Yes, vst plugin, used to be sold separately: PCM Native Random Hall | Lexicon Pro - Legendary Reverb and Effects
Now part of a complete bundle, needs an Ilok key.
http://www.davidgriesinger.com/aes99.pdf
Now part of a complete bundle, needs an Ilok key.
http://www.davidgriesinger.com/aes99.pdf
Last edited:
Thanks!
I had a few emails back and forth with Dave about headphone stuff and HRTF. I could never achieve what he was getting, unfortunately.
I had a few emails back and forth with Dave about headphone stuff and HRTF. I could never achieve what he was getting, unfortunately.
Wesayso, I found this post in your thread:
https://www.diyaudio.com/forums/ful...-driver-range-line-array-403.html#post5277331
Very helpful. Some questions:
1. I see you have a broad dip in the Mid (inside the MS loop) centered around 600 Hz, and then a peak at 1.8 kHz, and then a dip at 3.7 kHz and 7.2 kHz. That's what you were referring to in your earlier post here, correct?
2. For the S channel inside the MS loop, are you using the S-curve, i.e, boost centered at 400 Hz and dip centered around 3 kHz?
I am using Voxengo's linear phase EQ for the MS EQ, but I wasn't applying any EQ to the mid. I tried it but it seemed to defeat the improved imaging.
https://www.diyaudio.com/forums/ful...-driver-range-line-array-403.html#post5277331
Very helpful. Some questions:
1. I see you have a broad dip in the Mid (inside the MS loop) centered around 600 Hz, and then a peak at 1.8 kHz, and then a dip at 3.7 kHz and 7.2 kHz. That's what you were referring to in your earlier post here, correct?
2. For the S channel inside the MS loop, are you using the S-curve, i.e, boost centered at 400 Hz and dip centered around 3 kHz?
I am using Voxengo's linear phase EQ for the MS EQ, but I wasn't applying any EQ to the mid. I tried it but it seemed to defeat the improved imaging.
Hi ra7,
Yes, that is about it, for starters... as everything is adjustable to taste. the room target you have will be very important, see further below
I did make tweaks upon it myself.
In the side channels I haven't got a true dip at 3 Khz, but it starts to roll off ever so slightly above 1500 Hz. (basically extending the dip at 3-4 K out to the high frequencies)
I also play with level differences between mid and side and adjust the bass to remain equal in level. Currently my sides are up 1 dB compared to the mid part.
Check with a mic at the ear position with a foam ball acting as a head for shading and see where your specific dips & peaks are by using a Stereo sweep. I've done that part and my numbers are roughly based on that experiment.
I can give no guarantee what works for me needs to be the same for you. It would depend on what the room does etc.
I use a wide range of songs to judge any tweak. You can actually shift/shape the stage with these tweaks.
A song that should show what the level difference + mid/side EQ can bring would be:
"Hey Nineteen - Steely Dan"
While Donald's voice will be in the center (not very forward), the guitar part is past the left speaker position (in front of it). Halfway the song the guitar part will shift to just right of center and jump back eventually.
Once the background singers come in they image to extreme sides and are at about 150 degree angles (way to the side in front of the speakers). With these tweaks they should sound full bodied and real.
Also pick songs that have the same singer panned across the stage. Many Frank Zappa songs have this, like "Find Her Finer" from "Zoot Allures" and many more.
Find some songs that have a (staged forward) singer and see if you can easily follow the words. Actually the album Stripped, from Christina Aguilera has some different mic techniques that made it an interesting test album. Lots of interesting effects on that album for tests everywhere. Stripped Intro should be fun and quite 3D. Can't Hold Us Down is good as a test of positions (+ rapped part, does it sound natural etc.). Impossible has some neat effects featuring a lot of phantom positions plus wide panned background singers. Should be very easy to follow the words and to offset that, Underappreciated was a song that showed a remarkable difference (upgrade) with the mid/side EQ as far as following the words/ focal lines go.
You don't have to like all songs, 🙂 these are just examples that show clear differences.
Steven Stills - Treetop Flyer is my "girl with guitar" song (as I rarely like girl with guitar) and that song should knock you off your feet with the guitar part, as far as being natural and blooming. Steven Stills should sound like being there, I love that presentation.
