Re: Reference for typical SNRs is public spaces (John Culling )

Subject: Re: Reference for typical SNRs is public spaces From: John Culling <CullingJ@xxxxxxxx> Date: Thu, 23 Jan 2014 10:44:14 +0000 List-Archive:<http://lists.mcgill.ca/scripts/wa.exe?LIST=AUDITORY> --_000_48f0d89a520148ff9deb3a6744b3a723DB3PR02MB091eurprd02pro_ Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable Hi Andy. This is a very tricky question. I am not aware of any definitive data that = really addresses the issue adequately. Laboratory studies tend to use single sound sources in anechoic conditions. The auditory system copes very well in these conditions. The results indica= te that listeners can cope with very low SNRs (e.g. -10 dB for spatialised spe= ech- shaped noise interference, and lower for speech interferers). Moreover, som= e studies have used several interferers (e.g. Peissig and Kollmeier '97, Hawl= ey et al. '04), and shown a gradual elevation in SRT with increasing numbers of inter= ferers. Simulating a more complex scene, like a restaurant with multiple interferer= s and reverberation produces progressive degradation, though. We have been simulating up to eight interfering voices from a variety of speakers with r= everb based on real-room binaural room impulse responses. SRTs are around -2 to -= 3 dB with eigth interfering voices. I haven't begun to write this work up yet= , but the results are not disimilar to those from a cruder preliminary study publ= ished here. Culling, J. F. (2013). "Energetic and informational masking in a simulated = restaurant environment" in Moore, B. C. J., Carlyon, R. P., and Gockel, H.,= Patterson, R. D. and Winter, I. M.. (eds) Basic Aspects of Hearing: Physio= logy and Perception (Springer, New York) There remain limitations to this approach, of course. The technique remains dependent on standard target speech materials (IEEE/Harvard sentences) that are not very typical of normal conversation - particularly lacking a conversational context. It is also unclear whether 50% keyword intelligibil= ity is a tolerable level of comprehension for conversation. Karolina's study has other limitations. If I remember correctly, the materi= al was recorded from hearing impaired individuals, who may avoid the more severe listening conditions into which normally hearing people thrust themselves. Also, the method of establishing the SNR from the recordings would probably become impossible below a certain SNR, as it relies on a researcher judging from the recordings alone whether or not target speech is present. Noise level is collected from epochs without target speech, and speech level is derived by subtraction. Nonetheless, both approaches indicate that real-world SNRs are unlikely to be very near -10 dB, but be somewhere around 0 dB. Karolina's work suggests a bit above, mine a bit below. I guess what is really needed is for pairs of interlocutors to be wired up = with close microphones at the mouth (to establish reliably who is talking when) and at their ears, and then to go out for the night and try to produce norm= al speaking and listening behaviour. Perhaps after a few nights of this they would habituate to all the kit, and produce data that will get us closer to= a true answer. John From: AUDITORY - Research in Auditory Perception [mailto:AUDITORY@xxxxxxxx= ILL.CA] On Behalf Of Andy Sabin Sent: 22 January 2014 17:53 To: AUDITORY@xxxxxxxx Subject: Reference for typical SNRs is public spaces Hi List, Can anyone point me to a reference showing SNRs that are typically observed= in public spaces (e.g., restaurants, bars ...etc)? I can find this info fo= r overall SPL, but am having a hard time finding it for SNR. Thanks Andy Sabin --_000_48f0d89a520148ff9deb3a6744b3a723DB3PR02MB091eurprd02pro_ Content-Type: text/html; charset="us-ascii" Content-Transfer-Encoding: quoted-printable <html xmlns:v=3D"urn:schemas-microsoft-com:vml" xmlns:o=3D"urn:schemas-micr= osoft-com:office:office" xmlns:w=3D"urn:schemas-microsoft-com:office:word" = xmlns:m=3D"http://schemas.microsoft.com/office/2004/12/omml" xmlns=3D"http:= //www.w3.org/TR/REC-html40"> <head> <meta http-equiv=3D"Content-Type" content=3D"text/html; charset=3Dus-ascii"= > <meta name=3D"Generator" content=3D"Microsoft Word 15 (filtered medium)"> <style></style> </head> <body lang=3D"EN-GB" link=3D"#0563C1" vlink=3D"#954F72"> <div class=3D"WordSection1"> Hi Andy.