📚 Speech & Audio Exam Prep | Year 10 AQA 计算机：口语/听力备考专项
From voice assistants to digital music, sound is processed by computers as data. In this article, we break down every concept you need to ace AQA Computer Science questions on audio representation, speech technologies, and related hardware. Gain confidence in calculating file sizes, comparing sampling quality, and explaining how your device ‘hears’ and ‘speaks’.
从语音助手到数字音乐，声音在计算机中都被处理成数据。本文将逐一拆解 AQA 计算机科学关于音频表示、语音技术及配套硬件的所有考点，帮助你轻松回答采样频率、声音文件大小计算等问题，并理解设备如何完成“听力”与“口语”。
1. Sound as Data: Analogue vs Digital | 声音即数据：模拟与数字
Microphones capture continuous analogue sound waves produced by your voice or an instrument. The computer cannot store this infinite smooth curve directly; it must convert it into discrete digital data. This conversion process is called analogue-to-digital conversion (ADC).
麦克风捕捉你说话或乐器产生的连续模拟声波。计算机无法直接存储这类无限平滑的波形，必须将其转换为不连续的数字数据，这一过程称为模数转换（ADC）。
When you play a digital sound file, a digital-to-analogue converter (DAC) reconstructs an analogue signal for speakers or headphones. The quality of the reconstruction depends on two key decisions made during sampling: the sampling rate and the bit depth.
当你播放数字声音文件时，数模转换器（DAC）将数字信号重建为模拟信号，驱动扬声器或耳机发声。重建质量取决于采样过程中两个核心选择：采样频率和采样分辨率（位深度）。
2. Sampling Rate and Bit Depth | 采样频率与采样分辨率
Sampling rate defines how many times per second the amplitude of the analogue wave is measured. It is given in hertz (Hz) or kilohertz (kHz). A standard CD uses 44.1 kHz, meaning 44,100 samples are taken each second. According to the Nyquist theorem, the sampling rate must be at least twice the highest frequency you want to capture faithfully.
采样频率规定了每秒钟对模拟波形振幅进行测量的次数，单位为赫兹（Hz）或千赫兹（kHz）。标准 CD 采用 44.1 kHz，即每秒采集 44,100 个样本。根据奈奎斯特定理，采样频率至少要达到想要忠实捕捉的最高音频频率的两倍。
Bit depth (or sampling resolution) determines the number of bits used to store each amplitude measurement. Every bit added doubles the number of possible levels, making the digital approximation closer to the original wave. For example, 16-bit audio can distinguish 65,536 different levels.
采样分辨率（或称位深度）决定了用于存储每个振幅读数的比特数。每增加 1 位，可表示的级数就翻倍，使数字逼近更接近原始波形。例如，16 位音频可以区分 65,536 个不同的振幅等级。
Higher sampling rate and bit depth produce more accurate recordings but dramatically increase file size. You must be able to compare quality visually: a waveform sampled more frequently looks smoother and less jagged when zoomed in.
更高的采样频率和位深度能产生更精确的录音，但也使文件体积显著增大。你必须能从视觉上比较质量：放大后，采样更频繁的波形看起来更平滑，锯齿感更少。
3. Calculating Sound File Sizes | 声音文件大小计算
AQA exams frequently ask you to compute the size of an uncompressed sound file. The formula is: file size (bits) = sampling rate (Hz) × bit depth × duration (seconds) × number of channels. Common conversions: 1 byte = 8 bits; 1 kilobyte (KB) = 1024 bytes.
AQA 考试常要求计算未压缩声音文件的大小。公式为：文件大小（比特）= 采样频率（Hz）× 位深度 × 时长（秒）× 声道数。注意单位换算：1 字节 = 8 比特；1 千字节（KB）= 1024 字节。
Total bits = sample rate (Hz) × bit depth × time (s) × channels
Example: a 30-second stereo recording at 44.1 kHz with 16-bit depth. Stereo means 2 channels. Total bits = 44,100 × 16 × 30 × 2 = 42,336,000 bits. Convert to megabytes: 42,336,000 / 8 = 5,292,000 bytes; 5,292,000 / 1024 = 5,168 KB; 5,168 / 1024 = 5.05 MB (approximately).
