Yes, I agree that this might be better served in another thread. I made the same claims as both Jim and youknowyou several times on this thread and elsewhere. While it may sound easy, -10 to -15 dB isn't all that easy once you start picking apart that IR. Look a bit further than the standard representation you get from a measurement. That one usually only shows 8 kHz+ reflections etc...
Without bringing down the reflections by at least that amount you don't stand a chance of having a smooth FR response and well behaved phase plots.
That doesn't necessarily mean you get "bad sound", but it will be colored. As long as you realize that, do as you please. I know and realize my room still colors the sound to some (maybe even a large) degree. But at least it allows me to get quite a clean response:
Impulse at listening spot
1/12 smoothed FR plot, no gating...
Personally I'd still aim (and recommend) to get all those early reflections down by at least -10 to -15 dB, preferably more.
Do you want to know your room sound? Record a song at the listening position and play it back on headphones. You'll hear the room much more than sitting in that position. Play it back over the same speakers and you'll never forget that "room" sound!
If anyone wants more talk about the room, non specific to line arrays, lets open up a thread for that. After all I chose to use line arrays to have a shot at a well behaved room with minimal treatment. And guess what, it worked. (Synergy) Horns might have been an even better solution, but they would never get WAF approval. Believe me I tried.
Open Baffle simply takes up too much space for me to consider them. I need the speakers to be out of the way in the family room. I'm lucky to get away with approval for two huge towers anyway. I feel this has been discussed, maybe even more than once, on this thread already.
Can we get back to the inter aural comb problem now? I just spend 2.5 days straight trying to figure all of this out and nobody seems to be interested in any of it . Well it's still a big deal for me, so I'll carry on...
I'm not about to add a center speaker any time soon, I just might add subs and super tweeters at some point, just out of curiosity. So this inter aural cross talk has got my complete attention for the time being.
Without bringing down the reflections by at least that amount you don't stand a chance of having a smooth FR response and well behaved phase plots.
That doesn't necessarily mean you get "bad sound", but it will be colored. As long as you realize that, do as you please. I know and realize my room still colors the sound to some (maybe even a large) degree. But at least it allows me to get quite a clean response:
Impulse at listening spot
1/12 smoothed FR plot, no gating...
Personally I'd still aim (and recommend) to get all those early reflections down by at least -10 to -15 dB, preferably more.
Do you want to know your room sound? Record a song at the listening position and play it back on headphones. You'll hear the room much more than sitting in that position. Play it back over the same speakers and you'll never forget that "room" sound!
If anyone wants more talk about the room, non specific to line arrays, lets open up a thread for that. After all I chose to use line arrays to have a shot at a well behaved room with minimal treatment. And guess what, it worked. (Synergy) Horns might have been an even better solution, but they would never get WAF approval. Believe me I tried.
Open Baffle simply takes up too much space for me to consider them. I need the speakers to be out of the way in the family room. I'm lucky to get away with approval for two huge towers anyway. I feel this has been discussed, maybe even more than once, on this thread already.
Can we get back to the inter aural comb problem now?
I just spend 2.5 days straight trying to figure all of this out and nobody seems to be interested in any of it . Well it's still a big deal for me, so I'll carry on...I'm not about to add a center speaker any time soon, I just might add subs and super tweeters at some point, just out of curiosity. So this inter aural cross talk has got my complete attention for the time being.
That's EXACTLY what I'm talking about. The precise location of those dips and peaks depend on the setup of the speakers and head size. But it demonstrates the "problem" of stereo with 2 ears very well.
What I'm doing isn't canceling those dips, I'm merely postponing them to have a longer shot at hearing the phantom center signal.
The side signals, either panned hard left or hard right don't have the opposing FR to comb with at the ears and reacts slightly different. That's why, in a setup with a (very) low level of early reflections the phantom center can sound a bit more dark compared to sounds panned hard left or hard right. It's subtle but happening.
Earlier I noticed if I put some PEQ dips at ~3700 Hz and ~7400 Hz in my system some lyrics were easier to follow. A better intelligibility.
This inter aural cross talk delay I'm playing with has an even better effect than EQ can do by itself.
What I'm doing isn't canceling those dips, I'm merely postponing them to have a longer shot at hearing the phantom center signal.
The side signals, either panned hard left or hard right don't have the opposing FR to comb with at the ears and reacts slightly different. That's why, in a setup with a (very) low level of early reflections the phantom center can sound a bit more dark compared to sounds panned hard left or hard right. It's subtle but happening.
Earlier I noticed if I put some PEQ dips at ~3700 Hz and ~7400 Hz in my system some lyrics were easier to follow. A better intelligibility.
This inter aural cross talk delay I'm playing with has an even better effect than EQ can do by itself.
Seem like you need a center channel ear! Right over your eyes or perhaps on the top of your head!
I find the conversation interesting but in reality the fact we have two ears on either side of our heads is just something that our brain takes into account. How much this is a real issue is the real question, are we making something out of a little of concern problem? Even in the studio you have two channel separation of the speakers that are used to produce the music, so hard panned to the center is exactly how the music was heard when it was recorded. It seems you are trying to change what was originally accepted as two channel stereo signal. A center channel has nothing to do with normal two channel stereo, it is a different animal really. Multi-channel audio would solve this problem but that is just not the norm. And since we don't have perforated screens as in a movie theater we again end up with a different playback system even when we have center channels, they are not in the correct position really behind the screen. What do you get when you place two center channel speakers on each side of the flat screens above or below the screen, again you end up with the comb filtering you are spending so much energy to correct.
I find all of what Wesayso and Pano is doing interesting, at the same time I don't know how you can really ever achieve your end goals of no head shading of center images?
