Wesayso - I think you may be right! Could also be why you can hear improvements, whilst others with more traditional multiway set-ups may not, or even think it sounds worse with phase manipulation.
Well I'd definitely would advise them not to move up or down too much 😀.
It is also the reason I think a Synergy horn could do better here (than conventional multi ways) as it generally acts like a coaxial driver. It would not have the driver distances change as much compared to a more traditional setup. It behaves much like a single full range.
It is also the reason I think a Synergy horn could do better here (than conventional multi ways) as it generally acts like a coaxial driver. It would not have the driver distances change as much compared to a more traditional setup. It behaves much like a single full range.
Last edited:
I don't have any room measurements only pseudo-anechoic. I am not really in the market to changing my entire system just to play with some new software that I don't have and have never used. Its just not what I do. My main system is for listening and I don't play around with it. As I said I do all my psycho-acoustical work in a lab.
Yes there is group delay from the LP filters, but then the drivers are not in the same plane so this gets corrected. In the mid-band, i.e. through the crossover, there is no significant group delay, basically none (I could show you that.) As I said I have used a two channel DSP to bi-amp the system and the difference is not great, about the 5% that we have been talking about.
I think that it could be shown that if you have a flat DI and smooth frequency response then you have an optimized phase/group delay situation. This is clearly what Toole is saying and I can see now that it has to be true.
Last edited:
If not, then please stop haranguing us. I've offered to help, it's not hard to do the tests, the software is free and easy to use. How can we take you seriously when you refuse even a simple test?
Geddes - I still don't buy the story here. If there is a time change in the signal dependant on frequency content, how can it sound like the original recording of an instrument that covers the frequencies where the change happens?
In order to keep this thread a bit on the subject of the synergy it would still be interesting to see the semi-anechoic impulse, STEP and group delay to at least be able to compare it to what Nate is getting with the mini Synergy.
I'm just glad microphones are not multi-way. To bad most mixing/mastering probably involves speakers which have some form of phase rotation. And no telling where the crossovers were during the mixing process or what speakers were used.
For an example of crossovers in the vocal range Troels Gravesen had an interesting view on this:
Siri's Killer Note
Although this was a test of first order slopes compared to higher orders, something like this would also be very interesting to compare/test on time corrected speakers vs non time corrected speakers.
In order to keep this thread a bit on the subject of the synergy it would still be interesting to see the semi-anechoic impulse, STEP and group delay to at least be able to compare it to what Nate is getting with the mini Synergy.
I'm just glad microphones are not multi-way. To bad most mixing/mastering probably involves speakers which have some form of phase rotation. And no telling where the crossovers were during the mixing process or what speakers were used.
For an example of crossovers in the vocal range Troels Gravesen had an interesting view on this:
Siri's Killer Note
Although this was a test of first order slopes compared to higher orders, something like this would also be very interesting to compare/test on time corrected speakers vs non time corrected speakers.
Last edited:
Hi All,
I will prepare a test for you showing different phase influences. At the moment I am really busy (end of year stuff at the company) but I promise it will be here in the next couple of weeks.
I will open a new thread for this, since I think I am overblurring Nate's excellent thread here.
Two questions;
-Does anyone know where/how I can host WAV-files such that everyone can access them? the files will be too big for the forum.
-Earl, I've been checking the Linda Ronstadt track, nice track, but I can't find it in downloadable high quality anywhere. (best thing I have is spotify 320ogg which is good but not perfect). Would you accept this (I will also provide the "dry" example), or will you send me the track?
Kees
I will prepare a test for you showing different phase influences. At the moment I am really busy (end of year stuff at the company) but I promise it will be here in the next couple of weeks.
I will open a new thread for this, since I think I am overblurring Nate's excellent thread here.
Two questions;
-Does anyone know where/how I can host WAV-files such that everyone can access them? the files will be too big for the forum.
-Earl, I've been checking the Linda Ronstadt track, nice track, but I can't find it in downloadable high quality anywhere. (best thing I have is spotify 320ogg which is good but not perfect). Would you accept this (I will also provide the "dry" example), or will you send me the track?
Kees
Here is my little full range synergy's phase, impulse, and step. I swear it is easy to take and show with a few clicks of the mouse.
Here are the xo slopes at 500Hz (light blue is xo simulation from PCD):
Here are the xo slopes at 500Hz (light blue is xo simulation from PCD):
Earl, you could have those and still have an infinite number of different phase responses, are they all equally optimum? An FIR filter could be applied before the amp and make the phase, by itself, be wildly different. For example, random phase would make the IR very spread out. But DI would still be flat as would response magnitude.
If restricted to only using minimum phase filters ahead of it all, that might be close (though phase response can still vary since response magnitude outside the audio band will still affect minimum phase values in band some).
Look waveforms into pictures at post 320 (http://www.diyaudio.com/forums/multi-way/283068-mini-synergy-horn-experiment-32.html#post4556377) even picture 1 can be corrected on axis with FIR filters to look as picture 2 or 3.
Maybe not so easy for me. How do you effect the power response of a single section (let's say a 15" midwoofer like in the NS-15)? Obviously any eq done is global in terms of the power response. Are you just talking about basic eq of peaks/dips? A passive xo can't really fix a dip in power without pulling the whole response down, so I guess that's one thing that can be done with active. The only other thing I can think of is delay affecting the power response though the xo of the combined system.
Maybe I need to start drinking again, and I'll more clearly see your epiphany 😉
I guess my description gave it away ;-) But not quite true, hence the reason why it is interesting to experiment. I have posted elsewhere on diyAudio multiple measurements of my 3-way on and off axis around a 6' x 2' grid showing that the fr and step response stays virtually the same over a large sweet spot at the listening position. Aside from the time alignment, this is in part due to using frequency dependent windowing (FDW) for both amplitude and excess phase correction. Also in part to using constant directivity waveguides from 500 Hz on up.
