In virtual audio synthesis we use different excitation signals for listening tests. These tests are executed using headphone playback and real-time HRTF synthesis. Besides noise signals, speech is often used for tests especially for real life applications, such as mobile phones, various voice transmission lines, computer environments or accessories for the visually impaired, where speech intelligibility is important. This paper presents a summary of results of a listening test using different signals including speech in a virtual audio environment aimed at blind persons. Results show how speech contributes to the accessibility to computers and performs in a comparative test for virtual audio simulation. Furthermore, additional speech test signals are overviewed and introduced, such as speech-chorus signals, segmented spondees and the newly developed spearcons.