示例：一段 30 秒立体声录音，44.1 kHz、16 位。立体声代表 2 声道。总比特数 = 44,100 × 16 × 30 × 2 = 42,336,000 比特。换算：42,336,000 / 8 = 5,292,000 字节；5,292,000 / 1024 ≈ 5,168 KB；再除以 1024 约等于 5.05 MB。
Always show your working step by step. If the question provides a target file size, you may need to rearrange the formula to find an unknown variable such as the maximum possible duration or the bit depth required.
务必逐步展示运算过程。如果题目给出目标文件大小，你可能需要变换公式来求解未知量，比如最大可录音时长或所需的位深度。
4. Sound File Formats and Compression | 声音文件格式与压缩
Uncompressed formats like WAV or AIFF store every sample faithfully, resulting in huge files. Lossless compression (e.g., FLAC) reduces file size without discarding any audio information; the original wave can be perfectly reconstructed.
未压缩格式（如 WAV 或 AIFF）原样存储每一个样本，文件庞大。无损压缩（如 FLAC）可以在不丢弃任何音频信息的前提下缩小体积，原始波形能被完美复原。
Lossy compression (e.g., MP3, AAC) achieves much smaller sizes by removing sounds that humans are unlikely to hear, such as very high frequencies masked by louder tones. Once discarded, these details cannot be restored. This is a trade-off between quality and storage.
有损压缩（如 MP3、AAC）通过删除人耳不易察觉的声音（比如被较强声音掩蔽的高频细节）来实现大幅瘦身。一旦丢弃，这些细节便无法恢复。这是在音质与存储空间之间的折衷。
| Feature / 特性 | Lossy (MP3) / 有损 | Lossless (FLAC) / 无损 |
|---|---|---|
| File size / 文件体积 | Small / 小 | Moderate / 中等 |
| Quality / 音质 | Reduced / 有损 | Identical to original / 与原始相同 |
| Suitable for / 适合 | Streaming, portables / 流媒体、便携 | Archiving, editing / 存档、编辑 |
5. Input and Output Devices for Audio | 音频的输入与输出设备
Your voice enters the computer through an input device: a microphone. Microphones contain a diaphragm that vibrates when hit by sound waves, producing a tiny electrical analogue signal. This signal then goes to an analogue-to-digital converter.
你的声音通过输入设备——麦克风——进入计算机。麦克风内的振膜受声波撞击而振动，产生微弱的模拟电信号，随后送入模数转换器。
For output, computers use speakers or headphones. These receive a digital signal from the sound card, convert it through a DAC, amplify it, and push it to a driver (a magnet and coil system in a speaker) that moves a cone to create sound waves.
输出设备则为扬声器或耳机。它们接收来自声卡的数字信号，经 DAC 转换、放大，再驱动扬声器中的音圈与磁铁系统推动纸盆振动，从而产生声波。
Modern smartphones combine both functions into a single headset, often using Bluetooth to transmit digital audio wirelessly. This requires pairing of devices and introduces codec compression to fit the data into limited bandwidth.
现代智能手机常将两种功能集成在一副耳机中，通常通过蓝牙无线传输数字音频。这需要设备配对，并引入编解码压缩以适应有限的带宽。
6. Speech Recognition: Teaching Computers to ‘Listen’ | 语音识别：教会计算机“听”
Speech recognition converts spoken words into text or computer commands. The system captures audio, breaks it into short frames, extracts acoustic features, and matches them against a large database of known phonemes and language models using machine learning.
