Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

WIN 10 64 bit - App starts process but slow to presents GUI #44

Open
motioncircus opened this issue May 17, 2020 · 77 comments
Open

WIN 10 64 bit - App starts process but slow to presents GUI #44

motioncircus opened this issue May 17, 2020 · 77 comments
Assignees
Labels

Comments

@motioncircus
Copy link

Downloaded and installed autoEdit-3-1.2.7.exe

Ran the exe (with Admin Rights), which started a Digital Paper Edit app (32 bit) process in Task Manager

To Reproduce
Steps to reproduce the behavior:

  1. On a 64 bit WIN 10 Pro N machine, download the latest Pre-Release version
  2. Run the EXE
  3. Check Task Manager to see process Digital Paper Edit app (32 bit)
  4. Notice there's no application presenting

Expected behavior
The App GUI to appear

Desktop (please complete the following information):

  • OS: WIN 10 Pro N - 64 bit - Version 10.0.18363 - Build 18363
  • Chrome -
  • Version Version 81.0.4044.138 (Official Build) (64-bit)
@motioncircus motioncircus added the bug Something isn't working label May 17, 2020
@motioncircus
Copy link
Author

After writing this post, I find the App is now open... Perhaps it's just a LOT slower than I expect.
I will report back if this is an intermittent issue
Cheers

@pietrop
Copy link
Owner

pietrop commented May 17, 2020

Thanks for flagging this @motioncircus ! And thanks for the update

Interesting, it should probl be the 64bit version of the app rather then the 32bit.

based on electron builder docs it should be xcode10.2 on Travis, while in the current setup I have xcode 11 But not sure if this would effect the 32bit vs 64 bit.

And/or based on same docs, could add --x64 flag to make the arch explicit, in the npm script npm run build:mwl:publish:always

Following up on my comment on a related issue #43 (comment) in v 1.2.7 I removed the electron remote module for the most part, but I suspect that there’s also some re architecting to do, to speed up performance. the electron “backend” at the moment is connect to the react client via a global window variable which might be effective performance. And need to figure out a better way for the two to communicate, while maintaining a modular architecture

some other clues in the electron docs about performance as well.

But back to the windows issue, since I don’t have a windows machine, would you be able to try and package/build locally with the 64bit arc flag and see if that improve startup time?

It would be a matter of following these instruction in the README#setup

to get setup

And modifying this line in package.json

- "build:w": "cross-env NODE_ENV=production electron-builder -w",
+"build:w": "cross-env NODE_ENV=production electron-builder -w --x64"

Before running

npm run build:w

@motioncircus
Copy link
Author

motioncircus commented May 17, 2020 via email

@pietrop
Copy link
Owner

pietrop commented May 17, 2020

Ok no worries, there should be another release coming up, in the release section 1.2.8 Where I added the 64 bit flag. 🤞
( as of now it’s still building, check back in a bit)

Re npm, never mind those instruction, is probably a bit more involved, you’d need to install node ( the programming language) and npm is the package manager (works cross platform) but we might be able to test these changes without going through all that trouble.

And yeah it be good to get some feedback when you get to reconnecting the media in DaVinci Resolve.

One thing tho, is that autoEdit (2 and 3) works best if you import the original source (eg the media from the cards, or the source media you want to work with in the edit, rather then proxies if that makes sense), and let it create the audio and video preview for the app and the STT engine from it (uses ffmpeg under the hood), so that it reads the metadata from the original source footage and uses that info to create the sequence so that it can reconnect to it. ( this is mostly the EDL option)

@motioncircus
Copy link
Author

motioncircus commented May 17, 2020 via email

@motioncircus
Copy link
Author

motioncircus commented May 17, 2020 via email

@motioncircus
Copy link
Author

motioncircus commented May 17, 2020 via email

@motioncircus
Copy link
Author

motioncircus commented May 17, 2020 via email

@pietrop
Copy link
Owner

pietrop commented May 17, 2020

Not sure re virus warning, probl not supposed to happen. they are packaged, built and shipped to the release section via Travis CI, a third party, with virtual machine for OSX.

Building another release 1.2.9, where tried to tweak the settings to see if it makes it for arch 64. 🤞

I'm wondering about your apparent suggestion that I could use the Autoedit
previews in place of proxies.. I'd love to be able to do that but doesn't
that involve uploading the hundreds of GIGS of my 4K camera files to
AssemblyAI or similar?

what autoEdit does under the hood, is that it creates an audio preview / version of your media, (that is very fast to do even on large files) and then only sends that to the speech to text engine, eg assemblyAI.

