This was a cross-sectional study, conducted at the Educational Audiology Laboratory in 2011, with approval by the Research Ethics Committee (Protocol 129/2010).

In this study, only individuals who agreed with the procedures necessary for conducting the examination and who signed the informed consent, after having received information on the purpose and methodology of the proposed study, were enlisted.

The adopted inclusion criteria were: age between 18 and 59 years, audiometric thresholds within normal limits,26 and absence of ear wax plug or other changes in the external or middle ear which could modify the test performance.

The sample was selected by convenience and consisted of 79 adult subjects, 29 males and 50 females, aged 19–44 years (mean, 24 years).

The audiological evaluation occurred following an otoscopic examination performed by the Service's otolaryngologist, and consisted of pure tone audiometry by air conduction at frequencies of 250–8000Hz. To obtain these measures, a digital two-channel audiometer (Madsen, model Midimate 622) was used, along with TDH35 supra-auricular headphones. The test was conducted in a soundproof booth.

To assess the in-noise speech perception, the following equipment was used: the HINTPro 7.2 Audiometric System; two stereo loudspeakers for the free field; a computer with CD recorder; printer; and a sound-treated room.

The test was conducted in the free field condition in a sound-treated room, according to the HINTPro 7.2 Audiometric System operating instructions manual. HINTPro consists of equipment with interface connected to the computer, which permits the use of HINT. The installation of the specific software for the test on the computer is required, and the free field stereo loudspeakers were coupled to HINTPro 7.2.

Each list of 20 sentences was applied in the following situations: quiet (Q), noise front (NF), noise right (NR), and noise left (NL) and compound noise (CN). The sentences were presented at 0–0° azimuth, and the level of presentation was initially set at 45dBA, and varied in steps of 4dB and 2dB, according to the correct repetition of the level.

However, the competitive noise was introduced at 0–0°, 0–90°, 0–180°, and 0–270° azimuth, at a fixed intensity of 65dBA.27 The lists of sentences and the order of noise presentation were selected and presented randomly. The score was expressed in dB, representing the S/N ratio threshold.

For data analysis, the paired t-test and two-way ANOVA were used.

There was no significant difference among ears, gender, and interaction among conditions. We did observe a difference among the conditions tested (Table 1).

The subjects of this study reported no difficulty undertaking the test; they understood the instructions and performed properly.

As shown in Table 1, the mean values obtained for the in-noise speech recognition threshold (SRT – HINT Brazil) ranged from −6.47 to −3.20 for different listening situations. The literature contains two studies16,24 that describe the perception of speech in-noise performance expected for adults with normal hearing in the same listening situation – in the free field. The first study16 proposes values approaching those findings of this study for all test conditions, while the other24 found better results; this can be explained by the fact that the characteristics of the speech material and of the individuals who participated in the test can affect intelligibility. These characteristics include the words’ phonetic similarity, the rate of speech and the clarity of the speaker, the naturalness of the speaker's voice, the gender of the subject, and his/her dialect.24,28

The language skills of the listener interact with grammatical and lexical properties of the material (speech). Likewise, the listener's cognitive and memory abilities interact with the complexity and duration of the discourse. In addition, the age at which the language was acquired and the primary or secondary status of the language affect its intelligibility.24 Different speakers who speak in different languages and tones are used to record the HINT sentences.

Therefore, the masking effects on a particular speech sample by spectral noise are unpredictable, contributing to an unwanted source of variability in threshold measures.24 Another factor that can interfere in the results is the calibration for the free field in situ, with different conditions from one test to another. In a study on speech recognition in the free field, and also based on the literature review, it was found that the present results were different from those of other studies,12,16,24 possibly because of variables that can be found in the evaluation of the free field environment, such as room size, acoustic conditions, whether or not a reflective surface exists, as well as reverberation and calibration levels.10 One of these studies was not conducted in Brazil.24

The worst S/N ratio is obtained when the speech and noise are in the same position, explaining the results found for NF. According to the literature, the highest (worst) thresholds occur when speech and noise are presented at the same location, that is, speech and noise at 0° in front of the subject. The results are better when speech and noise are separated by an angle of 90°, with speech at 0° in front of the individual, and the noise to the right or to the left at 90°,24 the best reported results are seen when the noise and speech conditions are separated by 90°, with speech at 0° in front of the assessed individual and the noise at 90° to the right or to the left of the subject,3,16,29 which corroborates the findings of the present study.

The statistical difference between S/N ratios obtained at 0–0° in relation to those obtained with the noise positioned at 0–90° occurs because, when moving the sound from the in-front position to the lateral position in relation to the subject assessed, the sound stimulus presented laterally is not equally perceived by both ears, contrary to what occurs when the sound source is located in front of the subject. This difference in perception is due to the presence of interaural time and intensity differences, that occur when speech and noise sources are spatially separated.7,29

Importantly, the results of the compound noise (CN) are generated automatically by the HINTPro software, according to the formula RC=(2×NF+NR+NL)/422. Although age per se is not an essential factor in speech perception, there are reports in the literature of performance deterioration in speech perception with increasing age – in association with the natural aging process.1 Considering that the present study group was composed of young adults (between 19 and 44 years, mean 24 years), future studies with older groups of adults or even with elderly patients would be welcome.

Subjects of both genders and with ages between 18 and 50 years took part in the standardization of HINT Brazil18 using headsets. In tests performed in-noise, the authors consider questionable the need for adjustments with the introduction of visual cues for the older population, even if only for previous training to the test application, since, when the stimulus is given only by ear, older adults demonstrate less poor performances in speech recognition, when compared to young adults or even to normal hearing individuals.30,31 The same occurs when only visual cues are used.30,32,33 Overall, regardless of age, it is common that in difficult listening situations the listener performs lip reading to facilitate his/her understanding.30,34

Values were developed for the standardization of HINT in the free field during different listening situations in a noisy environment. The results of speech perception in adults with normal hearing showed worse performance in the situation of NF, i.e., 0–0°.

These results could be used as a reference in the construction of protocols for use of speech perception tests in noise, and in monitoring individuals with hearing loss.

This study was supported by PIBIC (Programa Institucional de Bolsas de Iniciação Científica)/ RUSP (Reitoria da Universidade de São Paulo).

The authors declare no conflicts of interest.