I have many more, that I use for some reason or another. Mainly to judge what is happening and to find the stage presentation I want.
The usual Audiophile songs, "Hotel California" from "Hell Freezes Over", "Nils Lofgren - Keith Don't Go" should excel, it is not a coincidence these recordings are often used at shows. Pink Floyd and Roger Waters solo (like "Amused to Death") are out of this world.
First and foremost: have fun with it! Start with what works and go from there... Do a little tweak at a time. But remember, the mid content in mid/side EQ does have side panned sounds in that sum. Don't overdo it or you'll be left with a big blur. It's important to keep it within 3 or max 4 dB difference between mid and sides.
The room target you have will also play a large role in all of this. My room target mimics the "Trained Listeners" from the graph presented by Toole:
And the mid/side tweaks are built around that target(*). But even that depends on room specifics, taste of the listener etc. If you have your balance set favoring the phantom prior to using the mid/side EQ that might not work as well.
If all of this does not work out for you, drop back to what you know did work. I have an entirely different scheme I am testing right now, without mid/side EQ but based on the original shuffler and cross talk cancellation. I can get this to sound very close to what I'm doing with mid/side EQ. It's based on what I use for Home Theatre. As Home Theatre does have a distinct center. Maybe try Home theatre test material first and get a good tonal balance there. That should show you where the tweaks should be.
The lack of early reflections is very important though, for the tricks I use. I cannot mention that enough. As that lack of early reflections is what uncovers the tonal differences. If one does not hear tonal differences there's no valid reason to make up for it, right? 🙂
(*) - The total sum of the sides remains very close to that balance with the tweaks I have. It is the phantom part that needs the most help. You can test this with clever IR feeds running through your system. Something like a Stereo IR file with coherent summing left + right IR (phantom only), one with positive left and negative right (no phantom).
Yes, that is about it, for starters... as everything is adjustable to taste. the room target you have will be very important, see further below
I did make tweaks upon it myself.
In the side channels I haven't got a true dip at 3 Khz, but it starts to roll off ever so slightly above 1500 Hz. (basically extending the dip at 3-4 K out to the high frequencies)
I also play with level differences between mid and side and adjust the bass to remain equal in level. Currently my sides are up 1 dB compared to the mid part.
Check with a mic at the ear position with a foam ball acting as a head for shading and see where your specific dips & peaks are by using a Stereo sweep. I've done that part and my numbers are roughly based on that experiment.
I can give no guarantee what works for me needs to be the same for you. It would depend on what the room does etc.
I use a wide range of songs to judge any tweak. You can actually shift/shape the stage with these tweaks.
A song that should show what the level difference + mid/side EQ can bring would be:
"Hey Nineteen - Steely Dan"
While Donald's voice will be in the center (not very forward), the guitar part is past the left speaker position (in front of it). Halfway the song the guitar part will shift to just right of center and jump back eventually.
Once the background singers come in they image to extreme sides and are at about 150 degree angles (way to the side in front of the speakers). With these tweaks they should sound full bodied and real.
Also pick songs that have the same singer panned across the stage. Many Frank Zappa songs have this, like "Find Her Finer" from "Zoot Allures" and many more.
Find some songs that have a (staged forward) singer and see if you can easily follow the words. Actually the album Stripped, from Christina Aguilera has some different mic techniques that made it an interesting test album. Lots of interesting effects on that album for tests everywhere. Stripped Intro should be fun and quite 3D. Can't Hold Us Down is good as a test of positions (+ rapped part, does it sound natural etc.). Impossible has some neat effects featuring a lot of phantom positions plus wide panned background singers. Should be very easy to follow the words and to offset that, Underappreciated was a song that showed a remarkable difference (upgrade) with the mid/side EQ as far as following the words/ focal lines go.
You don't have to like all songs, 🙂 these are just examples that show clear differences.
Steven Stills - Treetop Flyer is my "girl with guitar" song (as I rarely like girl with guitar) and that song should knock you off your feet with the guitar part, as far as being natural and blooming. Steven Stills should sound like being there, I love that presentation.
I have many more, that I use for some reason or another. Mainly to judge what is happening and to find the stage presentation I want.
The usual Audiophile songs, "Hotel California" from "Hell Freezes Over", "Nils Lofgren - Keith Don't Go" should excel, it is not a coincidence these recordings are often used at shows. Pink Floyd and Roger Waters solo (like "Amused to Death") are out of this world.