<o:p= ></o:p> <o:p> <= /o:p> This is a ve= ry tricky question. I am not aware of any definitive data that really<o:p><= /o:p> addresses th= e issue adequately. <o:p></o:p> <o:p> <= /o:p> Laboratory s= tudies tend to use single sound sources in anechoic conditions.<o:p></o:p><= /span> The auditory= system copes very well in these conditions. The results indicate <o:p></o:p> that listene= rs can cope with very low SNRs (e.g. -10 dB for spatialised speech-<o:p></o= :p> shaped noise= interference, and lower for speech interferers). Moreover, some <o:p></o:p> studies have= used several interferers (e.g. Peissig and Kollmeier '97, Hawley et al. <o:p></o:p> '04), and sh= own a gradual elevation in SRT with increasing numbers of interferers.<o:p>= </o:p> Simulating a= more complex scene, like a restaurant with multiple interferers<o:p></o:p>= and reverber= ation produces progressive degradation, though. We have been<o:p></o:p></sp= an> simulating u= p to eight interfering voices from a variety of speakers with reverb<o:p></= o:p> based on rea= l-room binaural room impulse responses. SRTs are around -2 to -3<o:p></o:p>= dB with eigt= h interfering voices. I haven't begun to write this work up yet, but <o:p></o:p> the results = are not disimilar to those from a cruder preliminary study published<o:p></= o:p> here.<o:p></= o:p> <o:p> <= /o:p> Culling, J. F. (2013). “Energetic and informational masking in a simulated resta= urant environment” in Moore, B. C. J., Carlyon, R. P., and Gockel, H.= , Patterson, R. D. and Winter, I. M.. (eds) Basic Aspects of Hearing: Physiology and Perception (Springer, New Y= ork)<o:p></o:p> <o:p> </o:p> There remain limitations to this approach, of course. Th= e technique remains<o:p></o:p> dependent on standard target speech materials (IEEE/Harv= ard sentences)<o:p></o:p> that are not very typical of normal conversation - parti= cularly lacking a <o:p></o:p> conversational context. It is also unclear whether 50% k= eyword intelligibility<o:p></o:p> is a tolerable level of comprehension for conversation.<= o:p></o:p> <o:p> </o:p> Karolina's study has other limitations. If I remember co= rrectly, the material<o:p></o:p> was recorded from hearing impaired individuals, who may = avoid the more<o:p></o:p> severe listening conditions into which normally hearing = people thrust <o:p></o:p> themselves. Also, the method of establishing the SNR fro= m the recordings<o:p></o:p> would probably become impossible below a certain SNR, as= it relies on<o:p></o:p> a researcher judging from the recordings alone whether o= r not target <o:p></o:p> speech is present. Noise level is collected from epochs = without target<o:p></o:p> speech, and speech level is derived by subtraction.<o:p>= </o:p> <o:p> </o:p> Nonetheless, both approaches indicate that real-world SN= Rs are unlikely<o:p></o:p> to be very near -10 dB, but be somewhere around 0 dB. Ka= rolina's work<o:p></o:p> suggests a bit above, mine a bit below.<o:p></o:p></span= > <o:p> </o:p> I guess what is really needed is for pairs of interlocut= ors to be wired up with <o:p></o:p> close microphones at the mouth (to establish reliably wh= o is talking when) <o:p></o:p> and at their ears, and then to go out for the night and = try to produce normal<o:p></o:p> speaking and listening behaviour. Perhaps after a few ni= ghts of this they <o:p></o:p> would habituate to all the kit, and produce data that wi= ll get us closer to a <o:p></o:p> true answer. <o:p></o:p> <o:p> </o:p> John<o:p></o:p> <o:p> <= /o:p> From:<span= lang=3D"EN-US" style=3D"font-size:11.0pt;font-family:"Calibri",&= quot;sans-serif""> AUDITORY - Research in Auditory Perception [mailto:= AUDITORY@xxxxxxxx On Behalf Of Andy Sabin Sent: 22 January 2014 17:53 To: AUDITORY@xxxxxxxx Subject: Reference for typical SNRs is public spaces<o:p></o:p></spa= n> <o:p> </o:p> <div> Hi List, <o:p></o:p> <div> <o:p> </o:p> </div> <div> Can anyone point me to a reference showing SNRs that= are typically observed in public spaces (e.g., restaurants, bars ...etc)? = I can find this info for overall SPL, but am having a hard time finding it = for SNR. <o:p></o:p> </div> <div> <o:p> </o:p> </div> <div> Thanks<o:p></o:p> </div> <div> Andy Sabin<o:p></o:p> </div> </div> </div> </body> </html> --_000_48f0d89a520148ff9deb3a6744b3a723DB3PR02MB091eurprd02pro_--

This message came from the mail archive
/var/www/postings/2014/
maintained by:

DAn Ellis <dpwe@ee.columbia.edu>
Electrical Engineering Dept., Columbia University