📚 Year 10 CAIE Computer Science: Speaking and Listening Exam Prep | Year 10 CAIE 计算机：口语/听力备考专项
Although the CAIE IGCSE Computer Science exam does not have a spoken or listening test, concepts related to speaking (output) and listening (input) appear throughout the syllabus in the form of audio data representation, speech synthesis, and voice recognition. This revision guide focuses on exactly those topics: how computers capture, store, and reproduce sound, how they ‘speak’ to us, and how they ‘listen’ to our commands. Mastering the theory and calculations behind digital audio will prepare you for any related questions in Paper 1 and help you understand practical applications like voice assistants and telephony.
尽管 CAIE IGCSE 计算机科学考试没有口语或听力测试，但与“说话”（输出）和“听力”（输入）相关的概念贯穿整个考纲，表现为音频数据表示、语音合成和语音识别。本复习专项聚焦这些主题：计算机如何捕捉、存储和重现声音，如何对我们“说话”，又如何“倾听”我们的指令。掌握数字音频背后的理论与计算，不仅能帮你应对 Paper 1 中的相关试题，还能让你理解语音助手、电话等实际应用。
1. Sound in the World of Computing | 计算机世界中的声音
Sound is a continuous analogue wave created by vibrations in the air. Computers, however, are digital machines that can only process discrete values. In order for a computer to work with sound—whether recording a voice note or playing back a podcast—the analogue sound wave must be converted into a stream of digital data. This process is known as digitisation, and it consists of two main steps: sampling and quantisation. Understanding these steps is fundamental for answering exam questions about how sound is stored and transmitted.
声音是由空气振动产生的连续模拟波。然而计算机是只能处理离散数值的数字机器。为了让计算机处理声音——无论是录制语音笔记还是播放播客——模拟声波必须被转换成数字数据流。这个过程叫作数字化，主要包含两个步骤：采样和量化。理解这些步骤是回答关于声音存储和传输考题的基础。
2. Analogue to Digital: Sampling Demystified | 模拟转数字：采样揭秘
Sampling means measuring the amplitude (height) of the analogue sound wave at regular time intervals. The number of samples taken per second is called the sample rate, measured in hertz (Hz). A higher sample rate captures the sound more accurately because it records more detail from the original waveform. For instance, CD-quality audio uses a sample rate of 44.1 kHz, meaning 44,100 samples are taken every second. If the sample rate is too low, an effect called aliasing occurs, where high-frequency sounds are misrepresented, leading to distortion.
采样是指以固定的时间间隔测量模拟声波的振幅（高度）。每秒采集的样本数量叫作采样率，单位是赫兹（Hz）。采样率越高，能够捕捉到的原始波形细节就越多，声音也就越准确。例如，CD 音质使用的采样率为 44.1 kHz，即每秒采集 44,100 个样本。如果采样率过低，会发生混叠效应（aliasing），高频声音被错误表示，从而导致失真。
According to the Nyquist theorem, the sample rate must be at least twice the highest analogue frequency to reconstruct the signal perfectly. Human hearing ranges from about 20 Hz to 20 kHz, so a sample rate of 40 kHz or more ensures faithful reproduction. This is why 44.1 kHz and 48 kHz are common standards.
根据奈奎斯特定理，为了完美重建信号，采样率至少应是最高模拟频率的两倍。人耳听觉范围大约在 20 Hz 至 20 kHz，因此 40 kHz 或更高的采样率可以保证忠实还原。这也是 44.1 kHz 和 48 kHz 成为常用标准的原因。
3. Quantisation and Bit Depth | 量化与比特深度
After sampling, each measured amplitude must be assigned a discrete digital value. This step is called quantisation. The bit depth (also known as sample size) determines how many bits are used to store each sample. A higher bit depth offers more possible amplitude levels, resulting in finer volume steps and less quantisation noise. For example, 16-bit audio can represent 65,536 different amplitude levels, while 24-bit audio can capture over 16 million levels, which greatly improves dynamic range and clarity.
采样完成后，每个测得的振幅必须被赋予一个离散数字值。这一步叫作量化。比特深度（也称样本大小）决定了用多少比特存储每个样本。比特深度越高，可表示的振幅级数就越多，音量阶梯越细腻，量化噪声也越小。例如，16 位音频可表现 65,536 种振幅级别，而 24 位音频能捕捉超过 1600 万个级别，从而极大提升动态范围和清晰度。
It is important not to confuse sample rate with bit depth. Sample rate affects frequency representation (how well high pitches are captured), whereas bit depth affects volume accuracy and background noise. Both contribute to overall sound quality and file size; exam questions often ask you to distinguish between the two.
