Interesting article:
https://pdfs.semanticscholar.org/0b1f/399269735384e8eca1f1bd288443639dde8f.pdf
https://pdfs.semanticscholar.org/0b1f/399269735384e8eca1f1bd288443639dde8f.pdf
Very interesting thread!
I have read 35 pages of it so far...
I am striving for a large sweetspot. So this pase shuffling seems interesting, making the sound more consistent between the exact main-sweetspot, and other nearby places.
But, I can also see the point of view of gedlee, that this correction would be best done in the mix-stage in the studio, if the effect is wanted by the artist. But maybe headphones have precedence in priority of mixing nowadays?
I have another use-case. I am downmixing 7.1 from movies into 2.0. Then this phase shuffling might be very useful for the C-channel going into L and R, right?
I have read 35 pages of it so far...
I am striving for a large sweetspot. So this pase shuffling seems interesting, making the sound more consistent between the exact main-sweetspot, and other nearby places.
But, I can also see the point of view of gedlee, that this correction would be best done in the mix-stage in the studio, if the effect is wanted by the artist. But maybe headphones have precedence in priority of mixing nowadays?
I have another use-case. I am downmixing 7.1 from movies into 2.0. Then this phase shuffling might be very useful for the C-channel going into L and R, right?
Last edited:
A lot more about the phantom channel fixing here:
http://www.diyaudio.com/forums/full-range/242171-towers-25-driver-range-line-array-new-post.html
Not from the start, you will have to work your way backwards to get all that is written about the phantom channel
http://www.diyaudio.com/forums/full-range/242171-towers-25-driver-range-line-array-new-post.html
Not from the start, you will have to work your way backwards to get all that is written about the phantom channel
I've tried the phase shuffler, which works as advertised. Though not without some minor drawbacks. I also tried digital cross talk cancelation and linear phase mid/side EQ.
The mid/side EQ has my preference. Even though it "should" be fixed in the mix I prefer to have this fix at the listening position. Chances are a near field setup experiences slightly less cross talk which makes me feel this way about it. Would love to be able to verify this one day.
I use the mid/side EQ for movies too, however I use slightly more aggressive values for it. It's obvious there are tonal differences between material mixed for a real center compared to the phantom.
The mid/side EQ I use has been used to compensate for tonal differences between the side and phantom stream. It's not a large fix but it does make a rather big difference (to me). Intelligibility of phantom center voices improved. Background voices became very (even scary) real.
These perceptional differences are largely influenced by reflection levels within the room (I've done the best I could to bring those down). I suggest to experiment yourself and find your own preferred cure. What works for me might not be or feel the same for you. A worthwhile test though.
The mayor drawback for me with the phase shuffler was the imbalance of the seats next to the exact sweet spot. One seat would work almost as well as the sweet spot while the seat on the other side had a shifted tonal balance. Most noticeably with music. It might still work very well with voices in movies. A very critical analysis from me though, I just preferred the wider sweet spot of the linear phase mid/side EQ.
Cross talk cancelation can also be very persuasive. It's more geared for a single listening position. I never found an optimum that held up (for me) over time. I tried two different versions of it, one based on Ambio (without moving the speakers close together, adjusting it to a stereo angle) and one (loosely) based on Prof. Edgar Choueiri's work. I can't afford to try the real thing
.
The mid/side EQ has my preference. Even though it "should" be fixed in the mix I prefer to have this fix at the listening position. Chances are a near field setup experiences slightly less cross talk which makes me feel this way about it. Would love to be able to verify this one day.
I use the mid/side EQ for movies too, however I use slightly more aggressive values for it. It's obvious there are tonal differences between material mixed for a real center compared to the phantom.
The mid/side EQ I use has been used to compensate for tonal differences between the side and phantom stream. It's not a large fix but it does make a rather big difference (to me). Intelligibility of phantom center voices improved. Background voices became very (even scary) real.
