ContentFries will automatically sync captions with time, but the actual accuracy differs depending on several factors.

Automatically generated captions for English are the most accurate with +/- 0.1s offset on the detected word start / end time.

We also provide automatic text to audio sync for dozens of languages. Please read the article below to understand these differences, and how you can fine-tune your output.

What is a timestamp error?

Behind the scenes, ContentFries sync mechanism will try to detect the correct timestamp of all words in the transcription. Depending on the method used, the output accuracy will differ. The timestamp error is up to 100 milliseconds for Premium models, but for manual sync some words can have timestamp error up to a couple of seconds.

Automatically generated captions

Premium V2 Model - English only (any dialect)

This system generated very accurate captions with very little offset error on word timestamps. From our observation, this error should not be greater than 0.1 sec. That means, if you say word "Hello" in 02:10 in your video, ContentFries should detect that and store the right word metadata.

Premium Model

Available for these languages and dialects:

Arabic (Algeria)
Arabic (Bahrain)
Arabic (Egypt)
Arabic (Iraq)
Arabic (Israel)
Arabic (Jordan)
Arabic (Kuwait)
Arabic (Lebanon)
Arabic (Morocco)
Arabic (Oman)
Arabic (Qatar)
Arabic (Saudi Arabia)
Arabic (State of Palestine)
Arabic (Tunisia)
Arabic (United Arab Emirates)
Arabic (Yemen)
Bulgarian (Bulgaria)
Catalan (Spain)
Chinese, Mandarin (Simplified, China)
Chinese, Mandarin (Traditional, Taiwan)
Croatian (Croatia)
Czech (Czech Republic)
Danish (Denmark)
Dutch (Belgium)
Dutch (Netherlands)
Finnish (Finland)
French (Belgium)
French (Canada)
French (France)
French (Switzerland)
German (Austria)
German (Germany)
German (Switzerland)
Greek (Greece)
Hindi (India)
Hungarian (Hungary)
Italian (Italy)
Italian (Switzerland)
Japanese (Japan)
Korean (South Korea)
Latvian (Latvia)
Lithuanian (Lithuania)
Malay (Malaysia)
Norwegian Bokmål (Norway)
Polish (Poland)
Portuguese (Brazil)
Portuguese (Portugal)
Romanian (Romania)
Russian (Russia)
Slovak (Slovakia)
Slovenian (Slovenia)
Spanish (Argentina)
Spanish (Bolivia)
Spanish (Chile)
Spanish (Colombia)
Spanish (Costa Rica)
Spanish (Dominican Republic)
Spanish (Ecuador)
Spanish (El Salvador)
Spanish (Guatemala)
Spanish (Honduras)
Spanish (Mexico)
Spanish (Nicaragua)
Spanish (Panama)
Spanish (Paraguay)
Spanish (Peru)
Spanish (Puerto Rico)
Spanish (Spain)
Spanish (United States)
Spanish (Uruguay)
Spanish (Venezuela)
Swedish (Sweden)
Turkish (Turkey)

This model offers premium captions accuracy.

Standard Model

Available for these languages and dialects:

Afrikaans (South Africa)
Albanian (Albania)
Amharic (Ethiopia)
Armenian (Armenia)
Azerbaijani (Azerbaijan)
Basque (Spain)
Bengali (Bangladesh)
Bengali (India)
Bosnian (Bosnia and Herzegovina)
Burmese (Myanmar)
Chinese, Cantonese (Traditional Hong Kong)
Estonian (Estonia)
Filipino (Philippines)
Galician (Spain)
Georgian (Georgia)
Gujarati (India)
Hebrew (Israel)
Icelandic (Iceland)
Indonesian (Indonesia)
Javanese (Indonesia)
Kannada (India)
Khmer (Cambodia)
Lao (Laos)
Macedonian (North Macedonia)
Malayalam (India)
Marathi (India)
Mongolian (Mongolia)
Nepali (Nepal)
Persian (Iran)
Punjabi (Gurmukhi India)
Serbian (Serbia)
Sinhala (Sri Lanka)
Sundanese (Indonesia)
Swahili (Kenya)
Swahili (Tanzania)
Tamil (India)
Tamil (Malaysia)
Tamil (Singapore)
Tamil (Sri Lanka)
Telugu (India)
Thai (Thailand)
Ukrainian (Ukraine)
Urdu (India)
Urdu (Pakistan)
Uzbek (Uzbekistan)
Vietnamese (Vietnam)
Zulu (South Africa)

Automatic captions generated by this model provide good output if the audio quality is high enough. Sometimes this model may skip some words / sentences, and it is the worst performing from our collection.

Automatic sync

If you have the video script or captions file, you can use it instead of the automatic captions generation process.

Please note that your transcription should be as close to what is being said in the video as possible to get the best results.

Warning: word timestamp error can climb up to ~30 seconds, especially with long video content.

This sync system works on a very different principle than the automatic captions generation and is intended to be used for captions timestamps, not for word timestamps, but we will make improvements over time to make it more accurate.

Supported languages:

Afrikaans
Amharic
Arabic
Aragonese
Assamese
Azerbaijani
Bengali
Bosnian
Bulgarian
Catalan
Czech
Mandarin Chinese
Welsh
Danish
German
Greek (Modern)
English
Esperanto
Estonian
Basque
Persian
Finnish
French
Scottish Gaelic
Irish
Galician
Greek (Ancient)
Guarani
Gujarati
Hebrew
Hindi
Croatian
Hungarian
Armenian
Interlingua
Indonesian
Icelandic
Italian
Lojban
Japanese
Greenlandic
Kannada
Georgian
Kirghiz
Korean
Kurdish
Latin
Latvian
Lingua Franca Nova
Lithuanian
Malayalam
Marathi
Macedonian
Maltese
Malay
Burmese
Nahuatl
Nepali
Dutch
Norwegian
Oriya
Oromo
Panjabi
Papiamento
Polish
Portuguese
Romanian
Russian
Sinhala
Slovak
Slovenian
Spanish
Albanian
Serbian
Swahili
Swedish
Tamil
Tatar
Telugu
Thai
Tswana
Turkish
Ukrainian
Urdu
Vietnamese
Yue Chinese
Chinese

How to fine-tune your settings

If you need to fix the timing of fries generated by the trimmer, the solution is very simple.

Locate and adjust the actual starting and ending time on the audio timeline

Fry

If your video is long, zoom in on the audio timeline and find pieces of fries where you want to fine-tune the timings.

Locate the real starting time by listening to the video. If you're not sure what words the fry is starting with, double-click the yellow zone to open up the fry level, and then come back to adjust the timings like this:



You can use the same process for the ending time.

Locate and adjust timings of cut outs on the fry level

If you deleted some parts of the text on the fry level, you can use the same process to find and fine-tune their timings. Note that on the fry level the cut out zones have red background color instead of yellow, which applies to fries in the trimmer.
Was this article helpful?
Cancel
Thank you!