切勿混淆采样率和比特深度。采样率影响频率的再现（即高音能被多好地捕捉），而比特深度影响音量精度和背景噪声。两者共同决定整体音质和文件大小；考试中常要求区分这两个概念。
4. Bit Rate and Why It Matters | 比特率及其重要性
Bit rate is the amount of digital data processed per second during playback or recording, usually expressed in kilobits per second (kbps). It can be calculated using the formula:
Bit rate (bps) = Sample rate (Hz) × Bit depth (bits) × Number of channels. For example, a stereo CD track (44,100 Hz, 16-bit, 2 channels) has a bit rate of 44,100 × 16 × 2 = 1,411,200 bps, or approximately 1,411 kbps. This value directly influences the balance between audio fidelity and file size.
比特率是播放或录制过程中每秒处理的数字数据量，通常以千比特每秒（kbps）表示。它可以用以下公式计算：
比特率 (bps) = 采样率 (Hz) × 比特深度 (bits) × 声道数。例如，立体声 CD 音轨（44,100 Hz，16 位，2 声道）的比特率为 44,100 × 16 × 2 = 1,411,200 bps，约 1,411 kbps。这个数值直接影响音质保真度与文件大小之间的平衡。
Understanding bit rate is crucial for tackling exam-style calculation questions. You may be asked to compare uncompressed bit rates or to explain why a lower bit rate reduces file size at the expense of quality. Musicians and streaming services constantly make this trade-off.
理解比特率对处理考试中的计算题至关重要。你可能会被要求比较未压缩比特率，或者解释为什么较低比特率会减小文件但牺牲音质。音乐人和流媒体服务无时无刻不在做这种权衡。
5. File Size Calculation Step by Step | 文件大小逐步计算
To estimate the size of an uncompressed audio file, we extend the bit rate formula over time. The general equation is:
File size (bytes) = Duration (s) × Sample rate (Hz) × Bit depth (bits) × Channels ÷ 8. Remember that 8 bits make 1 byte, and 1024 bytes make a kilobyte (KB), 1024 KB make a megabyte (MB). In CAIE exams, sometimes 1 kB = 1000 bytes is used; always read the question carefully.
要估算未压缩音频文件的大小，我们将比特率公式扩展到时间上。通用等式为：
文件大小 (字节) = 时长 (秒) × 采样率 (Hz) × 比特深度 (位) × 声道数 ÷ 8。请记住 8 位为 1 字节，1024 字节为 1 千字节 (KB)，1024 KB 为 1 兆字节 (MB)。在 CAIE 考试中，有时会使用 1 kB = 1000 字节，务必仔细读题。
Worked Example: Calculate the file size in MB for a 3-minute mono recording with a sample rate of 44.1 kHz and 16-bit depth.
Step 1: Duration in seconds = 3 × 60 = 180 s.
Step 2: File size (bits) = 180 × 44,100 × 16 × 1 = 127,008,000 bits.
Step 3: Divide by 8 to get bytes: 127,008,000 ÷ 8 = 15,876,000 bytes.
Step 4: Convert to MB (using 1 MB = 1,048,576 bytes): 15,876,000 ÷ 1,048,576 ≈ 15.14 MB.
Always show your working to earn full marks.
例题：计算一段 3 分钟、采样率 44.1 kHz、16 位深度的单声道录音的文件大小（MB）。
步骤 1：时长秒数 = 3 × 60 = 180 s。
步骤 2：文件大小（位）= 180 × 44,100 × 16 × 1 = 127,008,000 位。
步骤 3：除以 8 得到字节数：127,008,000 ÷ 8 = 15,876,000 字节。
步骤 4：换算为 MB（1 MB = 1,048,576 字节）：15,876,000 ÷ 1,048,576 ≈ 15.14 MB。
务必写出中间步骤以获取全部分数。
6. Common Audio File Formats | 常见音频文件格式
Different audio formats use various compression techniques to manage file size and quality. The table below summarises the formats you are likely to encounter in the syllabus:
不同的音频格式采用各种压缩技术来平衡文件大小和品质。下表总结了考纲中可能出现的主要格式：
| Format | Compression | Quality | Typical Use |
|---|---|---|---|
| WAV | Uncompressed | Lossless, high | Professional editing, masters |
| MP3 | Lossy | Good at lower bit rates | Music download, streaming |
| AAC | Lossy | Better than MP3 at same bit rate | Streaming video/audio |
| MIDI | Not audio, instruction set | Depends on playback device | Synthesised music, ringtones |
MIDI (Musical Instrument Digital Interface) stores commands such as note-on, note-off, and instrument selection rather than actual sound waves. This makes MIDI files extremely small and editable, but playback can sound different on various devices. In contrast, WAV stores raw PCM data, making it large but perfect for archiving.