I haven't found a proxy settings dialogue and, when I considered trying
such an approach, couldn't even locate the previews on my system.

It also creates a lower res mp4 video preview, To use within the autoEdit programme to show you the preview. If that takes to long and the transcript is ready you’d be able to interact with it via the audio preview while the video preview is being created.

These audio and video previews are generally used only by autoEdit. I could make them available to “export”/“save” externally to the programme if useful tho.

If you have proxies made via the editing software, as I guess in a 4K workflow, you might not edit directly on those rushes (correct?), I’d be curious to look at the steps of that workflow in more details to see what might be an optimal way to integrate it with autoEdit, once we figure out the 32 vs 64 bit issue.

@motioncircus
Copy link
Author

motioncircus commented May 17, 2020 via email

@pietrop
Copy link
Owner

pietrop commented May 17, 2020

Also when you get a chance could you check if autoEdit 2 windows version was also in 32bit?

https://opennewslabs.github.io/autoEdit_2/

Releases page
https://github.com/OpenNewsLabs/autoEdit_2/releases

@motioncircus
Copy link
Author

motioncircus commented May 17, 2020 via email

@motioncircus
Copy link
Author

motioncircus commented May 17, 2020 via email

@motioncircus
Copy link
Author

motioncircus commented May 17, 2020 via email

@pietrop
Copy link
Owner

pietrop commented May 17, 2020

Thanks for looking into it @motioncircus !

download size

Download size is probably double because it is now trying to do 32bit and 64bit in the installer.

virus warning

Do you still get a virus warning with latest version of autoEdit?

4K tests

The tests you did was with autoEdit 2 or 3?

What would be different in a 4K file?

The other thing to try is make a short sequence in Resolve with a few 4K clips, and export an EDL from Resolve to compare. (You could copy paste the content of the file here, might be easier then sending an attachment via GitHub)

screenshots

Seems like you might be replying to the GitHub email notifications with the screenshots, and the images are not coming through.

If you visit
#44 you should be able to add them here directly.

@motioncircus
Copy link
Author

Hi Pietro,
I've edited my replies above by dropping in the images - they seem to work as links but not directly visible in the post. Perhaps editing is different than direct posting because I've checked this post and the images are visible directly in preview.

virus warning

No, I didn't get the warning with the latest download but I did have Win Defender's 'real time virus monitoring' turned off, which is how I got around the issue yesterday.

4K tests

Were done in autoEdit 3

The Resolve import error I'm getting (on an autoEdit created EDL), says the clips fail to link because the timecode extents do not match any clip in the media pool

Which makes me wonder if it's possible that autoEdit-3 creates and uses timecode in it's previews that are stored in a different way, or a different place, than the timecode in the 4K files?

Capture6

Resolve test EDL export and comparison

I will make a new set of tests, using 4K files directly in autoEdit-3 AND a simple timeline created in Resolve and export the EDLs for comparison.

Back soon
Nigel

@motioncircus
Copy link
Author

Here's a screengrab of my two test EDLs, side by side in Notepad++

Capture11

There are some minor differences, one of which rings a bell from the last time I tried to figure this problem out - the fact that Resolve timelines default start at 01:00:00:00

The other is that Resolve sees the start time of each source clip as 02:xx:xx:xx
where autoEdit sees source start times as 00:xx:xx:xx

I wonder if my rendering out proxy files eliminates this 4K two hour offset?

I hope this helps.

Nigel

@motioncircus
Copy link
Author

OK I've done some research and testing and it's definitely a timecode offset issue because if I hack the EDL output of autoEdit, to add the required offset (and it's not just hours but minutes too) then it works imported into Resolve.

It seems my camera, a Panasonic GH4, is a bit flakey with timecodes:

This is a quote from the manual.

  • Time codes are not recorded on Motion Pictures recorded when the [Rec Format] is set to [MP4].

And of course I've been recording in MP4.

The timecodes on my generated proxies start at 00:00:00:00, hence them working fine when the edl is exported from autoEdit.

I might make a test using my GH4 with MOV format, after I check why I was using MP4, it might have been.

If you can think of a better workaround I'd love to hear it.
Nigel

@pietrop
Copy link
Owner

pietrop commented May 18, 2020

That’s great, thanks!

4K EDL

Seems like it’s also expecting the audio to be on a separate EDL track?

