Smartphone Recordings are Comparable to "Gold Standard" Recordings for Acoustic Measurements of Voice.
Shaheen N Awan, Mohsin Ahmed Shaikh, Jordan A Awan and 3 others
PMID 37019804WHAT IT FOUND
Modern iPhone and Samsung phone recordings captured the same voice quality differences as a research microphone.
Scores can shift with distance and room noise. Keep distance and room consistent, or convert phone scores before comparing them.
Key findings
01Smartphone recordings of CPP, L/H ratio and CSID correlated very strongly with the sound level meter standard (r's > .90) across vowel and sentence samples.
02Device type had a small effect on CPP in vowels and no significant effect in sentences, but a large effect on L/H ratio and on CSID in sentences.
03A quiet office-like room reduced CPP by approx. 1.5 dB and increased CSID by approx. 7 to 9 severity units; moving from 15 cm to 30 cm reduced CPP by approx. 0.4 dB and increased CSID by approx. 2 to 4 severity units.
STILL TO COME
How it was doneWhat they foundWhat it means for SLPs
Read the rest of this summary
You get three full summaries a month, free, and we do not ask for a card. Search, the TL;DRs and your library stay unlimited either way.
What it does not show
The study used recorded voice samples played through a simulator, not patients holding phones in real clinical or home settings. Only four phone models were tested, one device per model, so results may not apply to all phones. The quiet room was one office-like room; other rooms may have different background noise. Only ADSV spectral and cepstral measures were examined, not fundamental frequency or other common voice measures. The study did not test clinical outcomes, treatment effects, or test-retest variability in patients. Recorded signal level was about -18 dB FS, which the authors note is less desirable than -12 dB FS.
The easy way to misread this
Do not conclude that a smartphone recording can replace a research microphone without controls. Absolute scores differed by device, distance and room, especially for L/H ratio and sentence CSID, so raw scores from different setups are not interchangeable.