I came up with something helpful (I think). First, bear with me.
Run a measurement with ambient channels engaged and call that "A". Now run it again without the ambient channels and call that "B". Now in REW, do a (A-B) and call that "C".
What "C" represents is the ambient information only. Whats nifty about this is you can call it up on a spectrogram (see attachment) and in one snapshot, see the relative spectra in time. You can see clearly when it comes into being, in my example about 27ms, but also how it decays.
This same formula could be used to analyze the kicker.
Run a measurement with ambient channels engaged and call that "A". Now run it again without the ambient channels and call that "B". Now in REW, do a (A-B) and call that "C".
What "C" represents is the ambient information only. Whats nifty about this is you can call it up on a spectrogram (see attachment) and in one snapshot, see the relative spectra in time. You can see clearly when it comes into being, in my example about 27ms, but also how it decays.
This same formula could be used to analyze the kicker.
Attachments
Last edited:
Explain.
edit: When I do the A-B, I get the exact same result in magnitude as running the ambient channel by itself. This seems self confirming that A-B provides the difference. What A-B gives you that a simple ambient channel sweep by itself wont is the proper time index.
Last edited:
If you run a sweep with ambient channels only, REW corrects the first wavefront to T=0, even when in reality, the first wavefront from such a source may not enter the equation until 26ms (for example in my case) when compared to the time arrival of the mains.
Where "A" = Mains + ambient and "B" = Mains only, A-B lets you see the ambient content WHEN it arrives.
Thanks that is clever tip, think reason that REW can't show real timing is because input "USB mic" run on separate physical clock relative to soundcard output and that will provide small mismatch from sweep to sweep therefor the zero align IR thing. Recently changed to analog mic setup because of above so to have synced real time of flight info, it now uses left channel I/O as measurement chain and right channel I/O is physical loopback wired, because all I/O is on same clock REW can have ticked "Use loopback as timing reference" which ensures relative correct aligned IR timing from sweep to sweep and also info about time of flight.
I came up with something helpful (I think). First, bear with me.
Run a measurement with ambient channels engaged and call that "A". Now run it again without the ambient channels and call that "B". Now in REW, do a (A-B) and call that "C".
What "C" represents is the ambient information only. Whats nifty about this is you can call it up on a spectrogram (see attachment) and in one snapshot, see the relative spectra in time. You can see clearly when it comes into being, in my example about 27ms, but also how it decays.
This same formula could be used to analyze the kicker.
A very useful trick indeed! Especially for the current experiments. Even though I had used the functionality before I didn't think about using it like this, thanks for the idea!
In order to fix the high frequency lag on my ambient channels I was supposed to take some measurements yesterday. Only one problem. I sat down on the couch and started listening... that never ended... until it was time to do other things.
Today Jan Fekkes dropped by for a cup of coffee. I told him I had redone all processing, both the mains and the ambient channels, due to upgrading to JRiver V21. As long as he was there I figured to get his opinion on the new ambient sounds.
First we listened to the ambient speakers only, to give him an idea of what was going on there. We compared the old dry sound to the new type of ambience processing with about 30% added reverb.
Next I switched on the mains and switched off the ambient channels and let him listen for a while. Half way the song I turned on the ambient channels. He couldn't believe it could make that much difference in perception!
We tried it again on another song. Without the ambience, you get great sound with the stage clear in front of you. Good imaging etc.
But with the ambience added something happens... It's still very much in front of you, same positions but somehow it feels like you're in the middle of it. At least that's like Jan worded it. In the middle of the music without the stage physically moving. It images at the same spots, just way more involving. In other words, an added feel of envelopment. (much like listening in a better room, I hope)
You would never know it can make that much difference judging the sound that comes out of those ambient channels by themselves. Crazy how our mind works...
I still need to fix that timing, I did use a crude fix with some all pass filters.
It looks like the reverb is here to stay as well. I shortened and lowered the levels compared to earlier tests but it's quite captivating. I'm not completely done yet, it still needs some fine tuning.
For me it was worth it to try. In small doses it can really enhance that front stage. High doses are pleasing too, but take over and dominate the perception of that room that's added. Just a hint is enough to hear a distinct difference but still maximize/enhance that stage in the recording. Don't overdo it.
Today Jan Fekkes dropped by for a cup of coffee. I told him I had redone all processing, both the mains and the ambient channels, due to upgrading to JRiver V21. As long as he was there I figured to get his opinion on the new ambient sounds.
First we listened to the ambient speakers only, to give him an idea of what was going on there. We compared the old dry sound to the new type of ambience processing with about 30% added reverb.