First and foremost: have fun with it! Start with what works and go from there... Do a little tweak at a time. But remember, the mid content in mid/side EQ does have side panned sounds in that sum. Don't overdo it or you'll be left with a big blur. It's important to keep it within 3 or max 4 dB difference between mid and sides.
The room target you have will also play a large role in all of this. My room target mimics the "Trained Listeners" from the graph presented by Toole:
And the mid/side tweaks are built around that target(*). But even that depends on room specifics, taste of the listener etc. If you have your balance set favoring the phantom prior to using the mid/side EQ that might not work as well.
If all of this does not work out for you, drop back to what you know did work. I have an entirely different scheme I am testing right now, without mid/side EQ but based on the original shuffler and cross talk cancellation. I can get this to sound very close to what I'm doing with mid/side EQ. It's based on what I use for Home Theatre. As Home Theatre does have a distinct center. Maybe try Home theatre test material first and get a good tonal balance there. That should show you where the tweaks should be.
The lack of early reflections is very important though, for the tricks I use. I cannot mention that enough. As that lack of early reflections is what uncovers the tonal differences. If one does not hear tonal differences there's no valid reason to make up for it, right? 🙂
(*) - The total sum of the sides remains very close to that balance with the tweaks I have. It is the phantom part that needs the most help. You can test this with clever IR feeds running through your system. Something like a Stereo IR file with coherent summing left + right IR (phantom only), one with positive left and negative right (no phantom).
Attachments
Last edited:
The tweaks that work for Home Theatre (played back without an actual center) are the same that work well for Stereo, Stereo just needs a little less compensation.
A fun test movie I often use is Rango, as that has voices spanning the entire phantom width between speakers.
A fun test movie I often use is Rango, as that has voices spanning the entire phantom width between speakers.
FWIW, I don't remember any advantages to the image on my system, just bettertonal balance.
Might have to go back thru the thread to see if I mentioned any.
Might have to go back thru the thread to see if I mentioned any.
Tonal balance is probably the most important for me, personally. But I've grown to appreciate both. With what I have I can shape it to fit my environment. Hiding my room and let the music speak.
I still need to work on envelopment again. I do have it but I did have more at one point. I'd like to get that back, if I can remember all the tweaks I did at that point in time (lol).
I still need to work on envelopment again. I do have it but I did have more at one point. I'd like to get that back, if I can remember all the tweaks I did at that point in time (lol).
And tell me, which #^$&! version of JRiver do you have to own to have M/S encode AND decode? I just can't keep up with all their new versions. I'm always about 4 behind.
I still run version 22 and it seems to have both. Never used it though. Either I make my own sum/subtraction or I use the free Voxengo MSED plugin.
I do have a license for the newest JRiver, but didn't run it yet. Setting it up entirely within JRiver would also need to include linear phase EQ.
I do have a license for the newest JRiver, but didn't run it yet. Setting it up entirely within JRiver would also need to include linear phase EQ.
Last edited:
Thanks for the many tips and hints! I tried with the cuts you suggested for the M channel and that made a big difference. The tonal balance is much nicer. I'm keepin' it!
I usually starts with 'Birds on a Wire' and then move on to some choral and orchestral works that have depth.
Tell us more! I have been looking at the shuffler that member Pos implemented into RePhase, which is based on Pano's original shuffler. That together with some phase deltas in the sub 700 Hz range. Produces interesting effects. Would like to know what you've been experimenting with.
Another interesting thing is that with the arrays+Fountek Neo3.5H, the MS EQ actually makes things worse. Very interesting.
I usually starts with 'Birds on a Wire' and then move on to some choral and orchestral works that have depth.
Tell us more! I have been looking at the shuffler that member Pos implemented into RePhase, which is based on Pano's original shuffler. That together with some phase deltas in the sub 700 Hz range. Produces interesting effects. Would like to know what you've been experimenting with.
Another interesting thing is that with the arrays+Fountek Neo3.5H, the MS EQ actually makes things worse. Very interesting.