I had a similar problem with .mxf files. But those where easier to recognize, because of the file extension, and if there’s an MXF in one of the clips autoEdit creates 3 EDL Tracks, one for video and two for audio.

file metadata

If the time codes are not recorded then in theory it would threat it as zero in autoEdit, but you said you added an offset manually?

autoEdit (both 2 and 3) reads the file metadata to see if it needs to compute a time code offset when creating the sequence. I’d be curious to see what are the metadata it reads for that 4K mp4.

The easier way to do this, is If you go in autoEdit2, if you have imported a 4K file in there already (?), and in the transcript view, under export option you choose json. You can then share the content of that json file here; or if you are familiar with json, open that with notepad++ 📝 and only share values of the metadata attribute. If that makes sense?

Otherwise there’s also a way to do it in autoEdit3 but is a little more complicated to explain. But can do so if the autoEdit2 option is not feasible.

EDL format

These are some notes I made a while when working on the module to create the EDL programmatically about how the EDL format works
https://autoedit.gitbook.io/documentation/appendix/edl-format

@pietrop
Copy link
Owner

pietrop commented May 18, 2020

But yeah maybe test with mov first and see if that solves the problem before going down the metadata time codes rabbit hole 🐇 🕳

@motioncircus
Copy link
Author

OK,

Yes there was an offset which I edited, so perhaps the GH4 manual is wrong, or a firmware update changed the timecode to allow it to be generated on MP4s.

Here's the (full) json export ((both flavours of json), from one of my original (MP4) 4K files imported into autoEdit3

4k_SOURCE_jsonAutoEditTest.zip

FYI
Version 2.1.9 takes a full 4.5 minutes to start after double clicking the exe file - longest program load I've had for twenty years or so and very suspicious.

@pietrop
Copy link
Owner

pietrop commented May 18, 2020

Sorry, unfortunately in autoEdit 3, the transcript json doesn't return the metadata, unlike autoEdit 2.

But here's how to do it in autoEdit 3

  1. open autoEdit 3
  2. go to the transcript view
  3. click at the same time cmd+ option+i (not 100% sure keyboard combination on windows - maybe windows key + alt + i) OR in the menu under view --> Toggle Developer Tools
  4. This will show some dev tools. at the top, choose console if it's not already selected.
  5. in the console, you should see a json object that can be expanded, and you should be able to see the metadata info - see example below

dev tools

@motioncircus
Copy link
Author

file metadata
If the time codes are not recorded then in theory it would threat it as zero in autoEdit, but you said you added an offset manually?

That's right.. despite what the manual says, you can see a 2 hour odd offset (in the right hand side of the Notepad++ screengrab above)

I've just tested an MOV but the problem remains see below:

Annotation 2020-05-18 130050

It would appear that autoEdit isn't able to read the offset and add it to it's EDL output.

As you noticed, there is something different in the audio track creation too. Even when I bring in an EDL of a correctly adjusted timecode version, I can't see the audio in resolve until I manually create a new timeline and then switch back to my imported one.

Thanks for your help. I'm really hoping to be able to use this and I feel so close

N

@pietrop
Copy link
Owner

pietrop commented May 18, 2020

and, yeah re long startup time, might have to rollback to an earlier version

@pietrop
Copy link
Owner

pietrop commented May 18, 2020

and yeah see previous comment #44 (comment) if you could see what metadata are actually being read by autoEdit 3, it would help narrow it down 🤞

@motioncircus
Copy link
Author

I'm trying to follow your awesome guide GIF but my console seems a little different.. can't quite locate the json object

@pietrop
Copy link
Owner

pietrop commented May 18, 2020

Ok you can share a screenshot of where you got so far

@motioncircus
Copy link
Author

A loong slooow restart fixed the view

Annotation 2020-05-18 132148

@pietrop
Copy link
Owner

pietrop commented Jun 5, 2020

Question, did you have the same problem (slow startup time and 32bit instead of 64) in autoEdit2 windows version as well?

@sqr
Copy link

sqr commented Jun 5, 2020

AutoEdit2 loads fine (takes maybe around 10 seconds) and shows up in the Task Manager as autoEdit2. It also spawns a background process called Fast text based video editing app (32 bit).

@pietrop
Copy link
Owner

pietrop commented Jun 5, 2020

I wonder if this has something to do with it?

https://www.electronjs.org/docs/tutorial/windows-arm

And this thread

eugeneware/ffmpeg-static#35 (comment)

Or the signing of the app (altho autoEdit2 is not signed)
https://www.electronjs.org/docs/tutorial/code-signing

@sqr
Copy link

sqr commented Jun 5, 2020

Hello Pietro, after some more tests I think the problem might be that, as the app is built in portable mode, it has to unzip all the files within the exe into a folder in C:\Users\userAppData\Local\Temp .
I am pretty sure that this is what takes such a long time, as each folder pertaining to the app in my temp folder has around ~60k files and is several GBs in size. Even deleting it takes a while. This would also explain why the packed app takes so long to load but the unpacked version in the dist folder runs instantly.

