I wrote my last post... gonna do it now