My understanding of the center image issue, when there's no center speaker, is that in the upper midrange (1kHZ - maybe 8kHZ), the probability of comb filter cancellation at the listeners head, goes up, rendering the center image blurred, phasy, lacking clarity. We are talking about wavelenths that are between about 12 inches to about one inch, which means listening position gets critical, and is pretty much never going to be without cancellation issues at the couch. Added to that, the fact that we have 2 ears, each of which will receive a different "signature" of the comb filter cancellation effect. Because the exact FR will vary significantly for small changes in the listening position, I don't know how to correct this anywhere near perfectly. My best guess might be to accentuate the FR between 1kHZ and 8kHZ in order to get the center image energy back up to where it belongs, which throws off the FR of the L and R images... Not a real fix, but there's only so much I know how to do about this. I'm not familiar with some of the Processing you guys are playing with, and hesitate to make things more complicated, but adding a tiny bit of very short term reverb to the L+R, in this FR (1kHZ - 8kHZ), and not the L-R or R-L, intrigues me, because it might actually fix this issue better than anything I can think of. It might effectively decorrelate the center image a bit and somewhat perceptually fill in the comb filter cancellations, making it have a more solid sense of space. Coarse then you need a circuit that separates L+R from not L+R, and then re-combines them without new artifacts.
That's a step beyond adding phase shuffle to left and right, as was done at the beginning of this thread. Basically, you would be adding just one shuffle to the center, and I don't know if that would stop the comb filtering. Maybe. The stereo shuffle changes the phase between left and right eliminating the comb filtering that changes the tonal balance and causes center dialog to suffer.
I think much of the image blurriness simply comes from deep nulls at frequencies we need for localization. There may be other things going on, but I don't know.
I think much of the image blurriness simply comes from deep nulls at frequencies we need for localization. There may be other things going on, but I don't know.
I did some experimentation this weekend and was impressed! Excited about doing more.
I had very strongly looked into implementing a Trifield setup recently (Meridian from Gerzon), but didn't pursue it (wanted an identical center speaker and mine NLA, among other things).
My flow (using AUs in Audirvana):
1. L/R to M/S, +1.5db on the side
2. EQ:
Mid EQ: 400Hz, Q 0.8
3500Hz, Q 0.58
Side EQ: Same frequencies, Q
3. M/S to L/R to DAC (RME) where I have a couple of low frequency EQ bands for residual room modes
I currently have +2b at 400Hz and - 2db at 3500Hz on the side channel. I like the widening of the soundstage with the side level boost (less clustering around the speakers), and the overall tonal balance.
I tried inverse values on the mid channel, but found that centered vocalists didn't have enough "chest" sounds. Maybe I am just used to them sounding that way(?).
I will keep playing!
Bill
I had very strongly looked into implementing a Trifield setup recently (Meridian from Gerzon), but didn't pursue it (wanted an identical center speaker and mine NLA, among other things).
My flow (using AUs in Audirvana):
1. L/R to M/S, +1.5db on the side
2. EQ:
Mid EQ: 400Hz, Q 0.8
3500Hz, Q 0.58
Side EQ: Same frequencies, Q
3. M/S to L/R to DAC (RME) where I have a couple of low frequency EQ bands for residual room modes
I currently have +2b at 400Hz and - 2db at 3500Hz on the side channel. I like the widening of the soundstage with the side level boost (less clustering around the speakers), and the overall tonal balance.
I tried inverse values on the mid channel, but found that centered vocalists didn't have enough "chest" sounds. Maybe I am just used to them sounding that way(?).
I will keep playing!
Bill
Thanks for joining Bill, that's what it's all about! I hope you're having fun with these experiments. I know I did!
To answer ra7: my latest tweaks involve the cross talk compensation, so a negative small peak in each mid signal (later to be split into left and right again) that creates very similar EQ results as the mid/side EQ I suggested. The cross talk signal (negative peak) is a linear phase band passed signal targeting the cross talk dips as I've measured them. Above a certain frequency head shading prevents the cross talk and below the mids we have near omni results that shade the effect as well.
It does act a little different, more pronounced vocals, excellent for movies. Too little listening time for a verdict on stereo.
I did try to add small reverb to the above but even without it can be powerful.
To answer ra7: my latest tweaks involve the cross talk compensation, so a negative small peak in each mid signal (later to be split into left and right again) that creates very similar EQ results as the mid/side EQ I suggested. The cross talk signal (negative peak) is a linear phase band passed signal targeting the cross talk dips as I've measured them. Above a certain frequency head shading prevents the cross talk and below the mids we have near omni results that shade the effect as well.