These perceptional differences are largely influenced by reflection levels within the room (I've done the best I could to bring those down). I suggest to experiment yourself and find your own preferred cure. What works for me might not be or feel the same for you. A worthwhile test though.
The mayor drawback for me with the phase shuffler was the imbalance of the seats next to the exact sweet spot. One seat would work almost as well as the sweet spot while the seat on the other side had a shifted tonal balance. Most noticeably with music. It might still work very well with voices in movies. A very critical analysis from me though, I just preferred the wider sweet spot of the linear phase mid/side EQ.
Cross talk cancelation can also be very persuasive. It's more geared for a single listening position. I never found an optimum that held up (for me) over time. I tried two different versions of it, one based on Ambio (without moving the speakers close together, adjusting it to a stereo angle) and one (loosely) based on Prof. Edgar Choueiri's work. I can't afford to try the real thing
.
Last edited:
Welcome to the thread, there is a lot to take in.
Put that same signal into a stereo pair of speakers and it will sound dull due to comb filtering - that's explained earlier in the thread. If the down mix isn't doing an EQ on the phantom center, it's going to be duller than a true center speaker. That EQ will be room, system and listener specific, but does follow some general guidelines.
On music systems we generally don't notice the dull center, because the music has been mixed and mastered with 2 channel in mind. The phantom center is king. But if you have a system with good left to right phase matching and suppressed early reflections, you'll notice that left or right alone sound too bright. That might not bother some people, but it does bother me and takes away from the illusion of reality. It more about matching the sides to the center than altering the center.
But even that's beside the point. I want to fix it because it sounds wrong to me on my system. I hear it as an annoying minor artifact of a coherent system that is rarely dealt with in the mix. Therefore, I correct my system to eliminate the artifact.
The final phase shuffler I made worked well on my speakers and room and was not limited to a single spot. The effect is system dependent, and should really be corrected system by system. What works well one one system, may not on another, even if the effect sounds similar.
Yes, and that's the main use in the paper linked to in the beginning, improved dialog clarity. My search was for a better tonal balance across L-C-R for music, but dialog is a big concern. As stated earlier, movies started out with a single speaker behind the screen that pinned the dialog right there where faces usually are. When stereo became popular, dialog was still primarily mixed to the center speaker. And that worked well for cinemas that were still mono - I worked in one up until about 1995. Dialog always in the center speaker.
Put that same signal into a stereo pair of speakers and it will sound dull due to comb filtering - that's explained earlier in the thread. If the down mix isn't doing an EQ on the phantom center, it's going to be duller than a true center speaker. That EQ will be room, system and listener specific, but does follow some general guidelines.
On music systems we generally don't notice the dull center, because the music has been mixed and mastered with 2 channel in mind. The phantom center is king. But if you have a system with good left to right phase matching and suppressed early reflections, you'll notice that left or right alone sound too bright. That might not bother some people, but it does bother me and takes away from the illusion of reality. It more about matching the sides to the center than altering the center.
Yes - "IF". But in over 40 years working with musicians, actors, mixing and master engineers - I can't think of a single one who wanted it. Mostly it would just be considered a blip in an otherwise huge range of other issues.
But even that's beside the point. I want to fix it because it sounds wrong to me on my system. I hear it as an annoying minor artifact of a coherent system that is rarely dealt with in the mix. Therefore, I correct my system to eliminate the artifact.
The final phase shuffler I made worked well on my speakers and room and was not limited to a single spot. The effect is system dependent, and should really be corrected system by system. What works well one one system, may not on another, even if the effect sounds similar.
Is your system altec direct radiator bass with 1505 multicell horns? How big is the room?
It was the horn loaded 15 AKA the 828 cabinet. 1005 ten cell horns, not the 15 cell.
Room was small. See here post #11 here:
A last! My listening room. A good day.
Room was small. See here post #11 here:
A last! My listening room. A good day.
OK, this is what you have now with this shuffler? That's what I'm asking. What do you have now that works with this shuffler? The speakers and room?
