Subject: Re: Data Acquistion Standards From: David Isherwood <David.Isherwood(at)NOKIA.COM> Date: Fri, 15 Jun 2001 16:45:06 +0300Hi all, On 14 Jun 01, at 15:43, Bradley Wood Libbey wrote: <snip> > However, I am concerned about noise. I suspect that a majority of the > noise actually exists as acoustic energy, but I don't want my > measurement > system to be adding to this. The first possible source of > noise is from > aliased high frequency signals. What is the standard in > auditory research > for antialiasing filters? If they are necessary, do most > researchers use > stand alone filters or do they simply acquire digital data on > acquisition > systems with built in filters? Firstly, previous discusions have talked of the use of Sound Blaster cards for sample acquistion. Using such a device as a microphone amplifier and A/D means that any artifacts associated with the anti-aliasing filter will be very small when taking a holistic view of the noise and non-liniarities inherent in the system as a whole. The amount of artifacts created by the anti-aliasing filter will depend on the sampling frequency used. For 44.1/48kHz the usual negative effects of any low-pass filter (e.g. ripple, phase distortions, ringing) are mainly present in frequencies above 18kHz and are thus of little effect given the small amount of energy present in speech at these frequencies. I am not fully familiar with your experimental needs and computing resources, but on modern PCs the falling costs of audio cards that sample at higher sampling rates (upto 96kHz) mean that if this is an issue, then sampling at higher rates and then software resamling to lower sampling frequencies will give better results as the effect of the anti-aliasing filter will be confined to the frequencies you are dumping. > Secondly is the analog circuitry or the analog to digital converter > noisier on an inexpensive card than on a "professional" acquisition > system? The short answer is yes, but the longer one will depend on the source signal. If speech is being recorded under anechoic conditions then 'noise' created by the acquisition system will become an issue. One of the main problems for digital audio acquisition boards on computers is the amount of RF energy produced by the other components in a computer inducing noise in the input amplifier and A/D stage. This can be kept to a minimum by choosing a card which is well shielded or by placing the card away from 'noisy' sources (e.g. hard drives). Better still would be to use an external A/D and D/A connected to a digital input on the computer(even cheap Sound Blasters have these now). There are other fundamental issues that can effect speech acquisition but I can't make recomendations without knowing more about your needs. Cheers Dave David Isherwood Research Engineer Speech and Audio Systems Laboratory Nokia Research Center Visiokatu 1 P.O. Box 100 FIN-33721 Tampere Tel: +358(3)2725760 Mob: +358(40)7499293 Fax: +358(3)2725899 Email: david.isherwood(at)nokia.com