It does act a little different, more pronounced vocals, excellent for movies. Too little listening time for a verdict on stereo.
I did try to add small reverb to the above but even without it can be powerful.
Last edited:
Thanks for testing and reporting Bill. 
Of course for center channel dialog it's another thing.
I think so, and a lot of that might be because they were mixed and mastered that way. For me, it was always the sides that were too bight when the center sounded right. At least for music.
Of course for center channel dialog it's another thing.
I would expect that changing phase, using a shuffler, between left and right, would mess with image location below about 1.5kHZ. Maybe above that frequency too. Am I understanding that correctly? What if you use very short term reverb (several mono delays) on the extracted L+R signal, that is designed to add a comb filter effect that is very similar to the one you necessarily get because of the physical arrangement of the 2 speaker system, and because the two would have slightly differing delay times, and therefore null frequencies, they might tend to largely fill in each others nulls, giving the phantom center image better solidity and clarity. This would be analogous to how typical room acoustics comb filter effects, somewhat fill in each others nulls in the frequencies above about 1kHZ (?), because of the varied delay times of the many room reflections.
With a very short reverb on the separated L+R only, with strategic delay times (to have the potential of filling in the original comb filter nulls in the 1kHZ - 8kHZ FR), it seems like the result might be like high frequency room reflections largly filling in each others nulls, because of the differing delay times, leaving you with a relatively flat FR.
With a very short reverb on the separated L+R only, with strategic delay times (to have the potential of filling in the original comb filter nulls in the 1kHZ - 8kHZ FR), it seems like the result might be like high frequency room reflections largly filling in each others nulls, because of the differing delay times, leaving you with a relatively flat FR.
I don't remember it messing with the location at all. But maybe on some systems. Some others will chime in with their own experiences.
The original shuffler as posted is basically a series of short reverbs with differing phase left and right. You could download that file and use just one channel applied to L+R. Or maybe apply one to mid and the other to side. It's easy to experiment!
There are also some test files posted that allow you to find out if you actually have the tonal shift on your systems. Many people in live rooms do not. The shuffler is built into the room. 🙂
The original shuffler as posted is basically a series of short reverbs with differing phase left and right. You could download that file and use just one channel applied to L+R. Or maybe apply one to mid and the other to side. It's easy to experiment!
There are also some test files posted that allow you to find out if you actually have the tonal shift on your systems. Many people in live rooms do not. The shuffler is built into the room. 🙂
What you're describing, Bob, is very close to what the original article was all about. With one difference. It was meant to be applied to the center channel of a multi-channel setup for systems that don't have an actual center but play it as a phantom center. The shuffler will enhance the phantom parts.
Stereo changes the rules somewhat. With an L+R sum, there is a fixed amount of side (L and R) channel in that sum. So whatever you do to the L+R signal will partially influence the side channels as well. This rule also applies to using mid/side EQ or cross talk cancelation.
It does make this game a little less straight forward though. In that regard it is easier to experiment with a Multichannel feed, as that has the dedicated center material and was mixed for that.
Within Stereo it won't ever be easy to get back the true 'phantom only', the question remains, do we need to go that far... I've done a lot of experiments, as the results are very promising.
Like Pano, when I had setup my system in my room, with reduced early reflections, the side panned sounds seemed bright compared to the phantom sounds, tonally. However, the fix was a little different than I first anticipated. The mid/side EQ should give side panned sounds a different tonal balance. Less bright/thin, more 'body' is how I would describe it. But I needed to rebalance everything, because what I had before was setup for great phantom sound, tonaly. That's where most of the action is. With a bit of care and probably a lot of time, you can find a balance that works well throughout the stage. From side panned sounds through the phantom center. More balanced than without these tweaks.
Improvements or changes in imaging I have encountered I'd say are a pleasant by product of the better balance throughout. Every way that filled in the ~1850 Hz cross talk dip has had a more true to life presentation, even the 'phase only' shuffler worked for that. However the 'phase only' shuffler changed the perceived balance of left and right for me. Favouring one side over the other. Mid/side EQ is the simplest way to get improvements, but I do expect there's more to be had with further tweaks/tricks to fill in those cross talk dips.
