Thanks Wesyso for kicking off the thread again. 
I had let it go to sleep because I was off in Canada for work. And it takes some time for these ideas to digest. 
I need to read your thread and figure out how you are separating, EQing, then recombining the center. That was my first idea, before the phase shuffler came along.
@youknowyou
The room does make a difference, for sure. But I beleive from my experiences, measurements and reading the phase shuffler paper that the better the phase response (or at least the more identical) of the two speakers, the worse the tonal dip. Better phase tracking equals more comb filtering. It does also mean a tighter, better phantom center - no doubt about that. But often the trade off is the tonal shift.
For me it was just one of those things. My system was working better and better and sounding very good. Getting the sides to match the tone of the center was one of those "last step" improvements.
I had let it go to sleep because I was off in Canada for work. And it takes some time for these ideas to digest. I need to read your thread and figure out how you are separating, EQing, then recombining the center. That was my first idea, before the phase shuffler came along.
@youknowyou
The room does make a difference, for sure. But I beleive from my experiences, measurements and reading the phase shuffler paper that the better the phase response (or at least the more identical) of the two speakers, the worse the tonal dip. Better phase tracking equals more comb filtering. It does also mean a tighter, better phantom center - no doubt about that. But often the trade off is the tonal shift.
For me it was just one of those things. My system was working better and better and sounding very good. Getting the sides to match the tone of the center was one of those "last step" improvements.
Pano,
I find this all very interesting though I haven't tried the shuffler, my system is in pieces and I am working with a single channel for test purposes only so no way to play with your ideas. But your comment about the problem being worse with increased accuracy of phase response from two speakers made me wonder what this would be like with an omnidirectional speaker such as the MBL or any other omnidirectional type speaker? Would there be an ideal phase null with a perfect set of omni speakers or is it only a problem with two very discrete and directional speakers? Now the idea of the very long line arrays would seem to have such a different directivity and since you can steer those through other means I don't even know how you could compare the two totally different polar type of patterns you would have as far as the phantom center channel is concerned. I've never really thought about it the way you are discussing on this thread, a real eye opener I must say. I guess I always just assumed that the center phantom image was due to a center panned recording with equal energy in each channel, mono signal really, and not the combination of left and right channels through comb filter effects.
I find this all very interesting though I haven't tried the shuffler, my system is in pieces and I am working with a single channel for test purposes only so no way to play with your ideas. But your comment about the problem being worse with increased accuracy of phase response from two speakers made me wonder what this would be like with an omnidirectional speaker such as the MBL or any other omnidirectional type speaker? Would there be an ideal phase null with a perfect set of omni speakers or is it only a problem with two very discrete and directional speakers? Now the idea of the very long line arrays would seem to have such a different directivity and since you can steer those through other means I don't even know how you could compare the two totally different polar type of patterns you would have as far as the phantom center channel is concerned. I've never really thought about it the way you are discussing on this thread, a real eye opener I must say. I guess I always just assumed that the center phantom image was due to a center panned recording with equal energy in each channel, mono signal really, and not the combination of left and right channels through comb filter effects.
I don't know and suppose I will need more coffee to think about it, or it will injure my brain.
One would imagine that as long as the phase (within a short time window) is closely matched when it gets to your head, you'll get comb filtering. Directivity should not matter.
Where things could get different is in early reflections.
I'm going over to the orchid show in Hilo. I shall stare at amazing flowers and ponder the question.
Thanks Wesyso for kicking off the thread again.
I need to read your thread and figure out how you are separating, EQing, then recombining the center. That was my first idea, before the phase shuffler came along.
This image might clear it up:
Using Metaplugin as a VST host and Voxengo's free MSED plugin as Encoder and Decoder lets me use (parametric) EQ (in between) on the mid channel only within JRiver.
Confirmed it was working by cutting the channel paths one by one.
As I EQ the comb results of each speaker using DRC I figured it should work on the centre channel as well.
Not sure what you're asking here...
I've experimented with IIR EQ and FIR filter based EQ. The FIR filter based EQ wins for me. The straight line I posted is not my preferred in room response though. I have a downward slope from low frequencies to high that works best in my room. I played with a lot of ideas and targets. So far this one wins for me.
Last edited:
Cool! That does help and I will look into it.
The paper I linked to says that eliminating the comb filtering is a better approach than EQ, but remains to be heard IMO. Nice to have some different methods to poke at it with.
The thing your phase shuffler does better is making the centre image more stable when moving side to side. I have tried to use an extra convolver within JRiver and figured I could only use the shuffler on the mid signal and combine the results with the sides without actually applying the shuffler on the side channel itself. Haven't managed to get that working though. Somehow I can't successfully run an extra convolver without messing everything up.
The orchids were stunning - of course! They don't call this "The Orchid Isle" for nothing.
As for work, I know a guy here who maintains the internal website of a very large company based in the midwest. He basically said to them "Either I telecommute from Hawaii, or I quit." He was lucky, they let him do it. He sits on his lanai looking out over the vast Pacific ocean doing web work. That's the dream come true, right?
Pano,
I've often contemplated just opening a small little shop over there and making custom speakers for the tourist with local Koa wood and things like that. Do design work there and just have the production done on the mainland. If I can get my speakers done and on the market then perhaps I can pull it off. I haven't been there in years, but last time I was there I sat on the beach and worked on my laptop doing design work and writing quotations. I sent them back to my office and nobody was the wiser that I wasn't sitting in California. I almost didn't go back that time, I was real close but I just couldn't do it then. Then I met my wife and that screwed everything up! She's passed away so besides the two kids left, and the one should be taking his physical next week I should be down to just my daughter. She is following in my footsteps and working on starting her own business so I have a feeling she will do just fine. Then I can look at getting out of here, I've had enough of So Cal.
I've often contemplated just opening a small little shop over there and making custom speakers for the tourist with local Koa wood and things like that. Do design work there and just have the production done on the mainland. If I can get my speakers done and on the market then perhaps I can pull it off. I haven't been there in years, but last time I was there I sat on the beach and worked on my laptop doing design work and writing quotations. I sent them back to my office and nobody was the wiser that I wasn't sitting in California. I almost didn't go back that time, I was real close but I just couldn't do it then. Then I met my wife and that screwed everything up! She's passed away so besides the two kids left, and the one should be taking his physical next week I should be down to just my daughter. She is following in my footsteps and working on starting her own business so I have a feeling she will do just fine. Then I can look at getting out of here, I've had enough of So Cal.
Interesting topic.
I wonder if LoCo (localisation correction) and Acourate FLOW try to solve the same problem as described in the paper by using a different approach?
I wonder if LoCo (localisation correction) and Acourate FLOW try to solve the same problem as described in the paper by using a different approach?
LoCo said:
Flow said:
interesting!
with jriver, do you guys have the
-normalize filter volume checked?
-automatically switch filter based on input checked?
Ive tried playing with the file in Jriver. very interesting and promising. indeed, the spatiality, and tonality of the phantom image seemed more natural, even maybe more detailed, more full sounding.
however, I have serious phase shift which cancel all the advantage...
with jriver, do you guys have the
-normalize filter volume checked?
-automatically switch filter based on input checked?
Ive tried playing with the file in Jriver. very interesting and promising. indeed, the spatiality, and tonality of the phantom image seemed more natural, even maybe more detailed, more full sounding.
however, I have serious phase shift which cancel all the advantage...