Back to studio application
Studio Test Spec
TTS-Grade Studio Test Spec
Record a single test clip following the spec below and upload it as one .wav file. The Onyx production team uses this clip to evaluate whether your studio meets the delivery standard for TTS / AI voice-data projects.
Recording spec & environment
- Sample rate 48 kHz / 24-bit / mono channel
- Line-in direct recording
- No processing — no EQ, compression, plug-ins, or noise reduction
- Turn off air conditioner, mobile phone, and any other noise sources
- Verify sound card driver and export settings are correct
Test content — three parts, combined into one file
A
Speech segment
- Any content, read naturally
- Pause ≥ 2 seconds between sentences
- Target speech volume: -9 ~ -3 dB
B
Room tone
- 10–15 seconds of room ambience
- Speaker leaves the room — no voice
- Same gain and settings as the speech segment
C
Clap test (reverb check)
- About 20 cm from the microphone
- At least 3 clear claps
- About 2 seconds between claps
- Avoid clipping
Delivery spec
- Single .wav file (A + B + C combined as one track)
- 48 kHz / 24-bit / mono PCM
- Upload to Drive / cloud storage and paste the URL in the application form
Troubleshooting
No high frequencies / high freq cut off
Fix
- Check export format — confirm 48 kHz / 24-bit / mono PCM .wav
- Check sound card driver settings; install if missing (matching your sound card model)
Background noise obvious
Fix
- Move speaker closer to mic to improve SNR
- Use absorption boards to isolate mic from speaker
- Locate the noise source (AC vent, phone, etc.) and retest after removal
Reverberation obvious
Fix
- Add absorption boards between mic and speaker
Linear noise (horizontal lines on spectrogram)
Fix
- Check and replace equipment cables
- Restart equipment; turn off all electronics (lights, AC, etc.) and retest