MIDI（乐器数字接口）存储的是音符开、音符关及乐器选择等指令，而非真实声波。这使得 MIDI 文件极小且易于编辑，但播放效果在不同设备上可能不同。而 WAV 存储原始 PCM 数据，因此体积大，但适合存档。
7. Speech Synthesis: How Computers ‘Speak’ | 语音合成：计算机如何“说话”
Speech synthesis, often called text-to-speech (TTS), allows a computer to convert written text into spoken words. This is the ‘speaking’ side of our exam focus. Early systems used concatenative synthesis, piecing together stored diphones or syllables. Modern TTS engines rely on deep learning models that generate incredibly natural speech with realistic intonation. Applications range from screen readers for visually impaired users to virtual assistants reading aloud a weather forecast.
语音合成，通常称为文本转语音 (TTS)，让计算机能将书面文本转换为口语。这正是我们备考专题中“说话”的一面。早期系统采用拼接合成，将存储的双音子或音节拼凑起来。现代 TTS 引擎基于深度学习模型，生成的语音极其自然，语调也逼真。其应用范围从为视障用户服务的屏幕阅读器，到读天气预报的虚拟助手。
In the exam, you may be asked to discuss the advantages of TTS, such as aiding those with reading difficulties, enabling hands-free operation, and delivering dynamic information (e.g., GPS navigation). You should also be aware of limitations: emotional expression can be limited, and unusual words may be mispronounced.
在考试中，你可能会被问到 TTS 的优势，例如辅助阅读困难人群、实现免提操作以及传送动态信息（如 GPS 导航）。同时也要了解其局限：情感表达可能有限，生僻词可能发音错误。
8. Speech Recognition: How Computers ‘Listen’ | 语音识别：计算机如何“倾听”
Speech recognition (or automatic speech recognition, ASR) enables a computer to interpret human speech and turn it into text or commands. This is the ‘listening’ part. The process involves capturing audio via a microphone, digitising the sound, and analysing acoustic patterns using algorithms. Modern ASR systems, like Siri or Alexa, use large neural networks and cloud processing to achieve high accuracy. Phonemes and language models help predict what was said, even with background noise or accents.
语音识别（或自动语音识别，ASR）使计算机能够理解人类语音，并将其转换为文本或指令。这就是“听力”部分。此过程包括通过麦克风捕捉音频、将声音数字化，并使用算法分析声学模式。像 Siri 或 Alexa 这样的现代 ASR 系统利用大型神经网络和云端处理来达到高准确率。音素和语言模型即使在背景噪声或口音下也能帮助预测所说内容。
Key exam points include understanding that speech recognition involves both an acoustic model (matching sounds to phonemes) and a language model (predicting word sequences). Factors affecting accuracy include the training data size, clarity of speech, and microphone quality. Security and privacy concerns are also relevant, as voices can be recorded and used without consent.
考试要点包括理解语音识别涉及声学模型（将声音匹配到音素）和语言模型（预测单词序列）。影响准确度的因素包括训练数据量、语音清晰度和麦克风质量。安全与隐私问题也需注意，因为语音可能被未经同意录制和使用。
9. Audio in Data Compression Context | 数据压缩情境中的音频
Lossless and lossy compression both appear in audio technology. Lossless codecs like FLAC (Free Lossless Audio Codec) reduce file size without discarding any data, similar to ZIP for audio. Lossy codecs such as MP3 and AAC achieve much smaller files by removing sounds that are less audible to the human ear, a process called perceptual coding. For instance, a quiet sound masked by a louder one at a similar frequency may be discarded. The compression ratio can be dramatic: a CD track might be compressed from 30 MB down to 3 MB without major perceived loss in quality.
无损和有损压缩都出现在音频技术中。像 FLAC（自由无损音频编解码器）这样的无损编解码器减小文件体积而不丢弃任何数据，类似于给音频打的 ZIP 包。而有损编解码器如 MP3 和 AAC 通过移除人耳不易察觉的声音（称为感知编码）来获得更小的文件。例如，被相近频率较大声量遮蔽的轻声可能被丢弃。压缩比可以十分惊人：一首 CD 音轨可能从 30 MB 压缩到 3 MB，而主观音质损失不大。
Being able to explain why lossy compression works well for casual listening but not for studio editing demonstrates a thorough understanding. Exam questions sometimes ask you to justify the choice of a lossy format for a streaming service or to identify the data lost in MP3 encoding.
