پروژه زیرنویس گذاری (111668)

---

شرح پروژه

سلام یک پروژه زیر نویس داشتم

این فیلم

این کارها باید روش انجام بشه

Project 1: Individual Subtitling Project (Subtitle Generation & Human 

Optimization) 

1. Objective 

The primary objective of this project is to develop a critical understanding of the strengths 

and limitations of contemporary AI-driven speech-to-text (STT) systems for subtitle 

generation. It also aims at developing skills in creating and refining subtitles. Students will 

generate English subtitles using Subtitle Edit’s automatic transcription, then translate them 

into Persian and perform a human correction process to refine the raw AI output into 

professional, broadcast-quality subtitles — emphasizing accuracy, timing, and linguistic 

naturalness.

2. Required Tools & Materials 

• Software: Subtitle Edit 

• Video Source: A short (3–5 minute) English-language video clip worth subtitling 

3. Project Workflow 

Step 1: Subtitle Generation (Using Subtitle Edit)

• Open Subtitle Edit and load the local video file (Video → Open Video File).

• Navigate to Video → Audio to text (automatic)...

• Select a transcription engine (e.g., Whisper, Vosk, or Google Cloud Speech-to-

Text). Use the default or a "large" model for best accuracy.

• Generate the subtitles and save this raw output as SE_Raw_Transcribed.srt.

Step 2: Initial Review & Selection

• Play the video alongside the generated subtitles in Subtitle Edit.

• Assess the raw output based on:

o Word Accuracy: Identify substitution, omission, and insertion errors. Note 

misheard words (e.g., "write" vs. "right") or nonsensical phrases.

o Timing & Synchronization: Evaluate alignment of subtitle events with speech. 

Check for excessive lead-in/lag-out times and overlapping subtitles.

o Punctuation & Segmentation: Assess sentence boundaries (periods, 

question marks) and line breaks.

• This raw file will proceed to the next step.

Step 3: Machine Translation (MT)

• With the SE_Raw_Transcribed.srt file open in Subtitle Edit, go to Tools → Auto-translate.

• Choose a translation engine (Google Translate, Microsoft Translator, etc.).

• Set the target language and perform the translation.

• Save this raw MT output as MT_Raw_TargetLang.srt.

Step 4: Human Correction & Technical Styling 

1. Linguistic & Semantic Correction:

o Ensure the translation is idiomatic and natural in the target language.

o Correct MT errors like literal translations, wrong terms, and grammatical 

awkwardness.

o Preserve original meaning, humor, and cultural context where possible.

2. Final Formatting:

o Perform a final review with the video playing.

o Export the final, corrected file. Use .srt for universal compatibility 

or .ssa/.ass for advanced styling.

o Save as Final_Corrected.srt (or .ssa).

4. Deliverables & Submission 

1. The video source file

2. File A: SE_Raw_Transcribed.srt – The raw, unedited subtitles generated by Subtitle 

Edit.

3. File B: Final_Corrected.srt (or .ssa) – The final, human-corrected, and technically 

optimized subtitle file in the target language.

4. Analytical Report (PDF): Report_LastName.pdf (1–2 pages, 250–350 words)

o Introduction: Brief project description, including

▪ The reason for selecting the specific video file (e.g., clarity of speech, 

length, topic, or variety of speakers).

▪ A short statement confirming that the video does not have any pre-

existing Persian subtitles available online, ensuring all Persian subtitle 

output is generated from scratch through AI transcription and 

translation.

o Error Analysis: List the 2–3 most common types of errors found in the STT 

and MT outputs and provide examples. 

o Correction Process: Briefly describe your main strategies for correcting and 

optimizing the subtitles.

Conclusion: Reflect on the role of human expertise in the age of AI-generated 

subtitles.

Best of Luck

این کار ها دقیق باید روش انجام بشه

چقدره هزینه چون یک گزارشم براساسش می خوام

۲۰ روز

این پروژه شامل 1 فایل مهم است، لطفا قبل از ارسال پیشنهاد حتما نسبت به بررسی این فایل اقدام فرمایید.

مهارت ها و تخصص های مورد نیاز

---

---