Test of Text-to-Speech (TTS) Abbreviations in Waze Clients

Map Editors Community United States
1

This is a list of abbreviations and acronyms that are being added to two sandbox test rigs, to be tested in the Waze clients to learn their TTS output results for North America. (Jenncard maintained a second test rig.)

All tests are with the Nuance female “Samantha” voice only as used in North America. While we expect that the other English voice choice for NA, “Tom”, is likely to use the same TTS databases, we have not tested for it.

The list includes all the official and alternate abbreviations of the United States Postal Service, all the English abbreviations for Canada Post, a subset of the French abbreviations for Canada Post, a subset from the U.S. Federal Manual on Uniform Traffic Control Devices plus a number of others suggested by Waze users. (Additional abbreviations from the MUCCD for individual U.S. States and from any equivalent list for Canadian Provinces, have not yet been compiled.)

Official abbreviation list sources for specific states or providences are linked from their respective names in the test list.

If you already know or learn any TTS results, please post them here. Please submit only results that you are absolutely sure of, with the exact spelling (including any capitalization) of the abbreviation, and the confirmed TTS output result.

Also submit any additional abbreviations that you feel should be part of the list for testing and/or for future inclusion in the TTS system.

Some current abbreviations such as “Penn” (which generates “Pennsylvania”) should probably be removed from the TTS system in favor of using the capitalized two letter State and Province abbreviation standards. We should also discuss others for consideration that may need to be removed from the TTS.

To hear the most current TTS results from your Waze client device, it may be necessary to clear your current TTS cache. Enter “cc@tts” without quotes in the Navigation search to clear the cache. Waze should reply with a “TTS cache has been cleared!” popup box. Then wait for “Searching …” to end.

This list will be updated periodically. (The initial list was submitted 2012-01-18, which took a month for Waze to process, and for us to test in the Waze Clients.)

Once we learn what the TTS currently knows, we can add the results of this list to the Wiki, plus
we can use it to submit our requested TTS “wishlist” changes to Waze. Please note that entries in this are for testing purposes and do NOT automatically imply that we endorse them. This list is primarily to research as many possible abbreviations that we can test to learn what Waze currently recognizes or not.

To learn how to hear the TTS files on your Android device manually, read the side notes in the following post: http://www.waze.com/forum/viewtopic.php?f=10&t=15178&p=144296#p144296

A section on Abbreviations has been added to the Waze Wiki. The link below takes you to a selection of countries. Select your country from the list and follow the links to the abbreviations in your country:
Editing Manual > How to label and name roads (by country)

In the United States you can go directly to the US Abbreviations and acronyms list.

Google Drive Spreadsheet:
Text-to-Speech (TTS) Abbreviations in Waze Clients - Test List for North America aka the “Big List” or the “Big TTS Test List”.

2

I’m not sure how waze plans to handle regional/national/language diffences, but I am assuming that currently we are dealing with US English. If that is the case I would add in the state abbrieviations. For instance “MD” is pronounced “emm-dee” by waze. “Penn” is a problem and is pronounced “Pennsylvania.”

I would also add the following:
Alternate - ALT
Bound - BND
Crossing - X-ING
Expressway - EXPWY
International - INTL
National - NATL
Route - RT
Throughway - THWY

I would also note that some abbrieviations are ambiguous: ST for Saint or Street and DR for Doctor or Drive. Also some streets are lettered thus there are “E St (or Ave, Blvd etc.),” “N St,” “S St” and “W St” out there. Ultimately, there will need to be a method for distinguishing “St” when it means Saint and “St” when it means Street and “N” when it means North and “N” when is should just be “N.” If this is done with punctuation then we need to test “St” and “St.” etc. If this is done by parsing the phrase, that is “N” is N if followed by a road designation, but north otherwise, then we will need to indicate the different results depending on the phrase.

3

Thanks for putting together the list. Very thorough. I look forward to seeing the results as they come in.

4

Was this also to end up being an official list to help them differentiate things like Penn vs. Pennsylvania, St vs. St. (saint) and Dr vs Dr. (doctor)? I think we need to add those too, unless I’m missing the purpose of the full abbreviation list.

5

EXPY also works for Expressway.

6

CBenson:
Excellent points CBenson. Some of the words you mentioned were already in the list I believe, however I’ll be sure to add the ones you mentioned. At the Palo Alto Meet-Up there was discussion about many of the issues you mentioned. We did kick around the idea of using “St.” for “Saint” and “St” for “Street” as you mentioned. I’ll be sure to add “St.” The list does have NE. NW, SE, & SW to see what they generate, but I forgot about the issue of differentiating roads named with a letter such as “E” from “east” as the current “E” abbreviation currently says. Since there are fewer roads named with a letter than roads that include a direction, we will have to think of a way to make them different. Maybe with quotes? Anyone else have ideas here?

I suppose I should test every letter of the alphabet to see what they say. N, E, S, & W may not be the only ones that produce a word.

mapcat:
Your welcome… altough I don’t think I got any points for making that list did I? :slight_smile:

AlanOfTheBurg:
Actually Alan, the list was to be multipurposed. Intially to learn what abbreviations the TTS already knows, and secondly to be a basis for creating our TTS “wishlist” for Waze. I’ll see if I can make that clearer in the lead post. Thanks.

jason300
Thanks Jason.

