Microsoft AI-102 (Designing and Implementing a Microsoft Azure AI Solution) Practice Test 43

Question 420

In a voice-enabled field-service application, engineers must detect a configured wake phrase before starting a longer speech-recognition interaction. The expected evidence is that the wake phrase triggers the subsequent voice-assistant operation. Which implementation is the best match?

  1. Use batch transcription
  2. Use Speech keyword recognition
  3. Use pronunciation assessment
  4. Train a custom speech model

Explanation

Correct Answer

B:

Keyword spotting detects a short activation phrase in streaming audio and can gate more extensive recognition. For this voice-enabled field-service application, the intended evidence is that the wake phrase triggers the subsequent voice-assistant operation. It does not by itself transcribe an entire dictation or identify a specific person from their voice.

Incorrect Answers

A:

Batch transcription submits recorded audio files for managed offline speech recognition at scale; it may be used to process a backlog of stored call recordings asynchronously into transcript artifacts in a voice-enabled field-service application, but the current requirement is instead to detect a configured wake phrase before starting a longer speech-recognition interaction. That separate capability does not prove that the wake phrase triggers the subsequent voice-assistant operation; It does not provide an immediate low-latency result for an active voice conversation, so selecting that alternative would not meet the stated acceptance check.

C:

Pronunciation assessment evaluates speech against expected reference content and exposes relevant feedback metrics; it may be used to score a learner's spoken reading against the supplied pronunciation task in a voice-enabled field-service application, but the current requirement is instead to detect a configured wake phrase before starting a longer speech-recognition interaction. That separate capability does not prove that the wake phrase triggers the subsequent voice-assistant operation; A standard transcript alone does not score pronunciation quality, so selecting that alternative would not meet the stated acceptance check.

D:

Custom Speech can adapt recognition using relevant text and, where applicable, audio with reference transcripts; it may be used to improve recognition of specialized equipment identifiers or unusual vocabulary in a voice-enabled field-service application, but the current requirement is instead to detect a configured wake phrase before starting a longer speech-recognition interaction. That separate capability does not prove that the wake phrase triggers the subsequent voice-assistant operation; It does not translate recognized English into an unrelated target language automatically, so selecting that alternative would not meet the stated acceptance check.

Question 421

A navigation assistant receives a string of directions but the mobile player has no audio to render. Which output step is required?

  1. Use a supported SSML break element with an appropriate duration
  2. Select the supported synthesis output format before requesting new audio
  3. Select a supported Spanish locale and compatible neural voice for synthesis
  4. Send the text through supported speech synthesis and play the returned audio

Explanation

Correct Answer

D:

Text-to-speech converts textual instructions into an audio signal; plain text generation alone creates no playable bytes. In the hands-free multilingual service desk, confirm the fix by retesting the affected behavior after sending the text through supported speech synthesis and play the returned audio.

Incorrect Answers

A:

That change addresses another failure mode: sSML controls speech rendering, including pauses, without inserting literal silence words into the script; it does not address this case, for which the relevant intervention is to send the text through supported speech synthesis and play the returned audio.

B:

The synthesis response format is determined during generation; changing client playback metadata cannot transcode bytes already returned in its own workflow, but that is a different correction from the present requirement to send the text through supported speech synthesis and play the returned audio. The reported issue would remain.

C:

Although useful elsewhere in the hands-free multilingual service desk, a voice's language and locale capabilities affect pronunciation; changing the text alone may not select a suitable voice; the present case is different and calls for the team to send the text through supported speech synthesis and play the returned audio.

Question 422

A team assumes speech translation guarantees synthetic speech in every language that text translation supports. What must be checked?

  1. Validate the returned audio container, codec, and sampling configuration on the device
  2. Verify the selected speech translation API's text and audio output capabilities separately
  3. Check voice availability and language support for the selected Speech region
  4. Manage synthesis request lifecycles and audio playback queuing explicitly

Explanation

Correct Answer

B:

Availability of translated text does not imply every language pair supports spoken target audio. In the hands-free multilingual service desk, confirm the fix by retesting the affected behavior after verifying the selected speech translation API's text and audio output capabilities separately.