Stumbled over this nice USB calibrating test jig head phone tool miniDSP intend release to market end of November, link Acoustic Measurement Tools : EARS, and looking into their notes it looks tool could also be used investigate what this thread is about
I do like that head phone tool from MiniDSP. My only fear is it doesn't represent any head shading that would take place.
To get an idea of what it would do, it would be similar to offset a single microphone half of that ear to ear distance (not quite true, because of head shading) and measure with both channels firing.
A good example of how that works out was provided in this thread of a measurement done by jim1961.
I do like to see this continuous effort to bring more useful gadgets to market.
To get an idea of what it would do, it would be similar to offset a single microphone half of that ear to ear distance (not quite true, because of head shading) and measure with both channels firing.
A good example of how that works out was provided in this thread of a measurement done by jim1961.
I do like to see this continuous effort to bring more useful gadgets to market.
Last edited:
Yeah, it would need the 2-3cc chamber to mimic the ear canal. Hard to know what the purpose is, unless its just to compare different headphones. That does have some value.
Interesting discussion you had going on here, as I recognize the issue. Although for me it's more crystal clear sounds like backing vocals or horns that are hard panned to the left or right that bother me. They tend to break the virtual stereo image by drawing the attention to the speakers.
I was impressed by the effectiveness and simplicity of Pano's shuffler (version 2 that is). It ads some presence to the center image (lead vocals), although very subtle, but what I liked the most is that it fixed the issue I have with the hard panned backing vocals. It draws them away from the speakers and brings their tonality more in balance with the center image.
I made a slight variation to Pano's shuffler. Just to tailer it to my personal needs, I don't claim it's a better one. I used higher Q filters to focus on the comb filter frequencies and reduced the phase for higher frequencies as measurements show the problem decreases in that range.
RePhase settings file: https://1drv.ms/u/s!AsglByBDKzPumED5pE9Q39UYLN3r
I was impressed by the effectiveness and simplicity of Pano's shuffler (version 2 that is). It ads some presence to the center image (lead vocals), although very subtle, but what I liked the most is that it fixed the issue I have with the hard panned backing vocals. It draws them away from the speakers and brings their tonality more in balance with the center image.
I made a slight variation to Pano's shuffler. Just to tailer it to my personal needs, I don't claim it's a better one. I used higher Q filters to focus on the comb filter frequencies and reduced the phase for higher frequencies as measurements show the problem decreases in that range.
RePhase settings file: https://1drv.ms/u/s!AsglByBDKzPumED5pE9Q39UYLN3r
An externally hosted image should be here but it was not working when we last tested it.
Thanks for trying the shuffler and for your comments. I'll take a look at what you've done, sounds like a good approach. I can't remember what my final version of the shuffler was, it's been awhile. Will have to go look.
That was my basic complaint, too.
If yawl didn't know by now I am totally into DML technology.
One of the main factors that drew me to DML technology is the imaging.
I have never heard another speaker produce a so realistic phantom center sound as these DML panels....I also have never heard another speaker produce so realistic of a surround sound with just 2 channels.....A lot of times I have to check to see if I accidently turned on my 5 channel mode because just a pair of DML speakers can mimic both a phantom center and rear surrounds so realistic you would believe all 5 speakers are playing.
Another strong point of DML panels is the VOCALS....That's usually when I get fooled into thinking I left my center channel on because the voice sounds like its only coming from the center.
One of the main factors that drew me to DML technology is the imaging.
I have never heard another speaker produce a so realistic phantom center sound as these DML panels....I also have never heard another speaker produce so realistic of a surround sound with just 2 channels.....A lot of times I have to check to see if I accidently turned on my 5 channel mode because just a pair of DML speakers can mimic both a phantom center and rear surrounds so realistic you would believe all 5 speakers are playing.
Another strong point of DML panels is the VOCALS....That's usually when I get fooled into thinking I left my center channel on because the voice sounds like its only coming from the center.