đź—Ł Sentence Audio and misc updates

Hey guys, wanted to give the heads up on the new audio – looking forward to feedback about it :slight_smile:

Sentence Audio

We went through and added audio for all of the sentences.
You’ll see an audio symbol next to it which you can click to hear.

image

Of course, this will also be seen in the Lessons, or under the Info Panel in Reviews:

You can set your voice preference for the sentences in your settings:

We plan to roll out new text-to-speech audio options for the application itself in the near future as we continue to evaluate the best TTS options out there that can fit our rather unique needs.

Respect Character Unlocking

When we rolled out the new lessons flow, we previously didn’t respect the character unlocking preferences.

We now do: when you decide to learn a character that has prerequisites, the character will be delayed until you are familiar with those prerequisites:

character-unlock

Block the “other character” as an answer for words

Some words in HanziHero inherit one of the meanings ascribed to the characters within it. For example, 設置 set up which has the characters 設 establish and 置 set up within it.

It can be frustrating to know both the characters but you aren’t precisely sure what the meaning is, and the alternate meanings haven’t covered it.

So, if you try to input the “other character meaning” in these cases, it’ll be blocked:

word-block

Extra Study small changes

The previous Extra Study date ranges were inconsistent for both Recent Mistakes & Recently Learned.

We’ve since changed this to always look at the last three days, inclusive.

The docs have been reflected to show this.

Word Skip Preferences

We added an additional setting which (by default) skips over words which contain only skipped characters.

This makes it so every word you’re learning is linked to a character you’re learning :slight_smile: – this especially helps in the case of those who decided to take advantage of the HSK skipping.

Turn off Lessons

If you find yourself with too many reviews, and yet the Lessons button is taunting you to do more, I’ve since added an option for “None” – giving you the satisfactory greyed out Lessons while you can focus on getting your Reviews down :sunglasses:


That’s it for now! We have a lot of interesting suggestions to work through in the backlog. As always, thanks for your support and feedback, HanziHero wouldn’t be where it is today without you :wink:

Happy studying!

9 Likes

We’ve noticed some strange behavior for the new audio for the simplified course. Looking into it now.

1 Like

Our new text-to-speech provider for the sentence audio is working just fine. However, for the small snippets of audio we have for sounds, characters, and words, it seems to have some buggy behavior. Mainly for the simplified course.

To solve this, we are going to make it so the “Preferred pronunciation voice” only works for sentence audio. It will be renamed to “Preferred sentence pronunciation voice” for the time being. Additionally, we are going to roll back to use the old female TTS we have for both the simplified and traditional course. We’ll look into adding out a male TTS speech for the application (separate from the sentence audio) in the near future.

I’ll post again once these quick changes are rolled out. Should be done in the next hour or so.

2 Likes

I’m not sure that I like the new audio better. Maybe I am just very used to the previous one and need some time to adapt. I dont’ know if the “strange behavior” is regarding the tones, but I noticed that the pronunciation of 丫 sounds a lot like ya4 instead of ya1.

Regarding the extra study and recently learned new features I’m also not so sure that it is the best for my specific study routine, as I like to review both twice a day. Once before I do the regular reviews in the morning and also before I go to sleep. With this new set up maybe it will start to few like too much.

1 Like

One thing we plan to add in the near future for both of those is an additional “filtering” modal before launching the study session.

One would click e.g., “Review recent mistakes” which would bring up a modal which contains a list of the recent mistakes. One could then uncheck the ones they don’t want to include in the review session before clicking “Confirm”.

Would that address your concerns?

Separately, that change was rolled out a week ago or so. It’s mainly a bugfix to bring Extra Study in-line with the rest of the application, so if you haven’t noticed any changes in the past week in review load then I would not worry about it too much.

1 Like

We’ve completed rolling out the change to go back to the old pronunciation audio for most of the application. The new text-to-speech engine will only be used for sentence audio for the time being. You may need to refresh your browser page to get the changes. :construction_worker_man: :wrench:

3 Likes

Yes, this would be a good feature to avoid overload. I was taking a break from new lessons so I didn’t notice the difference in the recently learned, but I did notice that my recent mistakes section was unusually bloated. I just thought that I was doing an exceptionally bad job at reviewing lately. :rofl:

I was growing fonder of the new voice, it definitely had some good aspects to it.

2 Likes

It would be nice to have an option for sentence audio speed as well. :slight_smile: Right now, it just blurs together for me!

4 Likes

So yeah, I am pretty sure the sentence audio is a little bugged atm and is faster that it’s supposed to be. You might already be aware of this, but I recorded what it sounds like on my end atm:

Hope this is useful!

3 Likes

They sound too fast with my devices too. I had been copying them to google translate. It sounds much better there. :grinning:

1 Like

I have nothing of value to add to this thread, but I felt compelling to say that I listened to the track, interested to hear the fast talking. I was merely expecting the sentence audio … instead I was swept away by the smoothest vocals I’ve heard outside of Stephen Fry, accompanied by a soft jazz backing track. I’ve heard songs on the radio with less production value than your bug report :stuck_out_tongue:

3 Likes

Thanks for the feedback! Here is the same sentence with a lowered prosody speed. How does it sound?

This particular phrase sounds very similar speedwise as my google translate version. I timed it @2.37 vs 2.43 . :grinning:

1 Like

I think what we will probably do as a first step is add a setting for sentence playback speed with settings of “normal”, “slow”, and “extra slow” (the one I have above is “extra slow”). We’ll look into this this week. :+1:

Honestly this is pretty much normal speed for native speakers in many situations. A slower option for beginners, as it is being considered, might be a good solution.

That’s easier to understand, sure.

The one other piece of software I currently use, Pleco, has a voice speed setting that goes in 10% increments from 10% to 200%. I’m not sure this much granularity is necessary, but I think explicitly giving the users the speed percentage might be useful.

1 Like

Thank you, that’s very kind!

You can listen to full 2 hour version on my ASMR channel, Almost As Smooth As Stephen Fry ASMR. :wink:

2 Likes

Nice podcast by the way

Hey guys! We just added this setting. You can adjust the playback speed underneath the “Audio Preferences” with three options:

  • Normal
  • Slow
  • Extra slow

Let us know what you think :slight_smile:

3 Likes