语音识别将口述内容转化为文字或计算机指令。系统捕获音频，切成短帧，提取声学特征，再利用机器学习将其与已知音素和语言模型的海量数据库进行匹配。
Applications include virtual assistants (Siri, Alexa), dictation software, in-car systems, and accessibility tools for users with motor impairments. The process relies heavily on cloud servers, raising privacy concerns about recordings being stored and analyzed.
应用场景包括虚拟助手 (Siri、Alexa)、语音输入法、车载系统以及为行动不便者打造的辅助工具。该过程高度依赖云端服务器，引发了录音被存储和分析的隐私担忧。
Accuracy depends on clear pronunciation, low background noise, and the quality of the training data. Homophones (e.g., ‘there’ and ‘their’) remain a challenge that language context must resolve.
识别准确度取决于清晰发音、低背景噪音以及训练数据的质量。同音异义词（如“there”与“their”）仍然是一个需依靠语境解决的挑战。
7. Speech Synthesis: Teaching Computers to ‘Speak’ | 语音合成：教会计算机“说话”
Speech synthesis is the opposite process: converting text into audible speech. A text-to-speech (TTS) engine first analyses the text, applying pronunciation rules and handling abbreviations, then generates a waveform using methods like concatenative synthesis (joining pre-recorded speech fragments) or parametric synthesis (modelling the vocal tract).
语音合成则是相反方向：将文本转换为可听语音。文语转换（TTS）引擎首先分析文本，应用发音规则并处理缩写，然后通过波形拼接（组合预录语音片段）或参数合成（模拟声道模型）生成波形。
Modern neural TTS produces incredibly natural-sounding voices with appropriate intonation and rhythm. It is used in screen readers, navigation systems, and interactive voice response (IVR) phone menus. When combined with speech recognition, it enables fully spoken dialogue systems.
现代神经 TTS 能生成带有恰当地语调节奏、极为自然的语音，用于屏幕阅读器、导航系统和电话交互语音应答。与语音识别结合后，便可实现全语音对话系统。
Ethical implications include the ability to clone someone’s voice from a few samples, enabling deepfake audio. You should be prepared to discuss how this can facilitate fraud or misinformation.
伦理问题包括仅凭少量样本克隆某人声音，制造音频深度伪造。你应能讨论这会如何为诈骗或虚假信息提供便利。
8. MIDI: Representing Musical Performance | MIDI：表示音乐演奏
MIDI (Musical Instrument Digital Interface) represents sound differently from sampled audio. Instead of storing waveforms, it stores instructions like ‘note on’, ‘note off’, pitch, velocity, and instrument. This makes files extremely small.
MIDI（乐器数字接口）采用与采样音频完全不同的方式表示声音。它不存储波形，而是存储“音符开”、“音符关”、音高、力度和乐器等指令，因此文件极为小巧。
A MIDI file is played back by a synthesizer that interprets the instructions and generates sound. This approach is ideal for ringtones, backing tracks, and music education, but it cannot reproduce human speech or arbitrary live sounds. MIDI data is often manipulated using sequencing software.
MIDI 文件通过合成器解读指令并生成声音。这种方式非常适合铃声、伴奏和音乐教育，但无法再现人声或任意现场声音。MIDI 数据通常使用音序器软件进行编辑。
9. Audio Editing and Processing | 音频编辑与处理
Audio editing software, such as Audacity or Adobe Audition, allows users to manipulate sound data directly. Common operations include cut, paste, fade in/out, amplify, and noise reduction. These tools show the waveform visually, helping you identify silent sections or clipping (distortion caused by signal exceeding the maximum amplitude).
音频编辑软件（如 Audacity、Adobe Audition）允许用户直接操作声音数据。常见操作包括剪切、粘贴、淡入淡出、放大和降噪。这些工具以可视化波形方式展示，帮助你识别静音段或削峰失真（信号超出最大振幅导致的畸变）。
When editing, filters can be applied to boost or cut specific frequency ranges. For example, a high-pass filter removes low-frequency rumble, while a low-pass filter softens harsh treble noise. Understanding these helps in tasks like cleaning up voice recordings for podcasts.