Incorrect Answers

A:

Audio compatibility is an integration contract and can break despite successful synthesis requests in its own workflow, but that is a different correction from the present requirement to verify the selected speech translation API's text and audio output capabilities separately. The reported issue would remain.

C:

That change addresses another failure mode: not all voices and capabilities are available in every offering or region; it does not address this case, for which the relevant intervention is to verify the selected speech translation API's text and audio output capabilities separately.

D:

Although useful elsewhere in the hands-free multilingual service desk, duplicate playback can be a client orchestration issue rather than a voice-model defect; the present case is different and calls for the team to verify the selected speech translation API's text and audio output capabilities separately.

Question 423

An embedded device cannot decode a new synthesis format after an SDK update. What should be regression-tested?

  1. Use supported SSML pronunciation or phoneme controls for the code
  2. Validate the returned audio container, codec, and sampling configuration on the device
  3. Manage synthesis request lifecycles and audio playback queuing explicitly
  4. Check voice availability and language support for the selected Speech region

Explanation

Correct Answer

B:

Audio compatibility is an integration contract and can break despite successful synthesis requests. In the hands-free multilingual service desk, confirm the fix by retesting the affected behavior after validating the returned audio container, codec, and sampling configuration on the device.

Incorrect Answers

A:

Although useful elsewhere in the hands-free multilingual service desk, sSML can clarify specialized pronunciation where the selected voice supports the needed elements; the present case is different and calls for the team to validate the returned audio container, codec, and sampling configuration on the device.

C:

Duplicate playback can be a client orchestration issue rather than a voice-model defect in its own workflow, but that is a different correction from the present requirement to validate the returned audio container, codec, and sampling configuration on the device. The reported issue would remain.

D:

That change addresses another failure mode: not all voices and capabilities are available in every offering or region; it does not address this case, for which the relevant intervention is to validate the returned audio container, codec, and sampling configuration on the device.

Question 424

The solution brief for a hands-free multilingual service desk requires the system to deliver synthesized speech in a codec and sample rate compatible with the client. A successful test confirms that the client can decode and play the returned audio stream. What should the team use?

  1. Select an appropriate neural voice
  2. Use text-to-speech synthesis
  3. Configure speech audio output format
  4. Use speech translation

Explanation

Correct Answer

C:

Output format selection controls supported synthesized-audio encoding needed for downstream players or telephony. For this hands-free multilingual service desk, the intended evidence is that the client can decode and play the returned audio stream. Changing format cannot improve an incorrect language model or domain vocabulary.

Incorrect Answers

A:

Voice selection determines the available voice identity, locale, and supported synthesis properties; it may be used to choose a supported speaking voice and language for an accessible user experience in a hands-free multilingual service desk, but the current requirement is instead to deliver synthesized speech in a codec and sample rate compatible with the client. That separate capability does not prove that the client can decode and play the returned audio stream; A voice choice does not fix an upstream speech-recognition transcription error, so selecting that alternative would not meet the stated acceptance check.

B:

Speech synthesis renders text as spoken audio using a chosen supported voice and output format; it may be used to convert application-generated written instructions into playable audio in a hands-free multilingual service desk, but the current requirement is instead to deliver synthesized speech in a codec and sample rate compatible with the client. That separate capability does not prove that the client can decode and play the returned audio stream; It does not recognize a customer's microphone input or extract entities from text, so selecting that alternative would not meet the stated acceptance check.

D:

Speech translation combines speech recognition and translation for specified language configurations; it may be used to recognize incoming speech and produce translated text or supported translated speech in a hands-free multilingual service desk, but the current requirement is instead to deliver synthesized speech in a codec and sample rate compatible with the client. That separate capability does not prove that the client can decode and play the returned audio stream; Plain text-to-speech synthesis has no source-audio understanding, so selecting that alternative would not meet the stated acceptance check.

Question 425

Voice playback works on desktop but produces silence on mobile after a format change. What diagnostic is most relevant?

  1. Configure supported source-language identification or the correct input speech locale
  2. Pin a supported voice identifier and test voice changes before rollout
  3. Compare the returned media format against mobile decoder support
  4. Select a supported compressed audio output format suitable for the playback channel

Explanation

Correct Answer

C:

Platform codec support varies; a valid synthesis stream may still be unplayable by a particular client. In the hands-free multilingual service desk, confirm the fix by retesting the affected behavior after comparing the returned media format against mobile decoder support.