Next I switched on the mains and switched off the ambient channels and let him listen for a while. Half way the song I turned on the ambient channels. He couldn't believe it could make that much difference in perception!
We tried it again on another song. Without the ambience, you get great sound with the stage clear in front of you. Good imaging etc.
But with the ambience added something happens... It's still very much in front of you, same positions but somehow it feels like you're in the middle of it. At least that's like Jan worded it. In the middle of the music without the stage physically moving. It images at the same spots, just way more involving. In other words, an added feel of envelopment. (much like listening in a better room, I hope
)You would never know it can make that much difference judging the sound that comes out of those ambient channels by themselves. Crazy how our mind works...
I still need to fix that timing, I did use a crude fix with some all pass filters.
It looks like the reverb is here to stay as well. I shortened and lowered the levels compared to earlier tests but it's quite captivating. I'm not completely done yet, it still needs some fine tuning.
For me it was worth it to try. In small doses it can really enhance that front stage. High doses are pleasing too, but take over and dominate the perception of that room that's added. Just a hint is enough to hear a distinct difference but still maximize/enhance that stage in the recording. Don't overdo it.
WESAYSO:
Since we are working on identical concurrent ambient projects (although implemented in very different ways), there are a couple things about what your doing id like to be more clear on.
1) Do you have a room drawing showing where your ambient speakers are located?
2) How much of the left signal is mixed into the right and visa versa?
3) Do you have mdats of the ambient contribution(s) (I know you dont want to share them regarding your mains, but the ambients?)
Since we are working on identical concurrent ambient projects (although implemented in very different ways), there are a couple things about what your doing id like to be more clear on.
1) Do you have a room drawing showing where your ambient speakers are located?
2) How much of the left signal is mixed into the right and visa versa?
3) Do you have mdats of the ambient contribution(s) (I know you dont want to share them regarding your mains, but the ambients?)
Right now I can only answer question no 2.
My ambient channels are left minus right for the left channel and right minus left for the right channel. Both channels get (left plus right) mixed in at a lower level and slightly ahead in time (3 ms). This result get's 30% of reverb signal mixed in with emphasis on the late reverb.
I'll work on a drawing, I had planned that a while ago but haven't made one yet. For the mdats let me first clean up the ambient channel's timing.
My ambient channels are left minus right for the left channel and right minus left for the right channel. Both channels get (left plus right) mixed in at a lower level and slightly ahead in time (3 ms). This result get's 30% of reverb signal mixed in with emphasis on the late reverb.
I'll work on a drawing, I had planned that a while ago but haven't made one yet. For the mdats let me first clean up the ambient channel's timing.
Is there a particular reason you came to this formulation? Is it based on some psycho-acoustical theory?
I am asking because I began my experiment with the left ambient channel getting all left with no right, and visa versa. But i noticed some listening irregularities. I could sometimes hear the decay from the left, pan to the left, and ditto for the right channel. Then I realized that some amount of crosstalk was needed (some left in the right ambient channel and ditto for the left ambient channel). So I added another speaker in each of my ambient adjacent rooms and had them fed from the opposite channel. So for me, it works out something like this:
Right ambient:
Right channel magnitude -15db from direct
Left channel magnitude -22db from direct
Same, but reversed for Left ambient.
Put another way, the opposite channel mix level is at -7db.
Thought id share to see if where I am at translated into what you have done.
Last edited:
The reason for me to pick the L-R and R-L is what I remembered from my Car audio days. I've linked to that info a few times I think. There was a smart fellow over at DIYmobileaudio.com who posted under the name "werewolf" (an unregistered user, but google can find the posts), and a bit later as "Lycan" (still registered, but no posts after a certain eclipse, with a goodbye he announced). He supplied us over there with lots of interesting lessons. Interested as I was I have looked up almost every post by him (probably all of them), and adopted his views on ambient channels for my own experiments.
Here's an example thread with lots of his input: Rear Fill: do you use it? - Car Audio | DiyMobileAudio.com | Car Stereo Forum (still posting as "werewolf" on that thread)
The L-R is basically the left channel with inverted right channel. The other side has the opposite, meaning the ambient channels would be without the phantom center information as that gets nulled in the difference sum. That's why I mixed some of it back in on a lower level. Primarily to avoid the phantom center from sounding too different from the sides in tonality etc.
I've played with pure L and Pure R, but it didn't work as well. The difference signal successfully de-correlates the ambience info from the mains. As even Griesinger mentioned the importance of de-correlation I figured it might work.
You could try and invert the signal feeding the extra speakers you added and find out if that works for you in any shape or form. That would/could be similar I guess.
