2021 recap
Written by Sjors on 19 January, 2022
2021 is behind us, time for a recap.
This recap doesn't include data about visitors or pageviews, I don't collect that data.
Subtitle Tools doesn't use Google Analytics or other 3rd-party tracking software.
The only data I collect are anonymous statistics on how and how often tools are used.
Plain text tools
Most of the tools on the site work with plain text subtitles. Plain text means the subtitles are files with normal text inside, you can open them with Notepad and make changes.
- The plain text tools were used 2.864.255 times (an average of 7.869 times per day)
- A total of 4.952.672 files were uploaded
- Those files had a total size of 468.7 GB (that's a lot of text!)
These were the most popular plain text tools:
- Convert to Srt — 1.174.633 uses (41% of total)
- Subtitle Shifter — 548.864 uses (19% of total)
- Convert to UTF-8 — 296.048 uses (10% of total)
And unsurprisingly, the least used tool this year:
- Make Pinyin Subtitles — 8.759 uses (0% of total)
I made the Pinyin subtitles tool to help Chinese language learners. The tool converts Chinese subtitles to Pinyin, which is a phonetic alphabet for Chinese. It doesn't get used much because it is a niche tool. It also doesn't help that it is pretty difficult to find Chinese subtitles for Chinese tv-shows.
Subpicture tools
Subtitle Tools also converts subpictures to srt. A subpicture is a subtitle file that contains images of text. To convert subpictures to srt, we first have to extract all the images from the subpicture.
- We converted 235.066 subpicture subtitles (an average of 646 per day)
- Those subpictures had a combined size of 2.3 TB
- We extracted 139.723.542 images from subpictures (an average of 383.856 per day)
- Extracting those images took 4871 hours of work (an average of 13 hours per day)
The second step to converting a subpicture is reading the text from the image. This is done using OCR software. OCRing is very CPU-intensive, it takes up an entire CPU-core, using 100% processing power until the work is done. In 2021, Subtitle Tools ran 3 additional servers to do all the OCR work.
- OCRing all the images took 16.367 hours of work (an average of 45 hours per day, spread out over 3 servers)