Organising sentences and audio clips

[Flickserve]

Scenario

 

Sentences from different sources. Text and audio clips

 

Is there an elegant way of organising them?

 

 

[roddy]

Something like this so you can tag a file with source / date / content? 

Does every bit of text have audio and vice versa, or do you have audioless text and textless audio (although that'd just be  matter of typing, I suppose).

[Flickserve]

I want to have every audio clip with text...at this point in time. Some way of keeping track of useful sentences to shadow. 

[Yadang]

I'm assuming you've already thought of this but I "organize" them in Anki.

 

Would that not work for what you're looking for?

[Flickserve]

True. Do you think it's the best solution?

[edelweis]

If you want SRS-like functionality with only audio, you could search the forum for "gradint".

I used that for a while, it creates mp3 "lessons" from collections of sentences.

If you have translations you can set up those as prompts. Otherwise just use the Chinese audio as prompt.

No guarantees as to ease of use, I don't know whether it's been maintained.

[Yadang]

On 1/12/2019 at 10:35 PM, Flickserve said:

True. Do you think it's the best solution?

 

I like that I can use the browser to easily search across thousands of sentences if I'm not sure how a word is used. I also like that I can back it all up, sync it with my other devices, etc.

[pon00050]

On 1/13/2019 at 2:03 AM, edelweis said:

If you want SRS-like functionality with only audio, you could search the forum for "gradint".

I used that for a while, it creates mp3 "lessons" from collections of sentences. 

 

Have you figured out how to use this program?

It looks kind of intimidating and I wasn't totally clear about how I can make use of it the last time I tried it.

[edelweis]

@pon00050 what's your OS? I used it on Linux.

[pon00050]

On 1/16/2019 at 11:54 AM, edelweis said:

what's your OS? I used it on Linux.

I use MacBook now.. so Mac

 

If needed, I have access to Windows computers too.

[edelweis]

Well, I don't have those so I downloaded the Linux version again and installed it following the INSTALL.txt directives.

 

Since the default config is to use english as prompt and chinese as the target language, I didn't touch the config files.

 

I put some mp3s in the "samples" directory, following the README.txt in that folder:

Since I had only Chinese recordings, I copied each Chinese sentence file twice named like this:

xxxxx_zh.mp3 (contains a sentence in Chinese)

xxxxx-meaning_zh.mp3 (a copy of the same file)

Alternatively, if you have the same sentence recorded in English, you would name the files this way:

xxxxx_zh.mp3 (contains a sentence in Chinese)

xxxxx_en.mp3 (contains the same sentence in English)

(you can create subfolders for various collections of files if you don't want to mix everything together).

 

Then I launched the gradint.py executable from the top folder.

I got a very basic UI that proposes several choices like "speaker", "mp3", etc.

I pressed "mp3" and then I pressed "make MP3" and then "quit".

Gradint made the first "lesson" as an mp3 file in the "output" subfolder.

The first lessons is full of holes as you don't have any reviews yet, but as I remember later lessons contain new sentences interspersed with reviews.

 

This is the basics, I don't know about Mac but the basic Linux version uses a rather ugly text to speech audio for some prompts.

You can install a better TTS, or (which is what I did when I used gradint previously) you can record your own prompts, and then put them in the samples/prompts directory (there's a README there) to make it nicer.

 

And of course you can check the config files "settings.txt" and advanced.txt" for more options, like choosing the languages that you want to study...

 

Hmmm I think I will dust of my collections of sentences and restart using gradint. Thanks @Flickserve @pon00050 

[pon00050]

3 hours ago, edelweis said:

Since I had only Chinese recordings, I copied each Chinese sentence file twice named like this:

 

Suppose I want to learn hundreds, maybe even thousands, of sentences or vocabs.

Do I need to do this step for each and every sentence?

 

 

I would love to make use of the program. I wish that I won't have to spend too much time preparing the study materials.

[edelweis]

If you have access to a command window, you can do this step for a bunch of files at a time.

 

example with bash (I hear bash is the command window shell in MacOS ?):

go to the directory containing a bunch of mp3 files that you want to duplicate + rename (use the command "cd") and then type commands such as:

 

for ffile in *.mp3;

do nname=`echo $ffile|sed 's/.mp3$//'`;

    cp "$ffile" "${nname}_zh.mp3";

    mv "$ffile" "${nname}-meaning.mp3";

    echo $ffile done;

done

 

(might not work well if some files have double quotes " in their names)

 

The readme in the samples directory also contains special instructions for "poetry-like" groups of sentences.

You can have a subdirectory containing only _zh.mp3 files + the special !poetry file,

and the previous xxxx_zh.mp3 file (in alphanumeric order) will be used as a prompt for the next file.

I think I recorded a xxxx_en.mp3 prompt for the very first file (containing the name of the poem, or the title of the dialogue, something to introduce the sequence of mp3 files that are to be learnt in order.)