Many self-media creators may encounter a challenge in creating subtitles for their videos. In the early days, some would manually transcribe the content, but later, Open AI released their Whisper language model as open-source. While it runs efficiently, using it requires some knowledge of Python commands, which could pose a significant learning barrier for some users.
So, a passionate engineer named Konstantin created a new open-source project called “Whisper Desktop.” With Whisper Desktop, users no longer need to understand Python commands; instead, they can easily generate video subtitle files with just one click through a user-friendly GUI interface!
Whisper Desktop – A Brief Introduction
| The project name | Whisper Desktop |
|---|---|
| Purpose | One-click automatic video subtitle file generation. |
| Cost | Free |
| Project URL | https://github.com/Const-me/Whisper |
| Software download address | https://github.com/Const-me/Whisper/releases |
| Language model download address | https://huggingface.co/ggerganov/whisper.cpp/tree/main |
Whisper Desktop Installation Guide
Step 1 | Download the Whisper Desktop software zip package.
1.First, let’s open the software release page. Here, we will find download resources for various versions of the software. Typically, we will directly select the latest version for download. Once we locate the latest version, click on 【WhisperDesktop.zip】 to initiate the download.
2.After the download is complete, we will unzip it and see the following contents.
Step 2 | Download the Whisper language model.
1.First, let’s go to the Whisper development webpage to understand which model specification is most suitable for our use. On this page, we can find the following table.
Booster reorganized the table as follows to help us better understand the model specifications we need:
| Size | Training Parameters | English-Only Model | Multi-language Model | Required GPU Memory | Subtitle Output Speed |
|---|---|---|---|---|---|
| tiny | 39 million | tiny.en | tiny | Approx 1GB | Super Fast |
| base | 74 million | base.en | base | Approx 1GB | Fast |
| small | 244 million | small.en | small | Approx 2GB | Moderate |
| medium | 769 million | medium.en | medium | Approx 5GB | Slow |
| large | 1.555 billion | Not available | large | Approx 10GB | Super Slow |
Here, Booster recommends selecting the model based on the “Required GPU Memory.” For example, if your computer’s GPU memory is 8GB, you can choose the “medium” model. If it’s 16GB, you can opt for the “large” model.
2.Once you have decided on the model specification, you can proceed to the model download page. For example, if you want to use the “medium” model with multi-language support, locate and click on 【ggml-medium.bin】, then click on the download icon on the right to initiate the download.
3.After the download is complete, for ease of use, we can place the downloaded model file into the WhisperDesktop folder, as shown in the image below.：
Step 3 | Open Whisper Desktop and Load the Language Model.
1.Use the 【left mouse button】 to double-click on Whisper Desktop to launch the program.
2.During the first startup, the program will prompt us to provide the path to the language model. Here, we simply choose the downloaded model file in the 【Model Path】. For the 【Model Implementation】, we can use the default option 【GPU】. Once done, click the 【OK】 button. At this point, we have completed all the installation settings!
Whisper Desktop User Guide
Operation Example 1 | General Video Subtitle Generation
1.
After completing the installation and setup, you will see the following screen. Proceed to select the video language under 【Language】, the video file for which you want to generate subtitles under 【Transcribe File】, and choose the output subtitle format under 【Output Format】. For the subtitle format, we recommend selecting 【SubRip subtitle】, as it is a standard subtitle format compatible with various video software and YouTube.
Next, check the option 【Place that file to the input folder】, which will place the subtitle file directly in the same folder as the video file. Finally, click the 【Transcribe】 button to initiate the subtitle generation process.
2.After clicking the Transcribe button, you will see a progress bar on the interface. Please wait a moment…
3.Once the progress bar is completed, a “Transcribe Completed” window will pop up, indicating that the subtitle file has been successfully generated.
4.When you open the generated subtitle file, you will likely find that there are almost no transcription errors. Isn’t it fantastic?!
Operation Example 2 | Translate Video to English Subtitles and then Generate
Indeed, Whisper Desktop not only supports general subtitle file output but also enables translating videos into English subtitles and then exporting them. Let’s give it a try!
1.At the top right corner of the program interface, you will find the option 【Translate】. Check the box next to it. Next, click the 【Transcribe】 button.
2.After the output is completed, let’s open the subtitle file. Here, Booster has opened both the original subtitles and the translated English subtitles simultaneously. How is it? Can you accept the translation quality? Doesn’t it look great?
Conclusion
In conclusion, Whisper Desktop is an excellent tool that simplifies the process of generating video subtitles for content creators. With its user-friendly GUI interface, users can effortlessly create subtitle files with just one click. The software supports multiple model specifications, catering to various GPU memory capacities. Additionally, Whisper Desktop offers the unique feature of translating videos into English subtitles before exporting. Overall, it provides a seamless and efficient solution for subtitle generation and translation tasks, making it a valuable asset for content creators and video enthusiasts alike.