I have tried to build the app in a non-portable way but haven't found a way to make it work. I'm not familiar with electron-builder. I just changed the "portable" setting in the package.json to "target": "nsis". The packaging works and I get a .exe in my dist folder, however it says it's missing an autoEdit3.exe file once I run it.

I have tried to troubleshoot this issue but haven't found a solution.

I have read the links you provided but can't figure out if they have a connection with the long loading times.

@pietrop pietrop changed the title WIN 10 64 bit - App starts process but slow to presents GUI (+ 4K EDL issue for GH4 camera) WIN 10 64 bit - App starts process but slow to presents GUI Jun 9, 2020
@Klendat
Copy link

Klendat commented Jan 3, 2021

This seems to still be an issue. Is there any roadmap for fixing it?

@pietrop
Copy link
Owner

pietrop commented Jan 8, 2021

👋 I don't have a windows machine and therefore not able to properly look into this and troubleshoot. If there's interest interest in looking into it, I can assist with the effort. Let me know.

@digital-pet
Copy link

digital-pet commented Jan 8, 2021 via email

@pietrop
Copy link
Owner

pietrop commented Jan 8, 2021

If you got the time to look into this you are more than welcome to dive in. And I can assist when needed.

@digital-pet
Copy link

digital-pet commented Jan 8, 2021 via email

@motioncircus
Copy link
Author

motioncircus commented Jan 8, 2021 via email

@pietrop
Copy link
Owner

pietrop commented Jan 12, 2021

one thing worth trying, if anyone has the capacity, is to package/build the electron app with asar packaging enabled. From this comment electron-userland/electron-builder#428 (comment)

@nomatica
Copy link

nomatica commented Apr 5, 2021

So over the last week I began experimenting with Autoedit3. I am experiencing some issues expressed in this thread.

I am attempting to use Autoedit3 on a win10 64bit system.
Dowloaded ver 3-1.4.0

AutoEdit3 always takes a number of minutes to load. After loading, when attempting to perform a transcription, it is unpredictable. I have successfully performed one or two transcription, but most attempts result with the the progress arrow looping endlessly. I am using mozilla’s deepspeech. Autoedit3 confirms I have downloaded it and I am ready to go. When testing, I have limited myself to videos less that 15 minutes.

Any advice on troubleshooting would be appreciated. The file that was successful, showed just how amazing the application is, and I am hoping I can utilize it going forward.
The image below shows the progress arrow I mentioned.
autoedit issue

@pietrop
Copy link
Owner

pietrop commented Apr 7, 2021

Thanks for raising this @nomatica unfortunately I think this might be an issue with the underlying Mozilla Deepspeech module. Could you try with the free trial (5 hours a months) of assemblyAI and let me know if you encounter the same problem?

To add to that, Mozilla has discontinued support for deepspeech, But the good news is that there's a new initiative born out of that endeavor, called coqui (see on github.com/coqui-ai/STT). Which could be a worthwhile replacement 🤞 . I need to find the time (PRs also welcome) to look into whether they have a node moddule, and if it's ready to use, to see how to best integrate it into autoEdit3 (electron version).

@nomatica
Copy link

nomatica commented Apr 8, 2021

Thank you. That is good news about an alternative for deepspeech. Having a local option is important to me.

I can try AssemblyAI to trouble shoot. I will try and get it set up over the weekend. I have been able to to succeed using Mozilla DeepSpeech with 4 additional files. I then tried again with different files, and it failed in a same way as it had before. This lead me to believe that it might be something on my Windows 10 system. So I attempted to loaded autoEdit3 on another windows 10 system. However, AutoEdit3 did not seem to load fully. It loaded the menu bar [Application, Edit View… see attached photo]. I tried the “force reload” and it did not work. When I try loading the “speech to Text settings’ It comes up blank as well.

Do you have any idea what is going on there?
Thank you again
autoedit3

@pietrop
Copy link
Owner

pietrop commented Apr 9, 2021

Looking at Coqui STT I realized their npm module is not ready for use coqui-ai/STT#1830 😞

But looking at their docs, it seems to have same interface as deepspeech for now, and I think I might have found the issue but not 100% sure 🤞 💯 . Looking into it in this PR pietrop/deepspeech-node-wrapper#8

These lines might be needed to free up the memory after using the metadata and model option in the deepspeech node modules.

