Pytranscriber Free Download

[Annette Fazzari]

FORmacOS CATALINA AND ABOVE:

If you get a message "file damaged, you should move it to trash". You need to open a Terminal and type the following command to add an exception allowing the app to run:

For users in mainland China or other places where Google is blocked

The app will show a failure to connect to the internet message... meaning Google Servers.

You need to either use a proxy service and set it up at pytranscriber settings > proxy or install a desktop VPN app like Windscribe to bypass the blockage and be able to use pyTranscriber.

If you use Gnome 40 or higher with Wayland you need to use the special version for that. (pyTranscriber-v1.7-linux-gnome40-wayland*.zip)

Also make sure you have qtwayland5 package installed on your system (for Ubuntu apt install qtwayland5)

Note for Mac users:

There has been bug reports of problems running the app on the latest MacOS version. If you also experience the same issues the workaround for now is to install Ubuntu on VirtualBox for Mac, install pyTranscriber and run it from there... this setup shouldn't take more than 30 minutes.

Your donations will help me free up my time to do more direct work on pyTranscriber. If I can generate more income from sponsorship here, I can take time out of other jobs and commit more time to work on bug fixes, add new features, and make new releases, like possible future integration with Mozilla Deep Speech engine, which may improve accuracy and control over the speech detection.

For users in mainland China or other places where Google is blocked

The app will show a failure to connect to the internet message... meaning Google Servers.

You need to install a desktop VPN app like Windscribe to bypass the blockage and be able to use pyTranscriber.

For users in mainland China or other places where Google is blocked

The app will show a failed to connect to the internet message... meaning Google Servers.

You need to install a desktop VPN app like Windscribe to bypass the blockage and be able to use pyTranscriber.

Note for Mac users:

*Start the app clicking on launch.command

*As the app is not downloaded from the AppStore, it is necessary to allow it to run manually at System Preferences > Security and Privacy

I can use pytranscriber to create the subtitle text in a text file, and then I have to copy text one line at a time and paste into premiere text tool. Is there a way to automate that? There are hours of dialogue to add subtitle to.

It is a tool to translate movie subtitles from one language into another, or even show multiple language subtitles together. The tool is powered by Google Translate, thus even though the translations might not always be perfect, it supports a very wide range of languages!

The translator can be either used to automatically extract the subtitle from the video file (e.g. .avi, .mkv) and then perform translation on that subtitle file or process a subtitle file (e.g. .srt, .ass) instead. The first required an ffpmeg installed and setup to work from terminal. If you extract subtitle yourself, note that often the file format for subtitles is SRT. This only has minimal styling thus Anime usually uses ASS format, which can even do animations. I recommend naming output file as some_name.ass if you want the styling to remain and some_name.srt if you do not want styling.

Another really nice feature is being able to merge both the translation AND the original subtitles together. The original can be made smaller and slightly opaque to not distract and not take up too much space:

If you would like to learn a new language, you might as well show both, the language you would like to learn (Main) AND the one you speak very well (Secondary) (slightly smaller font and slightly opaque to not disturb, as shown in the example picture).

E.g. If the language you speak is English and you would like to learn Spanish for some series, that does NOT provide Spanish subtitles, simply use flag --merge and English subs will be translated into Spanish(translated) + English:

If, however, the series DOES have spanish subtitles, I would instead recommend translating the Spanish into English, since Google translate does not give 100% accurate text, thus by translating FROM Spanish you will get better Spanish subs quality, while English will not matter that much, since that is secondary subtitles. To make sure that you still see the Spanish on top and English at the bottom use flag --reverse, which will then generate Spanish + English(translated) subs:

Another flag, which is useful when merging two languages together is --line_char_limit. Often instead of showing one long line, two or even three lines are displayed in subs. While they all would still fit within a single line after being translated, when --merge is used, that would double the line count and would block a large portion of the video. To solve this add --line_char_limit with a number of around 70. This basically means that if there are less than 80 characters within a sub, remove all new lines. Of course, for some subs this number could be higher, or smaller, depending on the font size, thus might have to test a little bit before getting perfect result or count how many characters is safe to read within a single line.

You can also change the subs that are at the bottom fond scale using --secondary_scale. 100 will represent font size equal to the main subs at the top (100%) and smaller value will make them proportionally smaller:

Languages like Japanese, Chinese and many others use a non-latin characters. If you are learning a new language, it is likely you can't read the new alphabet as quickly as it is required to follow the subs. For that purpose --pronounce_translated to show pronunciation of the translation and --pronounce_original to show pronunciation of the original subs.

E.g. you might be learning Japanese and can't be bothered learning all the Hiragana, Katakana and Kanji, but you want to start understanding better Anime as you watch it. In that case it is best to find Japanese subs on kitsunekko.net and use --pronounce_original flag (also note --merge and --reverse):

If you have more advanced Japanese understanding (yet cannot read the characters), you can remove English subs altogether by performing Japanese to Japanese translation, but adding --pronounce_translated flag:

Alternatively, if you cannot find the japanese subs you can translate the english ones straight to Japanese, however can't guarantee that what is being spoken is exactly what you will be reading since Japanese language is more difficult to translate...

The tool supports a couple of translation libraries: googletrans and google_trans_new. In rare cases the translation might fail using one of the libraries. When that happens simply try another one :) You can choose which one to use with flag --translator:

In the future I would like to add official google translate API support, but that would require acquiring Google Translation API Key and passing it into the tool. If, however, you're translating 1-5 episodes per day, then using one of the two supported APIs is OK, however for very large amounts official API would be best, since then you could extend quota limits.

Note: google_trans_new ignores ALL new lines, meaning if there was some new lines \n within original subs, they will ALL get removed in both translations AND pronunciations. googletrans on the other hand keeps the new lines within translations, however removes them for pronunciations. Also note that the behavior might change in the future, since I am not responsible for maintaining these libraries.

The tool uses a free googletrans API, which uses one of the google domains e.g. translate.google.com or translate.google.co.uk to perform translation. After a couple of calls that domain gets blocked and thus another one is selected instead. I added 17 domains, which should ensure that you will always have a domain that still works, because after about 1h that domain gets unblocked. Don't worry, you can still go to chrome and use the google translate :)

The tool works best with English language, since some others might have strange characters that might make things funny... However the use of different separators selected automatically should ensure that things work (I did see Portugese fail for some reason, might have to investigate later). Although even in case of failure I made sure that even if it fails, it continues and produces the subs, just they might be misaligned with the main subs text...

3a8082e126