There is a lot of great software on here, but screenshots or videos would be helpful to see how it works or what it does.

Hello Chris,

I enjoyed the Bible Karaoke session today and thought I'd follow-up my question here if that's okay. It seems like an amazing tool for literacy learning IF the speed could be slowed down. Today, there were lots of solutions about slowing down the speed as is done on youtube. However, that solution would slow down the pronunciation of words which is not ideal.

Is there any way that the speed could be reduced by finding the silences between words and then increasing those? That would be a solution that would give the same speed for reading the word but slow the sentence speed down. Anyway, just an idea. This project has great potential for helping beginning readers - especially older people who think that they can't learn to read.

Thanks,
J

I tried out our recordings of Colossians, which HearThis numbers 0 (for the introductory material that is not part of the Bible), 1 (for chapter 1), 2, 3, and 4 (for chapters 2, 3, 4). Bible Karaoke numbers them 1-5 and no 0. I anticipate this being confusing for users. They might want to make a video of chapter 3, so they choose "3" but then they end up with a video that just contains chapter 2, and can't figure out why. If there would be a work-around that would allow what HearThis calls "0" to be labelled as "Intro" or "Introduction" or even "0" and the other chapters to be numbered 1-4, that would be really helpful.

This happened on JB's computer - I could not repro on my computer (worked fine). We are both Windows 10 and latest HearThis and Bible Karaoke.

Behavior Changed: Sending non-JS objects over IPC now throws an exception

In Electron 8.0, IPC was changed to use the Structured Clone Algorithm, bringing significant performance improvements. To help ease the transition, the old IPC serialization algorithm was kept and used for some objects that aren't serializable with Structured Clone. In particular, DOM objects (e.g. Element, Location and DOMMatrix), Node.js objects backed by C++ classes (e.g. process.env, some members of Stream), and Electron objects backed by C++ classes (e.g. WebContents, BrowserWindow and WebFrame) are not serializable with Structured Clone. Whenever the old algorithm was invoked, a deprecation warning was printed.

In Electron 9.0, the old serialization algorithm has been removed, and sending such non-serializable objects will now throw an "object could not be cloned" error.

This affects us, as we see the error "object could not be cloned" when trying to start the renderer, which uses IPC. We can use Electron 8.0 as a transition but to go beyond 8 requires the code changes recommended above.

• scrolling text
• subtitle (bottom 25%)

One of the nice features of BK is that it just knows where your HearThis data is (assuming HearThis was installed in the default folder). In #74, if a user accidentally (or even intentionally) changes the directory location it would be helpful if they could set it back to the OS default location.

To comply with google analytics policy.

This was reproduced multiple times on a Windows 7 computer. Here's what I tried, but none of these steps got past the error:

• closing HearThis (I thought this would work since I encountered a similar error like this on another machine and resolved it by closing HearThis)
• restarting Windows 7
• disabling Panda Dome antivirus and Panda USB

I copied the HearThis project folder to another computer and it didn't have this issue (but failed on a different step)

To make the color picker for the background (and font and speech bubble) even more useful for the average user who does not know how to look up number codes for colors or even that code numbers for colors exist, I suggest:

(a.) Add a few more color options to the current ones (my team would like some in the browns/natural colors range if possible)
or
(b.) Add a way to access a color palette selector that already exists in Windows (I seem to notice a number of programs pull from a similar selector that looks something like this:

or, even more options:

(screenshots from LibreOffice)

The user needs to be able to control the speed of the highlight animation, for example how long it stays on the word. The general feedback from users is that it is too "fast" or "chunky". The highlight needs to stay on the word longer and fade away smoother I believe.

What sort of analytics should we pursue?

Google Analytics for Electron desktop app tutorial: https://kilianvalkhof.com/2018/apps/using-google-analytics-to-gather-usage-statistics-in-electron/

In the Settings drawer, add a link to our license on GitHub: https://github.com/sillsdev/bible-karaoke/blob/master/LICENSE.md, e.g. "This software is released under the MIT License" and link to "MIT License". And add the text "This software uses libraries from the FFmpeg project under the LGPLv2.1".

Allow user to opt out of analytics.

This is not a major issue, just a convenience thing. I am wondering if there is any way that the program could remember which font you used last time you made a video so you don't have to go through the list and find it again.

It would also be nice for it to remember your other design selections, too, come to think of it.

Include some sample data that can be committed to the repo

From L, our first customer:

"As for our team, I feel like the software is not quite to the point of being usable without the section headings being bold (which is something I can't easily fix myself) "

In the user interface, it would be helpful if the area in which you can click around the words "1. HearThis" and "2. Display" would span the entire length of the window area (up to or even including the check mark that appears there after proceeding to "Next".

Note: this issue is MUCH lower priority than getting bold section headings and adjusting the animation, and for that reason I hesitated to even add it as an issue.