+ DeepSpeech.FreeMetadata(result);
+ DeepSpeech.FreeModel(model);

Will try and clean this up to make another npm release, and add it to a new release of the electron version of autoEdit3, then it be easier to test if it improves the situation with deepspeech 🤞 .


As for the other issue, of the blank screen, I am not sure what's going on there. What happens if you open the inspector, and look at the console? does it say it is disconnected or does it show some error messages?

@pietrop
Copy link
Owner

pietrop commented Apr 13, 2021

@nomatica could you try the latest version? https://github.com/pietrop/digital-paper-edit-electron/releases/tag/1.5.0

@nomatica
Copy link

I downloaded it the new version. I had to re-download the Mozilla DeepSheech files (1.8 gigs) I am curious what directory this downloads to and if there is a way to direct Autoedit3 to the already downloaded files?

Autoedit3 is currently running. It is working on an hour long audio file. I will report back.
Thank you for the work you put into this!
mozilla deep speech saved
Mozilla DeepSpeech

@pietrop
Copy link
Owner

pietrop commented Apr 14, 2021

Great @nomatica , let me know how it goes 🤞

Yeah it prompts to re-download because I've updated the underlying node module that wraps around deepspeech to use the latest version of deepspeech 0.9.3 pietrop/deepspeech-node-wrapper#8 (before it was using 0.6.1)

This is a stop gap waiting for Coqui STT to take over from DeepSpeech and replace it with that when it's ready coqui-ai/STT#1830

The folder the app stores user's data on a Mac, is at

/Users/${YOUR_HOME_FOLDER_USER_NAME}/Library/Application\ Support/digital-paper-edit-electron

There's an equivalent electron path for Windows and Linux (see electron docs).

In there the new Deepspeech model (0.9.3) is downloaded in models folder, while the old one is in the folder deepspeech-0.6.0-models (which could be be deleted)

This is due to inconsistency in the way the deepspeech releases changed the naming conventions of their releases 🤷‍♂️

@nomatica
Copy link

Ok thanks. I will leave the files for now, in case I want to try and compare the older version. It is still working almost 40 minutes in.

I do not know much about coqui.ai other than that they are new. Also not sure what is needed in order for it to be integrated.

@pietrop
Copy link
Owner

pietrop commented Apr 14, 2021

It should take roughly the length of the media, so if it's 1 hour, ~ 1 hour at least.
If it works well, in the future we could add some optimizations. Eg split the media in chunks (eg 5 minutes) and use the different CPU cores to transcribe some of these in parallel and then re-combine the results when done etc...
I had done something similar with IBM Watson STT for autoEdit2 and it worked quiet well, to create a fast turnaround.

TLDR on the rest, from what I understood, Mozilla is discontinuing DeepSpeech and Coqui is taking over (and re-branding). Coqui node STT should have the same interface as the DeepSpeech node module so after this last update the integration should be an easy module swap 🤞 if that makes sense 🙃

@nomatica
Copy link

A quick update: So the progress arrow was still circling. It had been over an hour so I deleted the file. I tried choosing a file that was 1 minute and it did not work. I am going to try and reboot my system and then try and test again. I will load a shorter file first.

@pietrop
Copy link
Owner

pietrop commented Apr 14, 2021

ok, cool, thanks.

if you are on a mac and do

option + cmd + i

It opens the inspector, and under the console tab you can see if there's any red errors

@nomatica
Copy link

I rebooted and ran a 1 minute audio clip and it was successful. Then I tried the 1hr audio clip again, and left it running over an hour and it was still working when I returned.

So I figured out that on a win 10 pc "option+ cmd +i" is "alt + CTRL+i"

To be honest I am not sure what I am looking at, but I will do my best to share what I think might have some clues.

