Re: low-latency audio I/O for Windows (Kelly Fitz )


Subject: Re: low-latency audio I/O for Windows
From:    Kelly Fitz  <kelly_fitz@xxxxxxxx>
Date:    Wed, 3 Oct 2007 15:55:44 -0700
List-Archive:<http://lists.mcgill.ca/scripts/wa.exe?LIST=AUDITORY>

I believe that RME still makes the lowest latency sound cards, but that they do not really approach the latency you require. The Hammerfall claims 1.5ms, but that is not throughput latency, it is the hardware latency. For simulating hearing aid algorithms, you need to pay that twice, once going in, and once going out. Moreover, most audio APIs that I know of use some kind of buffering, and you will therefore have those latencies on top of the sound card hardware latency. Other software systems like Max or Pd may impose additional buffering latencies. I am told that at the last IHCON, Giso Grimm (University of Oldenburg) demonstrated a system that achieved latency as low as 7ms, using a linux PC with ASIO drivers and the Hammerfall card. I think the system itself is described in Giso Grimm, Tobias Herzke, Daniel Berg, and Volker Hohmann. The Master Hearing Aid – a pc-based platform for algorithm development and evaluation. Accepted 2006 Acustica united with acta acustica, April 2006. I have never heard of a low-latency USB or 1394 interface, and I don't know enough about those protocols to know whether it is even achievable, but it is a reasonable question to ask. The low-latency RME cards are PCI or cardbus. If anyone can come up with a PC-and-soundcard-based system that has reliable and demonstrable <4 ms throughput latency, I'd love to know about it. HTH, Kelly Freed, Dan wrote: > Dear Auditory List Members: > > I'm looking for a low-latency audio interface device for use with > Windows computers. The goal is to perform real-time signal processing > on the PC for the purpose of developing and testing hearing aid DSP > algorithms. Consequently I need to keep the hardware-related signal > delay to a minimum. > > Unfortunately, manufacturers generally don't include latency in their > specifications, and tech support and sales people don't seem to have > latency information either. So I'm hoping that people on this list can > point me at some acceptable devices. > > My target is a total input-to-PC-to-output delay <= 3 ms at 48 kHz. I'd > prefer an external device (USB or FireWire). Other desirable features > include a mic input, a headphone output, analog I/O level controls, and > support for a wide range of sampling rates (16, 24, and 32 kHz would be > nice). But the critical feature is the low latency; I can work around > other limitations. > > Any suggestions will be greatly appreciated. Thank you! > > Dan Freed > Senior Engineer > Dept. of Human Communication Sciences & Devices > House Ear Institute > 2100 W. Third St. > Los Angeles, CA 90057 USA > Phone: +1-213-353-7084 > Fax: +1-213-413-0950 > Email: dfreed@xxxxxxxx > -- Kelly Fitz, DSP Research Engineer Starkey Hearing Research Center


This message came from the mail archive
http://www.auditory.org/postings/2007/
maintained by:
DAn Ellis <dpwe@ee.columbia.edu>
Electrical Engineering Dept., Columbia University