(I had a go at Spanish through Welsh a few years ago, and didnāt get on very well - might be a bit better now. I also tried Catalan through Spanish on Duo for a while, as I have a fair amount of passive Spanish, but that wasnāt very effective eitherā¦)
But both Basque and Catalan are pretty high on our to do list - weāre having a slightly hiccupy time with the Popty at the moment, which is why Bulgarian and Georgian havenāt been released yet - also, weāre struggling to find voices for Scottish Gaelic, so that may need to wait until we can get volunteer recording.
Iāve also just had a request for Latvian⦠so once we get the Popty behaving again, and finish doing the voices for Bulgarian, Iāll turn my beady eye in the direction of Basque and Catalan
I have the Spanish through Welsh downloaded - I wanted to make myself make use of some of the South Welsh stuff without trying to use it actively (and just fancied the challenge). It did sometimes reduce me to hysterical laughter when all higher brain functions just seized up⦠It also made me feel really impressed with everyone here whoās learnt Welsh when theyāre not 1st-language English speakers.
An update to checks that Iāve made on a Samsung A53 5G with Samsung OneUI 8.0 and Android 16ā¦
I have used Firefox, Chrome and Samsungās Internet Browser. I have tried using the app in each browser as well as installing the app from each browser. I have set battery usage to unrestricted for all. Additionally in the Samsung offering there is an option to allow background sound, which I have enabled.
The result is the same in all six scenarios. Once the screen is turned off, the sound continues for a phrase or two but once it gets to where the learner should be speaking it stops. It resumes where it left off when the screen is unlocked.
I wondered if it were some kind of buffering issue so I went to the download function but see that that has yet to be enabled. The next best thing was to turn off WiFi and mobile data, which I did but there was still plenty of content all the time the screen was alive.
I used Icelandic and Basque via Spanish and found no differences in behaviour.
Diolch, thatās super super helpful. We definitely have some particular issues with Samsungs and Android, this level of detail will help massively as we keep chipping away at itā¦
Steak, we think this is expected behaviour for Android when the app is in normal mode - to get expected behaviour of continued playing when screen goes off, you need to switch on ādriving modeā in the app, under settings
It would be super interesting to hear if that leads to any changes!
Oh, thatās fabulous news! I was beginning to pull some of my hair out, and I canāt really afford to lose any more
On another happy note, weāve just gone live with Bulgarian - same disclaimers, beta version, unproofed - but weāve got volunteer editing coming very soon now In the meantime, it should be usable and valuable
OK, reproducibility of errors is always a headache: my partner works with IT and has a line about x, y, or z only happening āif you try it on a Wednesday while standing on one leg and whistling Dixie,ā but I think there may be progress with the Greek. Iāve just done about half an hour or so, the last 15 minutes without input from me at all (screen off, hands off) without any gabbling.
I donāt know if this is because Tomās managed to fix things, or because I tried something different, or a bit of both, but this is what worked:
Android 14 OneUI 6.1
Chrome 146.0.7680.164
SSi build b6abc9a
Greek thru English Orange belt
In the browser (not the PWA) with Bluetooth speaker and Drive mode on.
What didnāt work when last tried:
PWA (installed from either Opera or the Samsung built-in browser, though, rather than Chrome) - not sure of the version bc itās now same as above, but it wasnāt when I tried it.
So it may just be a case of Browser Wars 2: this time itās mobile!
⦠Aaanyway: very very happy with the Greek, by and large - when I was first learning Welsh I kept thinking that I loved the method so much, I wished every language I wanted to learn could be taught the same way, and we really are inching closer to that - but I have spotted and managed to document the error I mentioned above. āHow to speakā seems to be translated as the 1sg with Ī»ĪĻ where it should be 2sg Ī»ĪµĻ (slightly irregular, hasnāt been taught AFAIAA) or 3sg Ī»Īει, and thereās a similar issue with the sentence γιαĻĪÆ Ī¼Ī±ĪøĪ±ĪÆĪ½ĪµĪ¹Ļ ĻĻĻ Ī½Ī± Ī¼Ī¹Ī»Ī¬Ļ Īλληνικά; (Why are you learning how for me to speak Greek?), which should have 2sg Ī¼Ī¹Ī»Ī¬Ļ (You speak).
Thereās also one slight issue with the English: if youāre not looking at the screen, alternate translations are read confusingly without pause, so that āher / of herā is pronounced āherofherā, etc.
Oookay, further to the above: exact same setup as āwhat workedā above, except that I forgot to put it in Drive mode, and it started falling over itself again within less than 10 minutes; enabled Drive mode & it behaved beautifully. Thatās definitely a clue of some sort!
The Romanian course is great! Iām almost finished with Orange belt.
One issue that started towards the middle of Orange belt is that every time a new seed is introduced, it goes through all of the drilling, and then at the end of that section, there are about a dozen sentences (the same ones every time) that, after the prompt, are spoken at double speed (or even faster). Once it goes through those sentences, the next seed is introduced and everything is normal until it starts with those sentences again.
This sounds similar to what I have experienced when the Greek is otherwise behaving itself - or possibly identical (I wouldnāt describe it quite like that, but we may be picking out different aspects of the same phenomenon). I find on my phone that the last few sentences of the spaced repetition bit immediately before the new seed have several that are given identically twice in succession, and that seem to be delivered a bit faster, and with almost no pause in which to speak before the first voice cuts in. I had taken it to be a deliberate attempt to put me under a bit of extra time pressure, but perhaps not. I havenāt experimented to see if the same happens on my netbook.
ETA - I said almost no pause, but I do find that if Iām really quick off the mark I can often just about get in there before the first voice. I suspect that what Iām noticing (so, possibly, not the same thing youāve found) is deliberate pressure, although itās possible it might want slight tweaking.