Here is a screenshot. There are no errors listed.
console

@nomatica
Copy link

nomatica commented Apr 14, 2021

Here is what is displayed in the console. You can see the short audio clip that was successful. However the second clip (the hr one) is a different story. There are a number of "violations"

prompt.js index.html:2 process.versions.node 12.8.1 index.html:3 process.versions.chrome 78.0.3904.130 index.html:4 process.versions.electron 7.1.14 C:\Users\David\AppDa…db\lib\diskdb.js:26 Successfully connected to : C:\Users\David\AppData\Roaming\digital-paper-edit-electron\db 9 index.js:16 In Electron Enviroment index.js:37 Running interval to check for transcripts TranscriptForm.js:148 (form.checkValidity() true index.js:16 In Electron Enviroment TranscriptForm.js:89 formData.get('path') F:\transcription test\Downloaded MP3s\115Earhart Amelia Women.mp3 2 index.js:16 In Electron Enviroment C:\Users\David\AppDa…Wrapper\index.js:22 getDefaultSttAndLanguage Object deepspeech C:\Users\David\AppDa…rapper\index.js:159 about to call ipcRenderer - deepspeech 1 deepspeech C:\Users\David\AppDa…Wrapper\index.js:22 getDefaultSttAndLanguage Object deepspeech C:\Users\David\AppDa…rapper\index.js:161 about to call ipcRenderer - deepspeech 2 2 index.js:16 In Electron Enviroment TranscriptForm.js:115 ApiWrapper.createTranscript-response Object 2 index.js:16 In Electron Enviroment index.js:84 handleSaveItem Object 3 index.js:16 In Electron Enviroment index.js:99 setState C:\Users\David\AppDa…rapper\index.js:153 readMetadataForEDL Object transcriptId e9106ccd2d3b483db9b68664cbcf47f0 C:\Users\David\AppDa…rapper\index.js:140 newTranscriptDataVideo Object transcriptId e9106ccd2d3b483db9b68664cbcf47f0 index.js:37 Running interval to check for transcripts index.js:39 interval: checking transcirpt update index.js:16 In Electron Enviroment index.js:37 Running interval to check for transcripts index.js:39 interval: checking transcirpt update index.js:16 In Electron Enviroment C:\Users\David\AppDa…rapper\index.js:164 asynchronous-reply {"_id":"e9106ccd2d3b483db9b68664cbcf47f0","description":"","id":"e9106ccd2d3b483db9b68664cbcf47f0","path":"F:\\transcription test\\Downloaded MP3s\\115Earhart Amelia Women.mp3","projectId":"dcfd8998336f47fcad9b8bf14eb67044","status":"in-progress","title":"115Earhart Amelia Women.mp3","url":"C:\\Users\\David\\AppData\\Roaming\\digital-paper-edit-electron\\media\\115Earhart Amelia Women.e9106ccd2d3b483db9b68664cbcf47f0.wav","transcript":{"words":[{"start":0.76,"end":0.8,"text":"if"},{"start":0.92,"end":1.08,"text":"mod"},{"start":1.16,"end":1.28,"text":"an"},{"start":1.38,"end":1.68,"text":"werld"},{"start":1.74,"end":1.92,"text":"of"},{"start":2.02,"end":2.32,"text":"sien"},{"start":2.42,"end":2.52,"text":"fil"},{"start":2.58,"end":3,"text":"invent"},{"start":3.1,"end":3.2,"text":"en"},{"start":3.26,"end":3.36,"text":"es"},{"start":3.44,"end":3.54,"text":"af"},{"start":3.62,"end":4.1,"text":"particular"},{"start":4.22,"end":4.6,"text":"interest"},{"start":4.68,"end":4.78,"text":"to"},{"start":4.84,"end":5.14,"text":"women"},{"start":5.9,"end":6,"text":"for"},{"start":6.04,"end":6.14,"text":"the"},{"start":6.22,"end":6.56,"text":"lions"},{"start":6.58,"end":6.66,"text":"te"},{"start":6.74,"end":6.98,"text":"women"},{"start":7.1,"end":7.22,"text":"had"},{"start":7.28,"end":7.44,"text":"been"},{"start":7.52,"end":7.74,"text":"mare"},{"start":7.88,"end":8.24,"text":"fected"},{"start":8.36,"end":8.48,"text":"by"},{"start":8.56,"end":8.7,"text":"its"},{"start":8.8,"end":9.18,"text":"newmol"},{"start":9.3,"end":9.72,"text":"eisen"},{"start":10.2,"end":10.4,"text":"and"},{"start":10.46,"end":10.72,"text":"those"},{"start":10.74,"end":10.84,"text":"of"},{"start":10.88