编辑时可应用滤波器增强或削弱特定频率范围。例如，高通滤波器能去除低频隆隆声，而低通滤波器可柔化刺耳的高频噪音。理解这些原理有助于完成播客人声录音的降噪任务。
10. Audio in Transmission and Streaming | 传输与流媒体中的音频
Streaming services (Spotify, YouTube) deliver audio over the internet in real time. To cope with varying connection speeds, adaptive bitrate streaming switches between different quality levels automatically. If your bandwidth drops, the server sends a lower bitrate version to prevent buffering.
流媒体服务（Spotify、YouTube）通过互联网实时传输音频。为应对波动的网速，自适应比特率串流会在不同质量等级之间自动切换。一旦带宽下降，服务器会发送较低比特率版本，避免播放卡顿。
Audio on web pages is embedded using the <audio> HTML5 tag, which can play back MP3, WAV, and Ogg files. Understanding codecs and containers is part of the wider digital media knowledge expected at this level.
网页上的音频使用 HTML5 <audio> 标签嵌入，可播放 MP3、WAV 和 Ogg 文件。理解编码与容器是本阶段数字媒体知识的一部分。
Digital audio broadcasts (DAB) use multiplexing to transmit multiple radio stations on the same frequency. Data is compressed with the AAC codec, providing better reception and more stations compared to analogue FM.
数字音频广播（DAB）利用复用技术在同一频率上传送多个电台。数据采用 AAC 编解码压缩，与模拟 FM 相比，接收更清晰、电台数量更多。
11. Multimedia Systems and Synchronisation | 多媒体系统与同步
In multimedia projects, audio must be synchronised with video or animations. Lip sync errors occur when the sound does not match the movement of a speaker’s lips. Computers achieve sync by timestamping audio frames and video frames, then playing them according to a common clock.
在多媒体项目中，音频必须与视频或动画同步。当声音与说话者嘴唇动作不匹配时，就会发生唇音不同步。计算机通过对音频帧和视频帧加时间戳，并根据统一时钟播放，从而实现同步。
When designing revision aids, consider recording yourself explaining algorithms aloud and listening back. This helps reinforce memory and mimics the auditory learning that complements visual diagrams. It is a practical study technique bridging the ‘speaking and listening’ approach with computer science content.
在复习时，不妨把自己的算法讲解录下来再听回放。这能强化记忆，并模拟听觉学习，与视觉图示相辅相成。这就是一种将“口语与听力”技巧与计算机科学内容相结合的学习方法。
12. Legal, Ethical and Environmental Considerations | 法律、伦理与环境考量
Copyright law protects sound recordings and musical compositions. It is illegal to download or share copyrighted music without permission. Streaming services pay royalties to artists based on complex licensing agreements. The Digital Rights Management (DRM) technology prevents unauthorised copying.
版权法保护录音制品和音乐作品。未经许可下载或分享受版权保护的音乐是违法的。流媒体服务通过复杂的授权协议向艺术家支付版税。数字版权管理（DRM）技术则防止未经授权的复制。
Voice data collected by smart speakers is processed on remote servers, raising issues around consent, data storage, and surveillance. The environmental impact of audio streaming includes the energy consumption of data centres that deliver content globally.
智能音箱收集的语音数据在远程服务器上处理，引发了知情同意、数据存储和监控等方面的问题。音频流媒体的环境影响包括向全球分发内容的数据中心所消耗的能源。
As consumers, we can make more sustainable choices by downloading music for offline playback instead of re-streaming frequently, and by using energy-efficient devices. Consider how upgrading audio hardware less often can reduce e-waste.
作为消费者，我们可以通过下载音乐离线播放（而不是反复在线流播）以及使用节能设备来做出更可持续的选择。想一想，减少音频硬件的更换频率可以怎样减少电子垃圾。
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