7

Request to add (does not work currently)

C - Calle - The abbreviation would be the first word in the street name

We’ll need to add a bunch of Spanish abbreviations since there are loads of streets in many states where the street name is in Spanish.

Pso - Paseo - confirmed working in TTS
Av - Avenue - confirmed working in TTS

8

Geez, I wonder who we could ask that might know those!

9

Jr. = Junior in TTS, as in Martin Luther King Jr. Blvd. I didn’t see it in your list. I found that without the period, “Jr” induces TTS to say “jay are”.

I expect the same result for “Dr.” as in “Dr. Martin Luther King Jr. Blvd”. Since most every town has a street honoring MLK, it’s appropriate to include it in your extensive list. Nice work, by the way.

10

Basically there needs to be positional awareness for handling things like St, Dr, N/S/E/W. Can anyone think of a name where St isn’t Saint unless it’s at the end of the name and therefore Street? Same thing for Dr.

N/S/E/W is the letter name if it’s in the middle of the street name, or the only “word” before the St/Rd/Blvd/etc. So you can have “W N St” for “West Enn Street” and “N St” for “Enn Street”, if you need “West North Street” you’ll just have to use “W North St”.

CR-### should be County Road, I’m not sure what we should use for State/County Route, maybe RT-### and TTS be “Route ###”. (Rowwt or Root? it’s regional)
SH-### should be State Highway ###.
US-### should be You Ess ### or You Ess Highway ### (I’d prefer to not have Highway but that’s just me)
I-### should be Eye ###

11

Worst case (for single letters):

N N St N

Should be:
North Enn Street North

:smiley:

12

According to this post, it may not work that way:

13

I’ve added everyone’s suggestions to the list, however I have not added anything to my sandbox yet. (Nor have I achieved any editing points in the last day! :frowning: )

Among the items added were:
Doctor = Dr.
Junior = Jr.
Saint = St.

Following that convention I’ve also added the following for the time being:
C (cee) = C.
E (ee) = E.
N (en) = N.
S (ess) = S.
W (double-u) = W.

Additionally all the state and province abbreviations were added.

A link at the bottom of the spreadsheet to Post Canada shows their official abbreviations, that I will probably merge with the current list at a later time.

14

Having written my share of regexps and parsers, it occurrs to me that if we are going to use the period/full-stop ‘.’ character to differentiate ‘N’ (“north”) from ‘N.’ (“enn”) or ‘St’ (“street”) from ‘St.’ (“saint”), we also need to decide whether it should have to be followed by a space. For example, should it recognize ‘St.Louis’ (no space) as “Saint Louis”?

Unless someone can think of a good reason to allow ‘.’ mid-word, I think the ‘.’ should double as a delimiter (i.e., a space should not be required), if only for the sake of saving screen real estate. “N Street” would be written as either ‘N.St’ or ‘N. St’.

15

And I think the period after the single letter indicating that it should be the letter is confusing. Plenty of signs out there that use N. to mean North and S. to mean South etc.

BTW, maybe we should find out how these are handled:
So. (South)
No. (North)

16

For the sake of clarity and reliability, you are right that we should try to minimize overloading (e.g., ‘N’ vs. ‘N.’), and try to minimize positional semantics (e.g., handling “St” differently depending on where it appears). The more “common sense” we try to build into it, the more complicated the code is to implement it, and the more likely fixing one problem will break something else.

Those would be good options to reduce overloading in street names, but we still need to address how to write “enn street” in a way that is readable, without breaking the existing database that uses ‘N’ for “north”.

17

One of my cities out in the hinterlands is littered with “27 1/2 Rd”, “D 1/4 Rd”, or one of my personal favorites “E 50/100 Rd”. Unfortunately, I’ve actually never Wazed there so I have NO IDEA what TTS does! That said, I’ve never seen a UR saying “should be one-half Rd not one {pause} two Rd”.

I’d be happy to throw down some roads to test this out, but obviously don’t want to muck up the local map. Gizmo, any tips on how you’ve set up your sandbox?

18

I can tell you that when it comes to “Co Rd 3 1/2” what TTS says is “Co Road Three One Slash Two”, with the Co as the first syllable of “cooperative”, I have yet to change one of these to CR-3.5 or CR 3.5 to see what happens.

19

‘Storrow Dr E’ is supposed to be “storrow drive east”, not “storrow doctor east” or “storrow doctor ee”. And each time someone comes up with an example that we haven’t anticipated, the rules (and code) become messier and harder to maintain.

All these conditions and exceptions are easy enough for a human to follow, but to a computer, context-sensitive grammars are brittle. Trust me, I have done this for a living. Better to start now with as straightforward and literal a system as possible, or it will turn into a perpetual problem, where fixing one case will break several others.

20

Per The Chicago Manual of Style, the correct abbreviation for Doctor is “Dr.” and Saint is “St.” (just as Ph.D. and NOT PhD is the correct abbreviation for that which I seek) :geek: