📚 Year 10 SQA Computing Science: Oral and Listening Exam Preparation Special | SQA 计算机：口语/听力备考专项
In the SQA Computing Science course, topics such as sound representation, speech recognition, and audio processing form the foundation for understanding how computers handle oral and listening tasks. This article bridges the gap between exam preparation for speaking and listening components and the underlying computer science principles, offering a focused revision guide for Year 10 learners. We will explore how sound is digitised, how voice assistants work, and how to tackle exam questions confidently.
在 SQA 计算机科学课程中，声音表示、语音识别和音频处理等主题是理解计算机如何处理口语和听力任务的基础。本文在口语与听力备考和计算机科学原理之间架起桥梁，为 Year 10 学生提供一份重点突出的复习指南。我们将探索声音如何被数字化、语音助手如何工作，以及如何自信地应对考试题目。
1. Sound as a Wave | 声音作为波形
Sound travels through the air as a longitudinal pressure wave. When we speak, our vocal cords vibrate, creating variations in air pressure that a microphone can capture. In computing, we represent this continuous analogue signal as a series of discrete digital samples.
声音以纵波的形式在空气中传播。当我们说话时，声带振动产生气压变化，麦克风可以捕捉到这些变化。在计算机中，我们将这种连续的模拟信号表示为一组离散的数字样本。
The two key properties of a sound wave are amplitude (loudness) and frequency (pitch). Amplitude determines how loud a sound is, measured in decibels (dB). Frequency, measured in hertz (Hz), indicates how high or low a note sounds. Human speech typically ranges from 85 Hz to 255 Hz for the fundamental frequency, but consonants and vowels contain much higher frequency components essential for intelligibility.
声波的两个关键属性是振幅（响度）和频率（音高）。振幅决定声音的响亮程度，以分贝 (dB) 为单位。频率以赫兹 (Hz) 为单位，表示音调的高低。人类语音的基频通常在 85 Hz 至 255 Hz 之间，但辅音和元音包含了高得多的频率分量，这些对于语音的可懂度至关重要。
2. Sampling Rate and Bit Depth | 采样率与位深度
To convert spoken words into a digital file, an Analogue-to-Digital Converter (ADC) takes measurements of the sound wave at regular intervals. The number of samples taken per second is called the sampling rate, expressed in kilohertz (kHz). A common rate for speech is 8 kHz, while CD-quality audio uses 44.1 kHz. Higher sampling rates capture higher frequencies more accurately, which is vital for preserving the clarity of consonants in listening exams.
为了将口语转换为数字文件，模数转换器 (ADC) 会按固定的时间间隔对声波进行测量。每秒采集的样本数称为采样率，以千赫兹 (kHz) 表示。语音常用的采样率为 8 kHz，而 CD 音质的音频使用 44.1 kHz。更高的采样率能更准确地捕捉高频，这对于在听力考试中保持辅音的清晰度至关重要。
Bit depth refers to the number of bits used to store each sample. It determines the dynamic range — the difference between the quietest and loudest sound that can be recorded. An 8-bit depth offers 256 possible levels, sufficient for basic speech, while 16-bit audio provides 65,536 levels and is standard for high-quality recordings. In speech recognition, a bit depth of 16 bits is often used to ensure subtle variations in voice are captured.
位深度是指用于存储每个样本的比特数。它决定了动态范围——即可以录制的最安静和最响亮声音之间的差异。8 位深度有 256 个级别，足以满足基本的语音需求；而 16 位音频提供 65,536 个级别，是高音质录音的标准。在语音识别中，常使用 16 位深度来确保捕捉到语音中的细微变化。
3. Calculating Audio File Size | 计算音频文件大小
Understanding how to calculate the size of an uncompressed audio file helps predict storage requirements and is a common SQA exam topic. The formula is:
了解如何计算未压缩音频文件的大小有助于预测存储需求，这也是 SQA 考试中的常见主题。计算公式为：
ile Size (bits) = Sampling Rate (Hz) × Bit Depth × Duration (seconds) × Number of Channels
文件大小 (比特) = 采样率 (Hz) × 位深度 × 时长 (秒) × 声道数
To convert bits to bytes, divide by 8. For example, a 30-second mono recording of spoken instructions for a listening test, sampled at 11.025 kHz with 8-bit depth, would have a file size of:
要将比特转换为字节，需除以 8。例如，为听力测试录制的一段 30 秒单声道语音指令，采用 11.025 kHz 采样率和 8 位深度，其文件大小计算如下：
11025 × 8 × 30 × 1 = 2,646,000 bits ÷ 8 = 330,750 bytes ≈ 330.75 KB
If the same recording were in stereo (2 channels) and CD-quality (44.1 kHz, 16-bit), the file size would increase dramatically, demonstrating the trade-off between quality and storage.
如果同一段录音采用立体声（2 声道）和 CD 音质（44.1 kHz，16 位），文件大小将大幅增加，这体现了音质与存储之间的权衡。
4. Audio File Formats and Compression | 音频文件格式与压缩
Different audio file formats use various compression techniques to reduce file size. Lossless compression retains all original data and allows perfect reconstruction, while lossy compression removes some audio data deemed less perceptible to human ears. For oral and listening exams, it is important to know which format preserves clarity.
不同的音频文件格式使用各种压缩技术来减小文件体积。无损压缩保留所有原始数据并能完美重建，而有损压缩则移除人耳难以觉察的部分音频数据。对于口语和听力考试，了解哪种格式能保持清晰度非常重要。
| Format | Compression | Typical Use | Quality for Speech |
|---|---|---|---|
| WAV | Uncompressed | Professional audio, exam prompts | Excellent, full fidelity |
| FLAC | Lossless | Archiving, language learning apps | Excellent, identical to original |
| MP3 | Lossy | Music, online listening exercises | Good at high bitrates (192 kbps+) |
| AAC | Lossy | Mobile devices, video audio | Better than MP3 at same bitrate |
中文对照表
| 格式 | 压缩类型 | 典型用途 | 语音音质 |
|---|---|---|---|
| WAV | 未压缩 | 专业音频，考试提示音 | 极佳，完整保真 |
| FLAC | 无损 | 存档，语言学习应用 | 极佳，与原版一致 |
| MP3 | 有损 | 音乐，在线听力练习 | 高比特率下良好 (≥192 kbps) |
| AAC | 有损 | 移动设备，视频音频 | 相同比特率下优于 MP3 |
5. Speech Recognition Technology | 语音识别技术
Speech recognition, or Automatic Speech Recognition (ASR), allows a computer to convert spoken language into text. The process involves capturing audio through a microphone, digitising it, and using algorithms to match acoustic patterns with phonetic units. Modern systems employ machine learning models trained on vast datasets of diverse voices.
语音识别（自动语音识别，ASR）让计算机能够将口头语言转换为文本。其过程包括通过麦克风捕获音频、将其数字化，然后使用算法将声学模式与音素单元进行匹配。现代系统使用在大量不同声音数据集上训练过的机器学习模型。
Key stages include feature extraction, where the system identifies Mel-Frequency Cepstral Coefficients (MFCCs) that represent the short-term power spectrum of speech. These features are then fed into a Hidden Markov Model (HMM) or a deep neural network to predict words. In oral exams, ASR can be used to evaluate pronunciation or to transcribe spoken answers for automated scoring.
关键步骤包括特征提取，系统识别出代表语音短时功率谱的梅尔频率倒谱系数 (MFCC)。然后，这些特征被输入到隐马尔可夫模型 (HMM) 或深度神经网络中以预测单词。在口语考试中，ASR 可用于评估发音，或将口头回答转录为文字进行自动评分。
6. Speech Synthesis (Text-to-Speech) | 语音合成（文本转语音）
Text-to-Speech (TTS) technology converts written text into spoken audio, providing the voice output for virtual assistants and reading out listening comprehension passages. Early systems used concatenative synthesis, stitching together pre-recorded speech fragments. Modern TTS relies on parametric models and neural networks that generate more natural-sounding speech.
文本转语音 (TTS) 技术将书面文本转换为语音音频，为虚拟助手提供语音输出并朗读听力理解段落。早期系统使用拼接合成法，将预录的语音片段拼接起来。现代 TTS 依靠参数模型和神经网络生成听起来更自然的语音。
In an exam context, TTS can deliver listening test prompts consistently. The intelligibility of synthetic speech depends on factors like prosody — the rhythm, stress, and intonation of speech. Scottish English, for instance, has distinct prosodic features, and good TTS engines must model these to avoid distorting meaning during localised exams.
在考试场景中，TTS 可以稳定地传送听力测试提示。合成语音的可懂度取决于韵律等因素——即语音的节奏、重音和语调。例如，苏格兰英语具有独特的韵律特征，优秀的 TTS 引擎必须模拟这些特征，以免在本地化考试中曲解含义。
7. Voice User Interfaces | 语音用户界面
A Voice User Interface (VUI) lets users interact with a computer through voice commands. This technology integrates speech recognition and synthesis to create a conversational experience. Examples include Siri, Alexa, and Google Assistant. In education, VUIs can be used for interactive language practice, where a student speaks and receives immediate feedback.
语音用户界面 (VUI) 允许用户通过语音命令与计算机交互。这项技术集成了语音识别和合成功能，以创造对话式体验。示例包括 Siri、Alexa 和 Google 助理。在教育领域，VUI 可用于交互式语言练习，学生说话后能立即获得反馈。
Designing an effective VUI requires careful consideration of wake words, intent recognition, and error handling. For a listening and speaking preparation tool, the system must accurately understand a learner's utterance even if pronunciation is imperfect, and it must provide corrective feedback without interrupting the flow of conversation.
设计有效的 VUI 需要仔细考虑唤醒词、意图识别和错误处理。对于听力和口语备考工具而言，系统必须准确理解学习者的发音，即使发音并不完美，同时提供纠正性反馈而不打断对话流程。
8. Natural Language Processing in Exam Tools | 考试工具中的自然语言处理
Natural Language Processing (NLP) is the branch of artificial intelligence that deals with the interaction between computers and human language. In the context of oral and listening exams, NLP powers automatic scoring of speaking tests, content analysis of spoken responses, and comprehension checking of listening passages.
自然语言处理 (NLP) 是人工智能的一个分支，处理计算机与人类语言之间的交互。在口语和听力考试中，NLP 支持口语测试的自动评分、口头回答的内容分析以及听力段落的理解检查。
Techniques such as sentiment analysis, keyword spotting, and syntactic parsing allow a system to assess whether a candidate has understood a piece of audio. For instance, after listening to a short lecture, a student might answer questions aloud; NLP then compares the spoken answer against model responses using metrics like word error rate and semantic similarity.
情感分析、关键词识别和句法解析等技术让系统能够评估考生是否理解了某段音频。例如，听完一段简短的讲座后，学生可能大声回答问题；NLP 随后使用词错率和语义相似度等指标，将口头回答与标准答案进行比较。
9. Listening Test Design and Accessibility | 听力测试设计与无障碍性
Computer science principles enable the design of fair and accessible listening tests. Digital audio allows precise control over playback speed, volume, and repetition. For candidates with hearing impairments, alternative visual representations such as waveforms or captioning can be provided, ensuring inclusivity.
计算机科学原理使得设计公平且可访问的听力测试成为可能。数字音频允许精确控制播放速度、音量和重复次数。对于有听力障碍的考生，可以提供波形图或字幕等替代视觉表示，确保包容性。
Exam boards often specify that listening materials must be delivered in a lossless format to prevent compression artifacts from affecting comprehension. The Nyquist-Shannon sampling theorem is fundamental here: to accurately reproduce a sound, the sampling rate must be at least twice the highest frequency present. For speech covering up to 4 kHz, a minimum sampling rate of 8 kHz is mathematically sufficient.
考试机构通常规定听力材料必须以无损格式提供，以防止压缩失真影响理解。这里奈奎斯特-香农采样定理至关重要：为了准确重现声音，采样率必须至少是最高频率的两倍。对于最高频率为 4 kHz 的语音而言，8 kHz 的采样率在数学上就已经足够。
10. Bandwidth and Streaming for Online Exams | 在线考试的带宽与流媒体
As many oral and listening assessments move online, bandwidth management becomes a critical factor. Streaming high-quality audio requires a stable internet connection. Buffering techniques store a small portion of audio ahead of playback to prevent interruptions. Understanding how codecs and adaptive bitrate streaming work helps explain why audio quality might fluctuate during an online test.
随着许多口语和听力评估转移到线上，带宽管理成为一个关键因素。传输高质量音频需要稳定的互联网连接。缓冲技术会预先存储一小部分音频以防止播放中断。了解编解码器和自适应比特率流媒体的工作原理，有助于解释为什么在线测试中音频质量可能会有所波动。
A typical audio stream at 128 kbps uses only about 0.0156 MB per second. This seems small, but a class of 20 students streaming simultaneously can place significant demand on a school network. Buffering algorithms, such as those using a circular queue data structure, keep a few seconds of audio in memory so that momentary network drops do not disrupt the exam experience.
