Re: Uli Behringer of The Music Group Q&A
Interesting,
Andrew, I think you're saying that you don't use any FIR filters in the processing [or are you saying that you don't use FIR filters for the "final voicing"?] for the sake of keeping the overall latency down. I'm curious about how much time delay you figure you have been able to save versus using FIR?
Secondly, since I think you were answering a question about the new Turbo branded speaker line vs the discontinued Behringer branded line, are you saying that the Behringer branded line used higher latency FIR processing?
Thanks!
Hi Drew,
Thank you for your question.
First of all I apologise if my answers are considered “marketing”. I am trying to answer people’s questions in a technical way and in particular in a thread that is associated with our company. If my answers are considered “marketing” I would appreciate guidance from the admin.
Back to your question:
IIR, FIR and the different types thereof all have their benefits and trade offs of course. I believe that in some applications FIR: especially Linear Phase FIR filters are used because it's an easy shortcut to a 3/4 way decent system voicing. I've seen systems that have latency of 10ms or more due to FIR and Linear Phase being used down to relatively low frequencies. Correct implementation of IIR can in many cases give the correct phase components on and off axis without the need of FIR thus allowing low latency. Also there are some psychoacoustic benefits to IIR such as no pre-ringing of the impulse response and no theoretical end to the impulse response, these are both factors present in 'real' sound that may come from acoustic instruments, voices etc. Although, for example, Linear Phase filters may more accurately reproduce signals on an oscilloscope or when transfer functions are viewed in the frequency domain, in some quite intensive listening tests we have performed, a well designed IIR solution sounded better and more accurate to the human ear - which after all is what it's all about. I other cases, especially with more complex systems, the FIR solution may give a better overall sound.
Another goal with TLX is to be able to provide a set of equalisation points that can be used in a controller with a relatively low DSP horsepower so that the systems would be available to a wider audience and the user could utilise already purchased equipment. Simplicity done well usually just sounds better.
We have always gone to great lengths to create the best possible electro-acoustic systems with the minimum need for equalisation. Better sound quality is always achieved by making the components and acoustic systems work correctly in the first place rather than having to 'fix' them with DSP afterwards.
In answer to your question: No FIRs are used in the system. Causal IIR methods + short delays are used in TLX to match and better an equivalent FIR method, without latency. ELX had a much more standard IIR method.
~