Only at the sweet spot do we get these dips at exactly the same frequency at both ears. Move a little and everything changes. Lots of simulations presented in pictures in this thread about that! 🙂
So far, the mid/side EQ has been able to run the longest in my listening room/living room. Even late reflections, still within the Haas limit can help fill the cross talk dips and thus help in perception of the stage. Making it more believable. That is and will remain my primary goal.
Stereo changes the rules somewhat. With an L+R sum, there is a fixed amount of side (L and R) channel in that sum. So whatever you do to the L+R signal will partially influence the side channels as well. This rule also applies to using mid/side EQ or cross talk cancelation.
It does make this game a little less straight forward though. In that regard it is easier to experiment with a Multichannel feed, as that has the dedicated center material and was mixed for that.
Within Stereo it won't ever be easy to get back the true 'phantom only', the question remains, do we need to go that far... I've done a lot of experiments, as the results are very promising.
Like Pano, when I had setup my system in my room, with reduced early reflections, the side panned sounds seemed bright compared to the phantom sounds, tonally. However, the fix was a little different than I first anticipated. The mid/side EQ should give side panned sounds a different tonal balance. Less bright/thin, more 'body' is how I would describe it. But I needed to rebalance everything, because what I had before was setup for great phantom sound, tonaly. That's where most of the action is. With a bit of care and probably a lot of time, you can find a balance that works well throughout the stage. From side panned sounds through the phantom center. More balanced than without these tweaks.
Improvements or changes in imaging I have encountered I'd say are a pleasant by product of the better balance throughout. Every way that filled in the ~1850 Hz cross talk dip has had a more true to life presentation, even the 'phase only' shuffler worked for that. However the 'phase only' shuffler changed the perceived balance of left and right for me. Favouring one side over the other. Mid/side EQ is the simplest way to get improvements, but I do expect there's more to be had with further tweaks/tricks to fill in those cross talk dips.
Only at the sweet spot do we get these dips at exactly the same frequency at both ears. Move a little and everything changes. Lots of simulations presented in pictures in this thread about that! 🙂
So far, the mid/side EQ has been able to run the longest in my listening room/living room. Even late reflections, still within the Haas limit can help fill the cross talk dips and thus help in perception of the stage. Making it more believable. That is and will remain my primary goal.
Last edited:
Wesayo,
I agree re. the M/S EQ. Though we are always prone to responding to the new with "wow, amazing" then have it wear off, I am having a hard time picturing that happening at this point. I am thinking of three parameters to approach at this point:
- M/S EQ, "fleshing out" off-center images/making less thin
- A mild boost in the S to increase spaciousness a bit, free/detach the lateral images from the speakers, and
- Comb-filtering because of interaural crosstalk
I have been reading lots of papers on the last and gone down the rabbit hole searching for a plugin that will allow me try a subtle phase shift or time delay between the L and R channels of the Mid, maybe 10ms or so above about 1.5kHz to affect the strongest destructive interference at 1.8-2kHz(?). Not sure that would be fruitful.
I followed this thread when it started for a time then lost track. Working my way forward again looking for nuggets.
I have also followed a thread on this company- room correction and crosstalk cancellation filters:
Loudspeakers, room and headphones digital correction services - Home Audio Fidelity
Bill
I agree re. the M/S EQ. Though we are always prone to responding to the new with "wow, amazing" then have it wear off, I am having a hard time picturing that happening at this point. I am thinking of three parameters to approach at this point:
- M/S EQ, "fleshing out" off-center images/making less thin
- A mild boost in the S to increase spaciousness a bit, free/detach the lateral images from the speakers, and
- Comb-filtering because of interaural crosstalk
I have been reading lots of papers on the last and gone down the rabbit hole searching for a plugin that will allow me try a subtle phase shift or time delay between the L and R channels of the Mid, maybe 10ms or so above about 1.5kHz to affect the strongest destructive interference at 1.8-2kHz(?). Not sure that would be fruitful.
I followed this thread when it started for a time then lost track. Working my way forward again looking for nuggets.
I have also followed a thread on this company- room correction and crosstalk cancellation filters:
Loudspeakers, room and headphones digital correction services - Home Audio Fidelity
Bill