Incorrect Answers

A:

Although useful elsewhere in the hands-free multilingual service desk, translation quality depends on accurate source-speech recognition before target rendering; the present case is different and calls for the team to compare the returned media format against mobile decoder support.

B:

That change addresses another failure mode: voice choice affects audible identity; uncontrolled defaults may shift pronunciation and style; it does not address this case, for which the relevant intervention is to compare the returned media format against mobile decoder support.

D:

Synthesis quality and file size depend on the output encoding; text normalization cannot replace audio-format selection in its own workflow, but that is a different correction from the present requirement to compare the returned media format against mobile decoder support. The reported issue would remain.

Question 426

Users complain that the synthesized voice is in the wrong language even though the script is Spanish. What setting should be verified?

  1. Request supported speech translation into French during recognition
  2. Select a supported Spanish locale and compatible neural voice for synthesis
  3. Send the text through supported speech synthesis and play the returned audio
  4. Select the supported synthesis output format before requesting new audio

Explanation

Correct Answer

B:

A voice's language and locale capabilities affect pronunciation; changing the text alone may not select a suitable voice. In the hands-free multilingual service desk, confirm the fix by retesting the affected behavior after selecting a supported Spanish locale and compatible neural voice for synthesis.

Incorrect Answers

A:

Speech translation can map spoken input into target-language text as part of the recognition workflow in its own workflow, but that is a different correction from the present requirement to select a supported Spanish locale and compatible neural voice for synthesis. The reported issue would remain.

C:

Although useful elsewhere in the hands-free multilingual service desk, text-to-speech converts textual instructions into an audio signal; plain text generation alone creates no playable bytes; the present case is different and calls for the team to select a supported Spanish locale and compatible neural voice for synthesis.

D:

That change addresses another failure mode: the synthesis response format is determined during generation; changing client playback metadata cannot transcode bytes already returned; it does not address this case, for which the relevant intervention is to select a supported Spanish locale and compatible neural voice for synthesis.

Question 427

A synthesizer speaks the same announcement twice because the client starts two overlapping requests. What application behavior should be fixed?

  1. Check voice availability and language support for the selected Speech region
  2. Verify the selected speech translation API's text and audio output capabilities separately
  3. Validate the returned audio container, codec, and sampling configuration on the device
  4. Manage synthesis request lifecycles and audio playback queuing explicitly

Explanation

Correct Answer

D:

Duplicate playback can be a client orchestration issue rather than a voice-model defect. In the hands-free multilingual service desk, confirm the fix by retesting the affected behavior after managing synthesis request lifecycles and audio playback queuing explicitly.

Incorrect Answers

A:

Not all voices and capabilities are available in every offering or region in its own workflow, but that is a different correction from the present requirement to manage synthesis request lifecycles and audio playback queuing explicitly. The reported issue would remain.

B:

Although useful elsewhere in the hands-free multilingual service desk, availability of translated text does not imply every language pair supports spoken target audio; the present case is different and calls for the team to manage synthesis request lifecycles and audio playback queuing explicitly.

C:

That change addresses another failure mode: audio compatibility is an integration contract and can break despite successful synthesis requests; it does not address this case, for which the relevant intervention is to manage synthesis request lifecycles and audio playback queuing explicitly.

Question 428

A team responsible for a hands-free multilingual service desk needs to convert application-generated written instructions into playable audio; the acceptance check is that the application can play audio corresponding to the supplied text. Which capability should be selected?

  1. Select an appropriate neural voice
  2. Use text-to-speech synthesis
  3. Configure speech audio output format
  4. Use speech translation

Explanation

Correct Answer

B:

Speech synthesis renders text as spoken audio using a chosen supported voice and output format. For this hands-free multilingual service desk, the intended evidence is that the application can play audio corresponding to the supplied text. It does not recognize a customer's microphone input or extract entities from text.

Incorrect Answers

A:

Voice selection determines the available voice identity, locale, and supported synthesis properties; it may be used to choose a supported speaking voice and language for an accessible user experience in a hands-free multilingual service desk, but the current requirement is instead to convert application-generated written instructions into playable audio. That separate capability does not prove that the application can play audio corresponding to the supplied text; A voice choice does not fix an upstream speech-recognition transcription error, so selecting that alternative would not meet the stated acceptance check.