So far this signal seems to work best for me. I've also tried without a kicker and just going with the late reverb (100% wet) based on a normal stereo input. But I reverted back to 30% and my old channel mix. So far this just works better.
I couldn't tell you the levels I have right now. All I know is that I reduced the kicker SPL and added the reverb. I notice them less even if I'm very close to the ambient channel.
I never notice them at the regular listening position. Except when my reverb tail is too long. If there's a echo like reverb in the recording mixed in that would make it go to the back instead of bouncing off a side wall (in the recording) like before. A shorter tail remedies that. Lots of playing around still. But it has potential over the Haas Kicker by itself. The atmosphere in some songs is wonderful. Songs like Infected Mushroom - Pink Froid and Demons of Pain. It works on most material I've tried so far to add that sense of envelopment. But the above songs (probably because they are created in a synthetic way and not use many real room recordings) are a thrill. More so than a Haas Kicker by itself. It made them come alive. My opinion of coarse.
Edit: thinking about it some more your current setup with R added to L and vice versa might provide you with more phantom center in the ambience than you'd want to have. Worth an experiment I'd say. Easy enough to swap the leads on the cross talk speakers you added, just swap + and -.
Here's an example thread with lots of his input: Rear Fill: do you use it? - Car Audio | DiyMobileAudio.com | Car Stereo Forum (still posting as "werewolf" on that thread)
The L-R is basically the left channel with inverted right channel. The other side has the opposite, meaning the ambient channels would be without the phantom center information as that gets nulled in the difference sum. That's why I mixed some of it back in on a lower level. Primarily to avoid the phantom center from sounding too different from the sides in tonality etc.
I've played with pure L and Pure R, but it didn't work as well. The difference signal successfully de-correlates the ambience info from the mains. As even Griesinger mentioned the importance of de-correlation I figured it might work.
You could try and invert the signal feeding the extra speakers you added and find out if that works for you in any shape or form. That would/could be similar I guess.
So far this signal seems to work best for me. I've also tried without a kicker and just going with the late reverb (100% wet) based on a normal stereo input. But I reverted back to 30% and my old channel mix. So far this just works better.
I couldn't tell you the levels I have right now. All I know is that I reduced the kicker SPL and added the reverb. I notice them less even if I'm very close to the ambient channel.
I never notice them at the regular listening position. Except when my reverb tail is too long. If there's a echo like reverb in the recording mixed in that would make it go to the back instead of bouncing off a side wall (in the recording) like before. A shorter tail remedies that. Lots of playing around still. But it has potential over the Haas Kicker by itself. The atmosphere in some songs is wonderful. Songs like Infected Mushroom - Pink Froid and Demons of Pain. It works on most material I've tried so far to add that sense of envelopment. But the above songs (probably because they are created in a synthetic way and not use many real room recordings) are a thrill. More so than a Haas Kicker by itself. It made them come alive. My opinion of coarse
.Edit: thinking about it some more your current setup with R added to L and vice versa might provide you with more phantom center in the ambience than you'd want to have. Worth an experiment I'd say. Easy enough to swap the leads on the cross talk speakers you added, just swap + and -.
Last edited:
Matrix Surround for Music
I think something along the lines of this article (link) and attached pic is what you guys are talking about?
I did just try swapping the polarity of the crosstalk speakers. It did seem to enhance the sense of space. But since I am doing this acoustically rather than electronically, I am not achieving a true R-L/L-R. In fact, I walked into the ambient rooms and could still clearly hear the vocals (mono) although they were a bit diminished and seemingly "phased" a bit.
I thank you both because it hadn't occurred to me that mono content in the ambient channels could adversely affect what I am trying to do.
On the other hand, these matrix models are designed around the source being about 120 degrees to the rear on both sides where in my case, the ambient content arrives at 90 degrees instead (directly left and right from the LP point of view). Not sure how that affects the desired model.
I think something along the lines of this article (link) and attached pic is what you guys are talking about?
I did just try swapping the polarity of the crosstalk speakers. It did seem to enhance the sense of space. But since I am doing this acoustically rather than electronically, I am not achieving a true R-L/L-R. In fact, I walked into the ambient rooms and could still clearly hear the vocals (mono) although they were a bit diminished and seemingly "phased" a bit.
I thank you both because it hadn't occurred to me that mono content in the ambient channels could adversely affect what I am trying to do.
On the other hand, these matrix models are designed around the source being about 120 degrees to the rear on both sides where in my case, the ambient content arrives at 90 degrees instead (directly left and right from the LP point of view). Not sure how that affects the desired model.