能够解释为什么有损压缩适合日常聆听但不适合工作室编辑，可以展现你对知识的深刻理解。考题有时要求你论证流媒体服务选择有损格式的理由，或识别 MP3 编码中丢失的数据。
10. Practical Calculation Drills | 实践计算题精练
Here are some exam-style problems to test your skill. Try to solve them before looking at the solutions.
以下是若干考试风格的问题，用来检测你的技能。请先尝试解答，再看参考答案。
Q1: A 2-minute stereo recording is made at 48 kHz and 24 bits per sample. What is the uncompressed file size in MB? (Assume 1 MB = 1,000,000 bytes for this exercise.)
A1: (2 × 60) s × 48,000 Hz × 24 bits × 2 channels = 120 s × 48,000 × 24 × 2 = 276,480,000 bits. File size = 276,480,000 ÷ 8 = 34,560,000 bytes = 34.56 MB.
Q2: Which format would you recommend for a digitised speech archive that must preserve original quality? Explain.
A2: WAV or FLAC because they are lossless formats. WAV stores original PCM data, while FLAC offers lossless compression to save space without quality loss. Avoid MP3 as it discards data irreversibly.
Q3: A telephone system uses a sample rate of 8 kHz and 8-bit depth. Estimate the bit rate in kbps.
A3: For mono: bit rate = 8,000 × 8 × 1 = 64,000 bps = 64 kbps. This low bit rate is acceptable for voice but poor for music, which is why phone calls sound different from studio recordings.
通过反复练习这类题目，你将在计算和数据表示大题中更加从容。
11. Linking Speaking and Listening to Smart Devices | 说话与听力在智能设备中的连接
Virtual assistants like Alexa or Google Assistant combine both speech synthesis and speech recognition in a seamless loop. You speak a request (‘listening’ via speech recognition), the assistant processes the command in the cloud, and responds with a synthesised voice (‘speaking’). This integration relies heavily on fast internet, powerful servers, and accurate acoustic and language models. On a local device, such as a smartphone virtual keyboard with voice input, the same principles apply, though some lightweight models run offline.
像 Alexa 或 Google Assistant 这样的虚拟助手将语音合成和语音识别无缝整合进一个循环。你说出请求（通过语音识别“倾听”），助手在云端处理指令，然后用合成语音回应（“说话”）。这种整合严重依赖高速互联网、强大的服务器以及精确的声学与语言模型。在本地设备上，比如带有语音输入功能的智能手机虚拟键盘，同样遵循这些原理，但部分轻量模型可以离线运行。
This convergence is a perfect example of why both topics appear together under the ‘speaking and listening’ umbrella in computing. A solid understanding of these technologies helps you evaluate their reliability, ethical issues, and future potential.
这种融合完美展示了为什么在计算机领域中“说话和听力”总是一起出现。扎实理解这些技术有助于评估其可靠性、伦理问题和未来潜力。
12. Top Tips for the Exam | 考试高分技巧
- Always check the units. Convert minutes to seconds, and know whether the mark scheme uses 1000 or 1024 for MB. If in doubt, state your assumption.
- Recall Nyquist’s rule to justify sample rate choices: sample rate ≥ 2 × highest frequency.
- Differentiate sample rate from bit depth clearly in explanation questions. One is about frequency, the other about amplitude precision.
- Use correct terminology such as ‘sampling’, ‘quantisation’, ‘lossless’, ‘lossy’, and ‘perceptual coding’.
- For speech recognition, mention phonemes and language models to access higher band marks.
- 始终核对单位。将分钟转换为秒，并明确评分标准中采用的 MB 是 1000 还是 1024。如有疑问，写明你的假设。
- 记牢奈奎斯特定理，用于论证采样率选择：采样率 ≥ 2 × 最高频率。
- 在解释题中清晰区分采样率与比特深度。一个与频率有关，另一个与振幅精度有关。
- 使用正确术语，如“采样”“量化”“无损”“有损”“感知编码”。
- 在语音识别中，提到音素和语言模型以争取更高分数。
Published by TutorHao | Computer Science Revision Series | aleveler.com
更多咨询请联系16621398022（同微信）
屏轩国际教育cambridge primary/secondary checkpoint, cat4, ukiset,ukcat,igcse,alevel,PAT,STEP,MAT, ibdp,ap,ssat,sat,sat2课程辅导，国外大学本科硕士研究生博士课程论文辅导Cancel reply