,"end":11.06,"text":"any"},{"start":11.12,"end":11.32,"text":"other"},{"start":11.4,"end":11.78,"text":"groops"},{"start":12.54,"end":13.12,"text":"cofound"},{"start":13.2,"end":13.34,"text":"and"},{"start":13.52,"end":13.84,"text":"toring"},{"start":13.9,"end":14.02,"text":"at"},{"start":14.08,"end":14.2,"text":"had"},{"start":14.26,"end":14.38,"text":"been"},{"start":14.42,"end":14.52,"text":"a"},{"start":14.56,"end":15.16,"text":"complishment"},{"start":15.26,"end":15.36,"text":"in"},{"start":15.4,"end":15.46,"text":"the"},{"start":15.52,"end":15.96,"text":"remotiv"},{"start":16.08,"end":16.46,"text":"tiels"},{"start":16.54,"end":16.68,"text":"af"},{"start":16.78,"end":17.54,"text":"twonresearch"},{"start":18.12,"end":18.24,"text":"it"},{"start":18.28,"end":18.38,"text":"is"},{"start":18.42,"end":18.52,"text":"in"},{"start":18.56,"end":18.66,"text":"the"},{"start":18.74,"end":19.04,"text":"com"},{"start":19.5,"end":19.64,"text":"that"},{"start":19.7,"end":19.8,"text":"the"},{"start":19.86,"end":20.66,"text":"afprications"},{"start":20.7,"end":20.86,"text":"af"},{"start":20.98,"end":21.22,"text":"sin"},{"start":21.32,"end":21.54,"text":"ashi"},{"start":21.62,"end":22.28,"text":"cativement"},{"start":22.68,"end":22.86,"text":"at"},{"start":22.9,"end":23.28,"text":"perhaps"},{"start":23.42,"end":23.52,"text":"the"},{"start":23.64,"end":23.86,"text":"most"},{"start":24.08,"end":24.32,"text":"far"},{"start":24.42,"end":24.76,"text":"reaching"},{"start":25.54,"end":25.7,"text":"and"},{"start":25.86,"end":26.22,"text":"istrue"},{"start":26.38,"end":26.78,"text":"changing"},{"start":26.86,"end":27.42,"text":"conditions"},{"start":27.58,"end":27.96,"text":"there"},{"start":28.4,"end":28.56,"text":"hat"},{"start":28.62,"end":28.88,"text":"women"},{"start":29.02,"end":29.18,"text":"have"},{"start":29.24,"end":29.66,"text":"become"},{"start":29.76,"end":29.86,"text":"the"},{"start":29.9,"end":30.26,"text":"greatest"},{"start":30.34,"end":31.32,"text":"benefitaries"},{"start":31.6,"end":31.74,"text":"and"},{"start":31.78,"end":31.86,"text":"the"},{"start":31.96,"end":32.26,"text":"modern"},{"start":32.5,"end":32.86,"text":"scheme."},{"start":33.94,"end":34.32,"text":"syan"},{"start":34.42,"end":34.58,"text":"pad"},{"start":34.66,"end":35.1,"text":"relysed"},{"start":35.18,"end":35.38,"text":"them"},{"start":35.52,"end":35.68,"text":"from"},{"start":35.76,"end":35.94,"text":"much"},{"start":36,"end":36.1,"text":"of"},{"start":36.14,"end":36.22,"text":"the"},{"start":36.32,"end":36.6,"text":"age"},{"start":36.78,"end":37.02,"text":"old"},{"start":37.12,"end":37.54,"text":"gudgery"},{"start":37.66,"end":38.08,"text":"comnected"},{"start":38.14,"end":38.26,"text":"with"},{"start":38.32,"end":38.4,"text":"ha"},{"start":38.46,"end":38.9,"text":"process"},{"start":38.98,"end":39.06,"text":"of"},{"start":39.1,"end":39.44,"text":"living"},{"start":40.3,"end":40.7,"text":"tanding"},{"start":40.76,"end":40.84,"text":"the"},{"start":40.96,"end":41.24,"text":"pins"},{"start":41.32,"end":41.76,"text":"leaving"},{"start":41.86,"end":42.02,"text":"and"},{"start":42.1,"end":42.36,"text":"crud"},{"start":42.48,"end":42.74,"text":"letor"},{"start":42.84,"end":42.92,"text":"te"},{"start":43,"end":43.82,"text":"menufacturing"},{"start":43.88,"end":44,"text":"e"},{"start":44.04,"end":44.56,"text":"fested"},{"start":45.12,"end":45.26,"text":"i"},{"start":45.4,"end":45.66,"text":"things"},{"start":45.7,"end":45.78,"text":"of"},{"start":45.84,"end":45.96,"text":"the"},{"start":46.06,"end":46.4,"text":"past"},{"start":46.64,"end":46.76,"text":"for"},{"start":46.8,"end":46.88,"text":"an"},{"start":46.94,"end":47.52,"text":"increasing"},{"start":47.64,"end":48.2,"text":"majority"},{"start":49.16,"end