Salad language code, diarization and important note

[ericcpi]

1. Why is mp4 not supported in the upload file when in the site it can accept mp4? https://salad.com/transcription

2. Enter the language code here. Transcription is available in 97 languages. The default value is "en".

I reached out to support of Salad and got this information.
Thank you for contacting SaladCloud support. Here is a list of the languages we support at this time:

Afrikaans, Albanian, Amharic, Arabic, Armenian, Assamese, Azerbaijani, Bashkir, Basque, Belarusian, Bengali, Bosnian, Breton, Bulgarian, Burmese, Catalan, Chinese (Cantonese), Chinese (Mandarin), Croatian, Czech, Danish, Dutch, English, Estonian, Faroese, Finnish, French, Galician, Georgian, German, Greek, Gujarati, Haitian Creole, Hausa, Hebrew, Hindi, Hungarian, Icelandic, Indonesian, Italian, Japanese, Javanese, Kannada, Kazakh, Khmer, Korean, Lao, Latvian, Lithuanian, Macedonian, Malay, Malayalam, Maltese, Marathi, Mongolian, Nepali, Norwegian, Oriya, Pashto, Persian, Polish, Portuguese, Punjabi, Romanian, Russian, Sanskrit, Serbian, Shona, Sindhi, Sinhala, Slovak, Slovenian, Somali, Spanish, Sundanese, Swahili, Swedish, Tagalog, Tajik, Tamil, Tatar, Telugu, Thai, Turkish, Turkmen, Ukrainian, Urdu, Uzbek, Vietnamese, Welsh, Yiddish, Yoruba, Zulu.

Do I put the exact word Tagalog? I tried putting tagalog and the workflow returned english.

Transcription Text:
In the family, there is one who breaks the cycle of poverty. What if... it's you? The one who breaks the cycle of poverty. The one who breaks the stigma that we are just poor. The one who is like, from you... No, no, we are not poor anymore. Yes, it's hard now. It's a tough time right now. There are challenges. And challenges will never be gone. The good thing is, you know that even if you're facing challenges, there are those who will continue to fight. Just keep fighting. Because it's worth it. It's worth it to have a good life. There's a happy life waiting for you.

 

[ArshilAhmad]

Hi @ericcpi,

1. Why is mp4 not supported in the upload file when in the site it can accept mp4? https://salad.com/transcription
View attachment 49323

Have you tried using an MP4 file?

2. Enter the language code here. Transcription is available in 97 languages. The default value is "en".
View attachment 49324

I reached out to support of Salad and got this information.
Thank you for contacting SaladCloud support. Here is a list of the languages we support at this time:

Afrikaans, Albanian, Amharic, Arabic, Armenian, Assamese, Azerbaijani, Bashkir, Basque, Belarusian, Bengali, Bosnian, Breton, Bulgarian, Burmese, Catalan, Chinese (Cantonese), Chinese (Mandarin), Croatian, Czech, Danish, Dutch, English, Estonian, Faroese, Finnish, French, Galician, Georgian, German, Greek, Gujarati, Haitian Creole, Hausa, Hebrew, Hindi, Hungarian, Icelandic, Indonesian, Italian, Japanese, Javanese, Kannada, Kazakh, Khmer, Korean, Lao, Latvian, Lithuanian, Macedonian, Malay, Malayalam, Maltese, Marathi, Mongolian, Nepali, Norwegian, Oriya, Pashto, Persian, Polish, Portuguese, Punjabi, Romanian, Russian, Sanskrit, Serbian, Shona, Sindhi, Sinhala, Slovak, Slovenian, Somali, Spanish, Sundanese, Swahili, Swedish, Tagalog, Tajik, Tamil, Tatar, Telugu, Thai, Turkish, Turkmen, Ukrainian, Urdu, Uzbek, Vietnamese, Welsh, Yiddish, Yoruba, Zulu.

Do I put the exact word Tagalog? I tried putting tagalog and the workflow returned english.

Transcription Text:
In the family, there is one who breaks the cycle of poverty. What if... it's you? The one who breaks the cycle of poverty. The one who breaks the stigma that we are just poor. The one who is like, from you... No, no, we are not poor anymore. Yes, it's hard now. It's a tough time right now. There are challenges. And challenges will never be gone. The good thing is, you know that even if you're facing challenges, there are those who will continue to fight. Just keep fighting. Because it's worth it. It's worth it to have a good life. There's a happy life waiting for you.

Could you please share the Workflow URL where you have tried this?

 

[ericcpi]

Have you tried using an MP4 file?

I just tested and it worked on mp4.

Could you please share the Workflow URL where you have tried this?

Pabbly - Online Marketing & Sales Software

Automate your sales and marketing with Pabbly. Unite email marketing, lead capture, WordPress themes, subscription all at one place.

connect.pabbly.com

Pabbly - Online Marketing & Sales Software

Automate your sales and marketing with Pabbly. Unite email marketing, lead capture, WordPress themes, subscription all at one place.

connect.pabbly.com

Additional information which maybe a bug.
The same video in this example.

If I place empty on the the language code and have pabbly workflow run on the file, it returns english, which is wrong.