C:

Output format selection controls supported synthesized-audio encoding needed for downstream players or telephony; it may be used to deliver synthesized speech in a codec and sample rate compatible with the client in a hands-free multilingual service desk, but the current requirement is instead to convert application-generated written instructions into playable audio. That separate capability does not prove that the application can play audio corresponding to the supplied text; Changing format cannot improve an incorrect language model or domain vocabulary, so selecting that alternative would not meet the stated acceptance check.

D:

Speech translation combines speech recognition and translation for specified language configurations; it may be used to recognize incoming speech and produce translated text or supported translated speech in a hands-free multilingual service desk, but the current requirement is instead to convert application-generated written instructions into playable audio. That separate capability does not prove that the application can play audio corresponding to the supplied text; Plain text-to-speech synthesis has no source-audio understanding, so selecting that alternative would not meet the stated acceptance check.

Question 429

An assistant uses a voice unavailable in its intended deployment region. What selection criterion was missed?

  1. Manage synthesis request lifecycles and audio playback queuing explicitly
  2. Use supported SSML pronunciation or phoneme controls for the code
  3. Check voice availability and language support for the selected Speech region
  4. Validate the returned audio container, codec, and sampling configuration on the device

Explanation

Correct Answer

C:

Not all voices and capabilities are available in every offering or region. In the hands-free multilingual service desk, confirm the fix by retesting the affected behavior after checking voice availability and language support for the selected Speech region.

Incorrect Answers

A:

That change addresses another failure mode: duplicate playback can be a client orchestration issue rather than a voice-model defect; it does not address this case, for which the relevant intervention is to check voice availability and language support for the selected Speech region.

B:

Although useful elsewhere in the hands-free multilingual service desk, sSML can clarify specialized pronunciation where the selected voice supports the needed elements; the present case is different and calls for the team to check voice availability and language support for the selected Speech region.

D:

Audio compatibility is an integration contract and can break despite successful synthesis requests in its own workflow, but that is a different correction from the present requirement to check voice availability and language support for the selected Speech region. The reported issue would remain.

Question 430

In a hands-free multilingual service desk, engineers must choose a supported speaking voice and language for an accessible user experience. The expected evidence is that generated audio uses the selected language and voice. Which implementation is the best match?

  1. Add SSML prosody controls
  2. Configure speech audio output format
  3. Use text-to-speech synthesis
  4. Select an appropriate neural voice

Explanation

Correct Answer

D:

Voice selection determines the available voice identity, locale, and supported synthesis properties. For this hands-free multilingual service desk, the intended evidence is that generated audio uses the selected language and voice. A voice choice does not fix an upstream speech-recognition transcription error.

Incorrect Answers

A:

Speech Synthesis Markup Language provides structured controls for supported voice properties and pronunciations; it may be used to adjust speaking rate, pauses, and other supported rendering attributes of synthesized speech in a hands-free multilingual service desk, but the current requirement is instead to choose a supported speaking voice and language for an accessible user experience. That separate capability does not prove that generated audio uses the selected language and voice; SSML does not retrain a recognition model or translate the underlying meaning of text, so selecting that alternative would not meet the stated acceptance check.

B:

Output format selection controls supported synthesized-audio encoding needed for downstream players or telephony; it may be used to deliver synthesized speech in a codec and sample rate compatible with the client in a hands-free multilingual service desk, but the current requirement is instead to choose a supported speaking voice and language for an accessible user experience. That separate capability does not prove that generated audio uses the selected language and voice; Changing format cannot improve an incorrect language model or domain vocabulary, so selecting that alternative would not meet the stated acceptance check.

C:

Speech synthesis renders text as spoken audio using a chosen supported voice and output format; it may be used to convert application-generated written instructions into playable audio in a hands-free multilingual service desk, but the current requirement is instead to choose a supported speaking voice and language for an accessible user experience. That separate capability does not prove that generated audio uses the selected language and voice; It does not recognize a customer's microphone input or extract entities from text, so selecting that alternative would not meet the stated acceptance check.

Designing and Implementing a Microsoft Azure AI Solution (AI-102) Practice Tests