I have seen an instance where the render fails (error trying to read audio length) if HearThis is open at the same time as Bible Karaoke. We should provide a better error message like "Please close HearThis before continuing". The message in question is in #10

I think this is a bug report.
When I tried to make a video of Ephesians Intro in BK 0.3.2, I had first made a video of Ephesians 1, then went back and chose "Intro" and made another video. The resulting video had the audio for the Intro, but the text that appeared was the text of Ephesians 1 still.
Closing the program, re-opening it, and then making the video again solved the problem. However, it would be more convenient not to have to do that between every video.

I am getting different text in my video than what I see in HearThis. The text in the BK video is older, outdated, and contains lots of spelling errors that have since been fixed in Paratext, which is where HearThis is pulling from.

The node libraries use python

Auto-select the dropdown in the case that only one option is available. This should apply to the following dropdowns: project, book, chapter.

Bible Karaoke should detect the presence of Scripture App Builder (SAB) projects on the computer and offer to render a video for those projects present that have audio and timing files.

This is an extremely minor issue, and possibly a matter of opinion.
I noticed that the checkmarks on the right side of the UI seem to appear "checked" (turn green) a little too early for Step 2. For example, in the screen shot below you can see I finished step 1, and am working on step 2, but before I have even done anything in step 2, the check mark has already turned green.
I would expect it to turn green only after I push the "Next" button. But perhaps the designer intended it to mean "you are allowed to go on to the next step if you wish"?
Anyway, there is my view/understanding as a user.
Thanks!

...rather than image backgrounds.

User could select from a color palette in the UI and the rendering code would use that instead.

Especially useful if the background color is set (#1)

If the source audio has any silent audio at the beginning or ending of an audio clip it affects the word highlighting.

I have an ffmpeg command that will trim audio below a specified dB level at the beginning and ending of an audio clip. My local tests proved to help the timings so I will look next into adding it to the audio timings process.

#74 changes to using mobx-persist for persisting data. It would be a better UX if we migrated AppState data from the old store if it exists.

When I make the Ephesians 1 video, there is no chapter announcement in the audio, and so the highlighting animation is all off by a whole paragraph/sentence from the get-go.
For Colossians, there is a chapter announcement, so there is no problem.

The animation shows a single word highlight right now...it needs to show all previous highlighted words now to reflect the new animation style.

When I try to generate a video, I get this error:

Could not find Chrome

Error: Could not find Chrome at module.exports.record (C:\Users\Samsung\AppData\Local\Programs\bible-karaoke\resources\app.asar\build\cli\lib\rendering\record-frames.js:10:15) at processTicksAndRejections (internal/process/task_queues.js:85:5) at async render (C:\Users\Samsung\AppData\Local\Programs\bible-karaoke\resources\app.asar\build\cli\lib\rendering\render-frames.js:45:5)

Does that mean Chrome is a prerequisite for Bible Karaoke? I don't have it installed, but use it as a portable app from a Veracrypt hidden container, for security reasons.

W

I think this was determined to be an issue that needed to be dealt with by the HearThis team, technically, but I am putting it here because it is something that needs to be dealt with in order to make the UI ready for use by the average computer user. And in case there is a possibility that it could be dealt with by BK instead.
The word "chapter" is pulled from somewhere within HearThis. However, the speaker of the local language announces this in their own language. It isn't a problem in SAB because there is no word "Chapter" that appears, but in the video that is output from Bible Karaoke, "Chapter {1}" appears in English.
I like the fact that the text appears at all, we just need to figure out how to make it easier for it to appear in the local language.

Need to upload a LICENSE.md with a copy of the MIT license

Perhaps we could change the UI from using dropdowns to using a more complex selector (range slider or something) to select more than one chapter at a time.

The current code uses a run-time downloader to download missing dependences. This is great in theory, but we need to guarantee that the program can run offline once it is installed. The installer should include ffmpeg and ffprobe dependencies from the get-go.

See #11 and #16

Node errors and FFMPEG output / errors should be saved to a running log. Also debug messages in the program code should be output to a log file on the user's computer. This could be used as a postmortem identify what's going on.

From Lisa:

After the user clicks "Make My Video," I think it would be helpful to have a message that helps the user understand approximately how long it will take. I know this is variable depending on the hardware on the person's machine, but for local, not so computer-savvy users, they need some indication that it will take more than 3 seconds. Maybe something like, "Please wait, this will take 5-10 minutes, depending on your computer" (or whatever you estimate it to take).

In the Ephesians video output from 0.3.3, I initially tried chapter 5 but got an error, so I switched to chapter 3. The resulting video for chapter 3 contains the section heading and the first verse from chapter 5, then continues on with chapter 3. The audio only is chapter 3 audio.

Kudos -- The karaoke highlighting animation does a pretty good job of figuring out within a few words that the initial text is not right, and skips down to the correct text.

I realize this is probably a more difficult request, but the translators pointed this out when they saw a sample video I created, and I think bringing the accuracy in this area up would have a really good payoff.

The highlighting lagging behind the audio is a lot more noticeable than audio lagging behind highlighting animation.

It is not the end of every sentence, so I can try to do a little more observation to see if I can identify a pattern. I do know that for most languages in the world, this one included, the ends of most sentences will have lower relative pitch, lower volume, and slower rate of speech. So those might be someplace to start (since I don't know what kind of algorithms you're using on your end).