By your own logic, one may ask why not 100 times better, just for the sake of the challenge.
What you came up as a reference has nothing to do with the group delay audibility at 20Hz. Just desperate Google searches to justify the nonsense of your requirement. Anyway, good luck learning to design digital FIR filters to reach those numbers.
I don't do "critical listening test" by myself, I consider them useless by any metric, and I don't have the means to organize serious sensory testing.
The obvious answer, easily done these days, is get one's whole collection onto hard disk; work out precisely what needs to be done to get perfect frequency, phase response, and anything else, from your current system, and DSP the lot offline to form the normal play collection. Of course, this takes all the fun out of wrestling with getting speakers just right,
- so will never be done, .
RNM,
Here is a list of some relevant articles from the old AES Audio Anthology, Vol. 1- Vol.25 1953 to 1977. I imagine you may have a copy, mine is very old but these articles may still have some information you may find helpful. I also need to look for a white paper I have on Time Alignment in my files that defines some of the real problems with time alignment.
Loudspeaker Phase Characteristics and Time Delay Distortion. Part 1 & 2
By: Richard Heyser page 120- 139
Active Crossover Networks for Noncoincident Drivers
By Siegried H. Linkwitz page 367-373
Here is a list of some relevant articles from the old AES Audio Anthology, Vol. 1- Vol.25 1953 to 1977. I imagine you may have a copy, mine is very old but these articles may still have some information you may find helpful. I also need to look for a white paper I have on Time Alignment in my files that defines some of the real problems with time alignment.
Loudspeaker Phase Characteristics and Time Delay Distortion. Part 1 & 2
By: Richard Heyser page 120- 139
Active Crossover Networks for Noncoincident Drivers
By Siegried H. Linkwitz page 367-373
Hi,
Such a contrast in helpfulness between this an Waly's noise! I do have the Anthology issues. The phono HP filter GD issue isn't for me but for TT/LP users to implement. I dont have such a system to use it on.
The best and most current numbers I can find on just noticeable detection of GD is 3-5ms at 20Hz. So, that or better is what I myself would choose. But, Waly seems to think other-wise. So be it.
I have a book coming that goes into great detail on the subject and gives numbers also (they are low). If it has useful info for the HP filter, I'll pass it along.
The more popular use will be for other needs.... active speaker amp/cross-over and the like using DSP to mimimise Group Delay. I just put up one research paper that is current (2014) along these lines. So there is hope that in the not too distant future we will be listening to better systems than ever before at a low cost.
It just keeps getting better.
http://www-inst.eecs.berkeley.edu/~n247/sp08/lectures/L05.pdf
THx-RNMarsh
Last edited:
I was taught about group delay in the Army at Ft Monmouth, N.J. The data that was encrypted had to be group delay EQ'ed or you would get garbage data. A company called Singer Instrumentation is what I used to flatten the group-delay over the analog communication lines being used for data. This was practiced for my assignment as shift chief of Europes largest comm site (USA) at that time in Frankfurt around 1970-72.
So I am sure such gear exists today somewhere and uses DSP. Though it wouldnt be optimized for audio needs most likely.
But ever since then, I wanted to try it on audio. Maybe the time is finally approaching?
https://ccrma.stanford.edu/~jos/filters/
THx-RNMarsh
So I am sure such gear exists today somewhere and uses DSP. Though it wouldnt be optimized for audio needs most likely.
But ever since then, I wanted to try it on audio. Maybe the time is finally approaching?
https://ccrma.stanford.edu/~jos/filters/
THx-RNMarsh
Last edited:
Which must include you since frequency dependant delay has nothing to do with dispersion.
Vaccy,
If you look in the old classic acoustic texts you will find dispersion does not refer to coverage angle! It is used as a label for the behavior of wavelength dependent changes in propagation velocity! Really.
Now if you worry about the microsecond differences in home scale systems, well that is a different issue.
The other issue is the common error of confusing the center of radiation with the voice coil location. The velocity of propagation in the cone is greater than in the air. So the center is actually forward of the voice coil. Generally the place where the area of the cone forward is equal to the area behind.
The demonstration is to use a dual driver box and look at where the coverage angle narrows for a given frequency.
The demonstration is to use a dual driver box and look at where the coverage angle narrows for a given frequency.
According to the word definition, instant propagation in the cone, if it is working in piston, not ?
Anyway, I am not able to predict the "source" point of a speaker (at a given frequency). Cones are, indeed, a (more or less) little in front of their moving coils. Horns drivers too due to the higher velocity in compressed air.
So, my way is very pragmatic. First, minimize the numbers of the ways, reason why i use horns for medium trebles. Find the two transducers and horns for having one octave of margin (my actual problem) between horn cutoff and the frequency where boomer will fractionate (first visible little accident in the impedance curve that coincide often with a peak in the response curve.
Set my filters 24dB/oct acoustic, integrating the natural slope of the transducers.).
Then tune the vertical alignment for having the best square waves possible.
(something like this ?
Last edited:
)
Moonaudio.fr was interesting to read, merci pour me donner l'opportunité d'améliorer mon Français.
Blind painting expert ?
http://www.moonaudio.fr/introduction.html
This square wave came from their site.
They use active filters with time alignment. Could do better out of axis if they used physical alignment.
Last edited:
Nope, dispersion is the proper term. In wave and vibration physics it describes the case where phase velocity depends on frequency - that is when propagation velocity (and therefore delay over a set distance) is frequency dependant.
Delay, group delay, and dispersion are separate concepts. And, as I say, are very muddled up on this thread, or apparently not even understood in some cases.
Dispersion in loudspeaker lingo deals with off axis behaviour of drivers or horns and is determined by geometry.
You are right in the sense that in the xover region frequency dependant delay may play a role, although in audio you would still not use the word 'dispersion' but rather 'lobing' to describe the effect this may have.
In cables it is used in the longitudinal sense, is different.
You are right in the sense that in the xover region frequency dependant delay may play a role, although in audio you would still not use the word 'dispersion' but rather 'lobing' to describe the effect this may have.
In cables it is used in the longitudinal sense, is different.
- Status
- Not open for further replies.