典型的 128 kbps 音频流每秒仅使用约 0.0156 MB。这看起来很小，但一个 20 名学生的班级同时进行流媒体传输可能会对学校网络造成巨大压力。缓冲算法（例如使用循环队列数据结构的算法）在内存中保留几秒钟的音频，以便短暂的网络中断不会破坏考试体验。
11. Exam-Style Questions and Tips | 考试题型与备考技巧
To prepare for the SQA Computing Science exam, you should be ready to answer questions connecting sound concepts to real-world oral and listening scenarios. Here are some example questions:
为了准备 SQA 计算机科学考试，你应该能够回答将声音概念与真实的口语和听力场景联系起来的问题。以下是一些例题：
- Explain why a higher sampling rate improves the clarity of speech recordings. A higher sampling rate captures higher frequency components, preserving consonant sounds that are essential for intelligibility.
- 解释为什么更高的采样率能提高语音录音的清晰度。 更高的采样率能捕捉到更高的频率分量，从而保留对可懂度至关重要的辅音。
- Calculate the uncompressed file size of a 2-minute, mono, 16-bit, 22.05 kHz audio recording of a candidate's spoken response. File size = 22050 × 16 × 120 × 1 = 42,336,000 bits = 5,292,000 bytes ≈ 5.05 MB.
- 计算一位考生口语回答的 2 分钟、单声道、16 位、22.05 kHz 音频的未压缩文件大小。 文件大小 = 22050 × 16 × 120 × 1 = 42,336,000 bit = 5,292,000 字节 ≈ 5.05 MB。
- Describe how lossless compression differs from lossy compression, and state which is more suitable for archiving listening test recordings. Lossless compression reduces file size without any loss of data, allowing perfect reconstruction; it is more suitable for archival purposes to maintain exam integrity.
- 描述无损压缩与有损压缩的区别，并说明哪种更适合存档听力测试录音。 无损压缩在不丢失任何数据的情况下减小文件大小，可以完美重建原始音频；它更适合存档用途，以保持考试的完整性。
When revising, practise converting between bytes, kilobytes, and megabytes, and always check the units in questions. Mind the difference between bits (b) and bytes (B). A typical exam trap involves using bits-per-second for streaming rates but bytes for file sizes.
复习时，请练习比特、千字节和兆字节之间的转换，并务必检查题目中的单位。注意比特 (b) 和字节 (B) 之间的区别。典型的考试陷阱是在流媒体速率中使用比特每秒，而在文件大小中使用字节。
12. Revision Checklist and Final Thoughts | 复习清单与总结
Use this checklist to ensure you have covered the essential topics for the oral/listening section of your SQA Computing Science exam:
使用这份清单来确保你已经掌握了 SQA 计算机科学考试口语/听力部分的核心主题：
- I can explain how sound is digitised using sampling rate and bit depth. / 我能解释如何使用采样率和位深度将声音数字化。
- I can calculate uncompressed audio file sizes. / 我能计算未压缩音频文件的大小。
- I understand the difference between lossy and lossless compression for audio. / 我理解有损和无损音频压缩的区别。
- I can describe how speech recognition and text-to-speech systems work. / 我能描述语音识别和文本转语音系统的工作原理。
- I know how NLP supports automated scoring of spoken answers. / 我知道 NLP 如何支持口语答案的自动评分。
- I am aware of the importance of appropriate file formats and bandwidth for online listening tests. / 我意识到在线听力测试中适当文件格式和带宽的重要性。
By linking computer science fundamentals with practical oral and listening exam scenarios, you build a deeper understanding that will serve you well both in tests and in designing future voice-enabled technologies. Remember to practise with past paper questions that combine calculations and theory.
通过将计算机科学基础与实际的听力和口语考试场景联系起来，你能建立起更深层次的理解，这不仅对考试有益，也对将来设计语音技术有帮助。请记得练习那些结合计算和理论的历年真题。
Published by TutorHao | Computing Science Revision Series | aleveler.com
更多咨询请联系16621398022（同微信）
或截图后用「扫一扫 → 相册」选择本图