Hi. I live in Japan, and I notice when I’m designing routes on the site, the Liberty map has place names in Chinese in the roman alphabet (sorry, I don’t know which Chinese), whereas the OSM has them in Japanese kanji.
I mention this because when I export to a gpx track to put on my phone to use with the OSMAnd+ app, the TTS says road names as Chinese. The choice of map doesn’t seem to affect the gpx export much. Is this correct?
I’ve been asking on the app support Telegram but there doesn’t seem to be a reason for it in the app, or at least we’ve eliminated suggestions.
Which OSM map did you compare liberty to? Kurviger OSM? Are the names of Kurviger OSM correct?
This is correct, map choice does not influence the routing.
If you export a GPX track, the track should not contain any road names. So in this case it is very likely that OSMAnd stores the Chinese names as well?
To explain a bit more, we use OSM data, here is one city for example. Not sure how your sample data looks like, maybe you can share more info on that?
all my app settings in OSMAnd to make sure I only had Japanese and English installed
tried with both OSM FAUand Liberty selected (and checked the .gpx in a text editor) - same result.
OSM as base maps are the only things in common between the two. I wonder then if in OSMAnd I need to find what map code/tag it’s using - the image appears in Japanese, but it is speaking them via TTS in Chinese. Thanks; I’ll do some more digging.
It’s the same thing with the Kurviger OSM map (also pinyin). OpenStreetMap (FAU) features no romanisation at all, only the kanji, but interestingly the OpenStreetMap DE (FAU) map does have (the correct) rōmaji, so I’ll just use that one.
Please note, the OSM Fau maps are not hosted by us, this is a mirror of the official OSM map styles and we can’t really change anything about these.
Kurviger OSM and Kurviger Liberty is hosted by us, so we could change something here.
Unfortunately I can’t read Chinese nor Japanese characters. I am happy to check if we can improve something, but I would need more details, like what is shown and what it should be like. Ideally 2-3 samples, so I can check if we can change it or not. (Can’t promise anything here, sorry )
Thanks for the response, and offering to look into the matter (which does not, by any means, hinder my use of the maps; it’s simply jarring to see Japanese kanji translitered in pinyin).
The location that I highlighted (circled) in the screenshot above, 市木, is romanised in pinyin as “Shì Mù” when it should be “Ichigi” (or “Ichiki”, both are possible).
Near to that place is 千町, romanised as “Qiān Tīng” (Chinese pinyin), though it should be “Senjyō” (Japanese rōmaji).
Earlier I said that the OpenStreetMap DE (FAU) map uses the correct rōmaji, but actually that’s not true. It has 市木 as “shi ki” (mixing on and kun readings, perhaps based on frequency, as 市 is most often read as “shi”) and 千町 as “sen machi” (same thing, mixed readings, perhaps based on frequency analysis).
I did want to provide the examples, as requested, but having slept on this matter I’m not sure that it’s reasonable to expect Kurviger to ensure that the maps it uses will have Chinese characters transliterated correctly depending on the country/region in which they’re used. So if you are able to include proper rōmaji on Japanese maps, that will be very much appreciated; and if you can’t, that’s OK.
I should note that major place names are transliterated properly in rōmaji, so for example 倉吉市, north of my screenshot, is correctly shown as “Kurayoshi”. It’s only the names of smaller places that are transliterated in pinyin, so this really is a… small issue.
It seems like the Shì Mù is generated. This is done by a software that is part of our map stack. So unfortunately there is not much we can do short term about this.
I think it’s great that we are aware of the issue. In general I am happy to fix issues like that, if it’s possible. In this case I fear it’s not possible for us to fix this easily.
Thanks for following up and explaining why some (bigger) towns are correctly named while most of the smaller ones are not. Not such a big deal; I will live (easily) with it.
Hi.
This is still happening for me, so I posted into the OSM support forum here, and they suggested a couple of things the app devs could do:
the app developers need to detect the language e.g. by setting it using country boundaries. What would be needed is to trace which part of the software stack is doing the transliteration and fix it there. It can be done. I think Organic Maps does this correctly.
Thanks for letting us know. As commented above, one option would be to add the correct name tags, that would show the correct naming. We still have this on our radar and we hope that the issue will be fixed/improved in the software we are using to render the map. We are also planning a few improvements and updates in the future but it’s nothing we can easily hot fix.
I searched for something to help me before posting this, but could not find anything specific.
I live in Japan, but mostly read/use English. I downloaded all Japan offline maps, but often when I am using the app, the places are not even in Japanese, they are in Vietnamese, Chinese and Spanish sometimes. Other times they show in English or Japanese.
The only tip I could find was to download the offline maps, but it did not change the issue.
I understand that not all of OSM will translate to my preferred language, but why would it translate to multiple others instead?
Any suggestions are appriciated. Thank you.
I moved your question, please see the explanation above. We are planning a bigger map update over the coming months so hopefully this will already improve the situation. For more info, also see the info above