Member
Joined 2003
The real problem is tweeter-midrange integration, not the order of the slopes per se. Wiggles at midrange step response at about 1.1ms and at 1.25ms.
In recent ASR loudspeakers reviews there are impulse and step response measurements.
Sometimes I wonder how people can see number of ways, polarities and acoustical integration by looking step response. For example this. What can you see?
Let's try. Three way, T reversed polarity, M+W normal polarity ?? Time coherent, smooth transitions.
Like solving Wordle 🤔🙂
Like solving Wordle 🤔🙂
I have seen few impulse responses and probably one step measured by Erin, but that's not adequate. Group delays are needed due to good readability for a human being, and ETC would be very nice also for acoustical timing issues such as diffraction. When I asked GD plot first time on ASR, some member - probably a donor replied that amirm does not care about timing and probably won't show it. He finally did, but maybe that was the last GD visible on the forum.
Anyway, this is not the right place to make suggestions or criticize some other forums. My ban continues, but I don't mind.
Quite good, but no. This was just a trick 🙂 It's ideal single way with one 4th order high-pass at LF and three 2nd order all-pass filters. Phase response looks vented 4-way with 4th order slopes.
I bet @DcibeL will confirm my guess that his speaker is a 3-way, with positive polarities on all drivers. Wiggles at 1.1ms and at 1.25ms do suggests bad tweeter-midrange integration.
In your example very evident were suspiciously clean impulse response and ETC.
Three all-pass filters are very unusual (improbable) for a loudspeaker. No wonder many will fail on this trick.
Member
Joined 2003
Well, you are correct on the first part, it is a 3-way speaker, and all drivers are in positive polarity. However, I don't know what is "bad" about the midrange to tweeter integration. In-room and power response for this speaker is very good, and phase between drivers is nearly spot on through the crossover region.
Here is phase response for midrange and tweeter, crossover is at 2700Hz.
Definitely more than one but definitely not all of them.
I’d say maybe 1/4. I have a script to help me process the information a bit more quickly and sometimes I run the “old” version which doesn’t include the GD and step and I forget to add them manually.
However, I took the time tonight to fix that. So the remainder should have both from here on.
Of course, server space being costly over time, I will see if this is useful enough to the community to continue after some time.
- Erin
Excellect. Here is reference information for required graphs, traces and ms scale. It's actual speaker which can run as a rabbit for other commercial products. Nothing really special - just decent design with standard components.
Remove some duplicate normalized or single frequency directivity charts if server space is limited.
Thanks for confirming! I apologize, "bad (integration)" is too strong word, better description is "not so optimal". Yes, phase response is very good, but I suspect there is something not so optimal with tweeter-midrange time difference in acoustical paths, from their acoustical centers (origins) to the microphone. If you are using VituixCad, try with different position "z" of the tweeter. Pay attention to "vertical directivity line chart".
Also the latest Genelec Ones are improved - I guess as a result of their latest studies about sensibility, audibility and significance of timing.
Very nice that manufacturer offers "different compromise" at least as an option. This was very different a decade ago. Proper studies seem to have more weight than quickly referred uncertainties e.g. in Toole's documentation.
Very nice that manufacturer offers "different compromise" at least as an option. This was very different a decade ago. Proper studies seem to have more weight than quickly referred uncertainties e.g. in Toole's documentation.
Forgive my naivity, but is it not still a large amount of group delay? 3ms is ~ 1m, and in low frequenties the wavefront is important for our hearing system to resolve and discriminate sounds.
How was this measured, considering the high resolution down to 20. In anechoic chamber, or with near+far blending? Klippel maybe?
Also, maybe some basic info about the system: closed/BR etc; LF extension;2/3/x ways; all FIR; hybrid passive/FIR EQ etc... Just to have some references.
Last edited:
Could it be the Taipuu speakers Kero active shown here:
http://www.htguide.com/forum/showthread.php?44128-VituixCAD-v2/page11
http://www.htguide.com/forum/showthread.php?44128-VituixCAD-v2/page11
That's XO simulation of Kero Active prototype with FIR preset. LF is near field + simulated baffle loss. Measurements are not final.
IR+SR of IIR preset looks different though timing of 2-way is almost never bad:
Minimum phase calculation is not reliable at very low frequencies due to noise and limited measurement range so group delay graph could/should start at 50 Hz (just like in Genelec's graph).
Attachments
I'd like to think positively and compare to first two Ones; 8260a and 8351a. Progress is no less than huge.