":49.28,"text":"the"},{"start":49.4,"end":49.64,"text":"day"},{"start":49.82,"end":50.12,"text":"lane"},{"start":50.24,"end":50.32,"text":"he"},{"start":50.46,"end":50.6,"text":"an"},{"start":50.76,"end":51.06,"text":"power"},{"start":51.2,"end":51.32,"text":"may"},{"start":51.4,"end":51.48,"text":"be"},{"start":51.56,"end":52.04,"text":"obtained"},{"start":52.16,"end":52.3,"text":"by"},{"start":52.4,"end":52.74,"text":"pushing"},{"start":52.88,"end":53.34,"text":"bucpes"},{"start":53.9,"end":54.1,"text":"and"},{"start":54.2,"end":54.66,"text":"comingly"},{"start":54.76,"end":55.52,"text":"manufactured"},{"start":55.54,"end":55.64,"text":"in"},{"start":55.7,"end":56.1,"text":"apinning"},{"start":56.24,"end":56.68,"text":"products"},{"start":56.74,"end":56.88,"text":"of"},{"start":56.98,"end":57.16,"text":"all"},{"start":57.22,"end":57.32,"text":"the"},{"start":57.4,"end":57.72,"text":"world"},{"start":58.36,"end":58.56,"text":"ar"},{"start":58.62,"end":59.08,"text":"available"},{"start":59.18,"end":59.32,"text":"at"},{"start":59.38,"end":59.44,"text":"the"},{"start":59.5,"end":59.72,"text":"house"},{"start":59.88,"end":60.02,"text":"by"},{"start":60.26,"end":60.64,"text":" store"}],"paragraphs":[{"speaker":"U_UKN","start":0.76,"end":32.86},{"start":33.94,"end":60.64,"speaker":"U_UKN"}]},"clipName":"115Earhart Amelia Women.mp3","sttEngine":"deepspeech"} C:\Users\David\AppDa…rapper\index.js:166 transcribe Object index.js:37 Running interval to check for transcripts index.js:39 interval: checking transcirpt update index.js:16 In Electron Enviroment TranscriptCorrect.js:38 json Object 6 index.js:16 In Electron Enviroment index.js:37 Running interval to check for transcripts 6 index.js:16 In Electron Enviroment TranscriptForm.js:148 (form.checkValidity() true index.js:16 In Electron Enviroment TranscriptForm.js:89 formData.get('path') F:\transcription test\1hr test sn0803 Comparative Smartphone Technology 15 minutes.mp3 2 index.js:16 In Electron Enviroment C:\Users\David\AppDa…Wrapper\index.js:22 getDefaultSttAndLanguage Object deepspeech C:\Users\David\AppDa…rapper\index.js:159 about to call ipcRenderer - deepspeech 1 deepspeech C:\Users\David\AppDa…Wrapper\index.js:22 getDefaultSttAndLanguage Object deepspeech C:\Users\David\AppDa…rapper\index.js:161 about to call ipcRenderer - deepspeech 2 2 index.js:16 In Electron Enviroment TranscriptForm.js:115 ApiWrapper.createTranscript-response Object 2 index.js:16 In Electron Enviroment index.js:84 handleSaveItem Object 3 index.js:16 In Electron Enviroment index.js:99 setState C:\Users\David\AppDa…rapper\index.js:153 readMetadataForEDL Object transcriptId 2645ce9d772c44959f59bebfcbe4ad85 index.js:37 Running interval to check for transcripts index.js:39 interval: checking transcirpt update index.js:16 In Electron Enviroment index.js:37 Running interval to check for transcripts index.js:39 interval: checking transcirpt update index.js:16 In Electron Enviroment C:\Users\David\AppDa…rapper\index.js:140 newTranscriptDataVideo Object transcriptId 2645ce9d772c44959f59bebfcbe4ad85 index.js:37 Running interval to check for transcripts index.js:39 interval: checking transcirpt update index.js:16 In Electron Enviroment index.js:37 Running interval to check for transcripts index.js:39 interval: checking transcirpt update index.js:16 In Electron Enviroment index.js:37 Running interval to check for transcripts index.js:39 interval: checking transcirpt update index.js:16 In Electron Enviroment index.js:37 Running interval to check for transcripts 
Here is a screenshot of the section saying violation.

violation

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Projects
None yet
Development

No branches or pull requests

6 participants