All else being equal, it is interesting to hear what adding delay to an XO leg response does by AB'ing the 2 filters.
wesayso, having spent 10 years recording/mixing in a number of pro studios, every one I was in had a pair of these: 1977 UREI 813 Studio Monitors | Mixonline While they certainly had diffraction and directivity issues, amongst other issues, the one thing they did well was time alignment. See Ed Long's work on that. As noted in the article, there are more of these monitors in studios than any other, certainly during that time frame.
One other point is that most of these pro studios had much superior acoustical design than any of our living rooms in terms of room ratios, NC ratings and low frequency diffusion. Every little bit helps 🙂
If I understand the doctor correctly, it's this:
1) When you apply EQ to a loudspeaker system, the EQ will change the power response and the frequency response at the same time. For instance, if I apply a boost of 3dB at 1500hz, the frequency response and the power response will both change.
2) If you use dedicated amplifier channels and dedicated EQ for every driver, you can "decouple" the two.
For instance, let's say you have a two-way like the Gedlee NS15. If it's using a single channel of amplification and a single channel of EQ, then the power response and the frequency response are tied together. You can't change one without changing the other. But if you use a dedicated channel for each, you can tweak the EQ and change the frequency response and the power response independent of each other.
For instance, you might apply 3dB of boost to the tweeter alone, or 3dB of boost to the woofer, or a fraction to both. You would monitor the power response to determine which combination is the most effective.
This would probably be especially powerful if you had an array of microphones, and some software that can try a series of combinations until you came up with the optimum power response.
Once you had delay into the mix, you can see that this gets REALLY powerful, because you have a LOT of variables to tweak.
For instance, even with three "tweaks", you're talking about hundreds of measurements. Again, using a hypothetical boost of 3dB at 1500hz, you have these measurements:
1) 3dB boost on woofer and tweeter, no delay
2) 3dB boost on tweeter alone, no delay
3) 3dB boost on woofer alone, no delay
4) 3dB boost on woofer and tweeter, .5ms delay on tweeter
5) ...
Once you factor in the requirement for doing these modifications and measurements from a number of angles, you start to see that's it's a fairly insane number of measurements.
But, again, do this via a computer and some software, and it's powerful. It could be the difference between having good looking polars and GREAT looking polars.
Or I may have misunderstood Geddes entirely 🙂
I think this goes back to the point about system design as a whole (and the benefit from such correction reducing as the "quality" of the system goes up).
i.e. you have a, iirc, a pretty well controlled room and have constant directivity speakers. You then apply an FDW which means that the vast majority of what you're seeing in the measurement is the speaker itself. As a result your listening window measurements should be pretty consistent and your subjective experience could also be expected to be similarly consistent.
It seems another q again as to whether this means the (excess) phase correction is worthwhile in your case.
I measured the Danley SH50 and my Gedlee Summas about a year ago.
Data starts here : http://www.diyaudio.com/forums/multi-way/244508-monster-massive-20.html#post4157679
I'd take the data with a grain of salt. I'd rented them for a couple days and I didn't have much time to do the measurements properly.
The thing that was REALLY vexxing with the Summa was the center-to-center spacing. I didn't realize it when I bought them, but due to the nearly two foot gap between woofer and tweeter, they're REALLY sensitive to vertical height. IE, if the pathlength between the woofer and tweeter reaches seven inches, that will create a suckout*. And due to the large center-to-center spacing, it's pretty easy to have that happen if you're not careful with the vertical tilt or your seating height.
This isn't a condemnation of the Summas - they're really great. Just something to keep in mind when you're listening to or measuring a speaker like it. You have to watch that vertical height carefully.
* If anyone's curious why the suckout occurs, it because that pathlength difference creates 180 degrees of phase shift at the xover frequency.
Data starts here : http://www.diyaudio.com/forums/multi-way/244508-monster-massive-20.html#post4157679
I'd take the data with a grain of salt. I'd rented them for a couple days and I didn't have much time to do the measurements properly.
The thing that was REALLY vexxing with the Summa was the center-to-center spacing. I didn't realize it when I bought them, but due to the nearly two foot gap between woofer and tweeter, they're REALLY sensitive to vertical height. IE, if the pathlength between the woofer and tweeter reaches seven inches, that will create a suckout*. And due to the large center-to-center spacing, it's pretty easy to have that happen if you're not careful with the vertical tilt or your seating height.
This isn't a condemnation of the Summas - they're really great. Just something to keep in mind when you're listening to or measuring a speaker like it. You have to watch that vertical height carefully.
* If anyone's curious why the suckout occurs, it because that pathlength difference creates 180 degrees of phase shift at the xover frequency.
Last edited:
Funny. I see so much discussion about phase, which in the end is going to come down to the age old "square wave in ,square wave out" A speaker too should be a straight wire with gain. Simplistic yes, but Ithink we all know in the end that the speaker that can do a square wave is pretty likely to sound great. (-:
There is only one reason to use a passive crossover anymore and that is financial, a reason which is rapidly becoming a non reason. 10 minutes with JRiver and this became clear as invisible glass. (-:
The real upshot of all of this is there are a few patents that have a stranglehold on the industry. It's a matter of time till the dam breaks. This is looking like the airplane over the dirigible. Somebody has to say this. My hat has long been off to Tom Danley, but you have to wonder where this is leading. Sorry about this, it is a whole subject. I will start a new thread.
Maybe consider the generic definition of a butterworth crossover filter. Responses are -3dB, power is flat and response is flat on axis. This is connected with phase being out by 90 degrees and the frontal lobe being tilted.