File Name: 11132024-1104AM-motivational speech for filipino's [ ezmp3.cc ].mp3

File URL: https://drive.google.com/uc?id=1SKbOTIQJAg7HXE0no9EBsn8NkOc1hUad&export=download

Transcription Text:
In the family, there is one who breaks the cycle of poverty. What if... it's you? The one who breaks the cycle of poverty. The one who breaks the stigma that we are just poor. The one who is like, from you... No, no, we are not poor anymore. Yes, it's hard now. It's a tough time right now. There are challenges. And challenges will never be gone. The good thing is, you know that even if you're facing challenges, there are those who will continue to fight. Just keep fighting. Because it's worth it. It's worth it to have a good life. There's a happy life waiting for you.

If I call it thru postman using the same video, the srt and the output text is correct in Tagalog.

"input": {
"url": "https://drive.google.com/uc?id=1SKbOTIQJAg7HXE0no9EBsn8NkOc1hUad&export=download",
"language_code": "",
"word_level_timestamps": false,
"diarization": false,
"srt": true
}

Here's the response. Which is in Tagalog. Looks like pabbly supplies an english value maybe by default if nothing is supplied.

"output": {
"srt_content": "1\n00:00:00,101 --> 00:00:03,140\nSa pamilya, may isang pumutol ng kahirapan.\n\n2\n00:00:05,344 --> 00:00:07,400\nWhat if, ikaw yan?\n\n3\n00:00:09,722 --> 00:00:12,340\nYung puputol ng cycle, yung puputol ng kahirapan.\n\n4\n00:00:14,525 --> 00:00:16,880\nYung puputol ng stigma na mahirap lang tayo.\n\n5\n00:00:18,921 --> 00:00:23,020\nYung tipong from you, hindi, hindi, hindi na tayo mahirap.\n\n6\n00:00:26,705 --> 00:00:29,540\nYes, naghihirap ngayon, may pinagdadaanan ngayon.\n\n7\n00:00:30,325 --> 00:00:31,320\nMay challenges pa rin ngayon.\n\n8\n00:00:31,420 --> 00:00:32,575\nAnd di mawawala yung challenges.\n\n9\n00:00:35,562 --> 00:00:36,545\nAng kagandahan is,\n\n10\n00:00:37,708 --> 00:00:38,390\nalam mo na,\n\n11\n00:00:38,410 --> 00:00:40,856\nkahit may dinadaan ng challenges,\n\n12\n00:00:41,217 --> 00:00:41,939\nmay nanguhusga,\n\n13\n00:00:41,979 --> 00:00:42,400\ntuloy lang.\n\n14\n00:00:44,330 --> 00:00:44,800\nLaban lang.\n\n15\n00:00:46,266 --> 00:00:47,080\nKasi worth it yun eh.\n\n16\n00:00:47,301 --> 00:00:48,577\nWorth it yung magandang buhay.\n\n17\n00:00:49,803 --> 00:00:52,000\nMay napakasayang buhay na naghihintay.\n\n",
"text": " Sa pamilya, may isang pumutol ng kahirapan. What if, ikaw yan? Yung puputol ng cycle, yung puputol ng kahirapan. Yung puputol ng stigma na mahirap lang tayo. Yung tipong from you, hindi, hindi, hindi na tayo mahirap. Yes, naghihirap ngayon, may pinagdadaanan ngayon. May challenges pa rin ngayon. And di mawawala yung challenges. Ang kagandahan is, alam mo na, kahit may dinadaan ng challenges, may nanguhusga, tuloy lang. Laban lang. Kasi worth it yun eh. Worth it yung magandang buhay. May napakasayang buhay na naghihintay.",
"duration_in_seconds": 52.741224,
"duration": 0.01,
"processing_time": 14.149931192398071
}

 

[Preeti Paryani]

Hello @ericcpi

Thanks for the confirmation.

I just tested and it worked on mp4.

As per the help text, the default language code is English, so if the field is left empty the output text will be in English.

If I call it thru postman using the same video, the srt and the output text is correct in Tagalog.

Have you tried adding Tagalog under the language code field? If not please try adding "tl" or"tgl" for Tagalog and let us know if it works fine for you.

If I place empty on the the language code and have pabbly workflow run on the file, it returns english, which is wrong.

 

[ericcpi]

Hi @Pretti

I reached out to support of salad and here's what they had to say.

Eric,

Thank you for your patience, I had to consult one of our developers for additional insight.

In this case, you will need to make two API calls in order to receive output in both English and Tagalog. For the Tagalog workflow, you do not need a language code. Our model will detect the language and return the transcription in that language.

To translate the audio into English, you will need to run a separate English Translation workload, which will return an English file. To do that, set "translate:" "to eng".

The salad API is smart enough to distinguish it from English and Tagalog and no language code is needed.
If I place Tagalog in your language code, it will return as English, which I was saying was wrong.

Create a job for Salad Transcription API - SaladCloud

Creates a job for Salad Transcription API

docs.salad.com

Even if it says english on their API, i think on your end you should not hardcode it to 'en' if the language code is blank.

 

[Preeti Paryani]

Hello @ericcpi

Our team has made some changes at our end. Kindly make a few tests and let us know if it's working fine for you now.

 

[ericcpi]

Awesome support team here in Pabbly Connect
Looks good on my end.

 

[Preeti Paryani]

Hello @ericcpi

Thanks for the update, we'd love to hear your thoughts! If you have a moment, please rate us on Google. Your feedback helps us improve.