I find the conversation interesting but in reality the fact we have two ears on either side of our heads is just something that our brain takes into account. How much this is a real issue is the real question, are we making something out of a little of concern problem? Even in the studio you have two channel separation of the speakers that are used to produce the music, so hard panned to the center is exactly how the music was heard when it was recorded. It seems you are trying to change what was originally accepted as two channel stereo signal. A center channel has nothing to do with normal two channel stereo, it is a different animal really. Multi-channel audio would solve this problem but that is just not the norm. And since we don't have perforated screens as in a movie theater we again end up with a different playback system even when we have center channels, they are not in the correct position really behind the screen. What do you get when you place two center channel speakers on each side of the flat screens above or below the screen, again you end up with the comb filtering you are spending so much energy to correct.
I find all of what Wesayso and Pano is doing interesting, at the same time I don't know how you can really ever achieve your end goals of no head shading of center images?
I was quite happy with a bit of mid/side EQ to correct the tonal imbalance I perceived. At least UNTIL I tried Pano's rephrase-2 shuffler .
That gave me a clear perspective on the problem. What's not to like if we can manipulate the signal at the speaker to perceive more reality?
I won't move to a center speaker. At least not for stereo content. But there is a big difference in perception of near field monitor speakers, like the setup used in most studio's, and the ones where you don't sit as close. That's why there's something to gain here.
If I had not heard it in person I would probably say the same things you are. I'd deny it because I am pleased (very pleased) with my speakers performance as is. But now that I have heard the difference. It is way too interesting for me to avoid or deny it. Yet improvements can be heard when you decide to make the room behave better. This is a step in that same direction.
I like every little tweak that get's me closer to the recording. This one definitely counts as such in my book. Our brain is indeed powerful enough to compensate. As it's also pretty well equipped and schooled to deny added room sound at your listening position.
I can say, from personal experience, it just works! Try not to judge it until you've heard what it can bring. I know that's pretty hard to do, but hearing this would make you a believer of this technical problem, I'm sure of it. At least it turned me around.
I'm not convinced adding a third channel would solve this. We'd need material mixed to include that third channel to benefit from that. It creates new problems in my opinion. This "thing" I'm working on does not detract from regular stereo at any other position. It merely creates a longer view on the phantom center for your brain to process.
The papers from Professor Edgar Choueiri show a way to theoretically solve the entire inter aural comb filter problem. But that brings new problems off axis. So ends up doing something in between. What he does, exactly I don't know (not willing or able to spend over $50.000,00 to find out), but it's that same puzzle.
I've experimented with various RACE algorithms (from Ambiophonics.net and elsewhere) a long time ago and I couldn't stand the side effects. But this simple solution I use does not have those side effects. I'm trying to keep it as simple and least intrusive as possible while gaining our brain that extra time at the ears. Intelligibility improves, tonal balance improves, imaging improves... what's not to like if you don't suffer from draw backs compared to the usual stereo signal?
.That gave me a clear perspective on the problem. What's not to like if we can manipulate the signal at the speaker to perceive more reality?
I won't move to a center speaker. At least not for stereo content. But there is a big difference in perception of near field monitor speakers, like the setup used in most studio's, and the ones where you don't sit as close. That's why there's something to gain here.
If I had not heard it in person I would probably say the same things you are. I'd deny it because I am pleased (very pleased) with my speakers performance as is. But now that I have heard the difference. It is way too interesting for me to avoid or deny it. Yet improvements can be heard when you decide to make the room behave better. This is a step in that same direction.
I like every little tweak that get's me closer to the recording. This one definitely counts as such in my book. Our brain is indeed powerful enough to compensate. As it's also pretty well equipped and schooled to deny added room sound at your listening position.
I can say, from personal experience, it just works! Try not to judge it until you've heard what it can bring. I know that's pretty hard to do, but hearing this would make you a believer of this technical problem, I'm sure of it. At least it turned me around.
I'm not convinced adding a third channel would solve this. We'd need material mixed to include that third channel to benefit from that. It creates new problems in my opinion. This "thing" I'm working on does not detract from regular stereo at any other position. It merely creates a longer view on the phantom center for your brain to process.
The papers from Professor Edgar Choueiri show a way to theoretically solve the entire inter aural comb filter problem. But that brings new problems off axis. So ends up doing something in between. What he does, exactly I don't know (not willing or able to spend over $50.000,00 to find out), but it's that same puzzle.
I've experimented with various RACE algorithms (from Ambiophonics.net and elsewhere) a long time ago and I couldn't stand the side effects. But this simple solution I use does not have those side effects. I'm trying to keep it as simple and least intrusive as possible while gaining our brain that extra time at the ears. Intelligibility improves, tonal balance improves, imaging improves... what's not to like if you don't suffer from draw backs compared to the usual stereo signal?
Last edited:
Jim if you don't mind, I'm borrowing your pictures for the Phantom Center thread as it's one of the clearest examples of what's happening. Your room is an excellent play garden for this kind of stuff!
Here's the link: http://www.diyaudio.com/forums/multi-way/277519-fixing-stereo-phantom-center-61.html#post4745331
Here's the link: http://www.diyaudio.com/forums/multi-way/277519-fixing-stereo-phantom-center-61.html#post4745331
Thanks for taking the time to do those measurements. It becomes pretty clear that any resemblance to my theoretical plots isn't coincidence.
I like your measurement even better than the ones I've seen from Toole. They clearly show the truth of what's happening at the ears. And are a testament of how well adapted our brain is to make something out of that, somthing that still makes sense to us!
One more question, how accurate is that mic position at the left ear and the size of dummy head? I could probably "tune" something for your setup to pré convolve some test tracks for you to listen too. If anyone has the room for it to hear a difference, you have.
I like your measurement even better than the ones I've seen from Toole. They clearly show the truth of what's happening at the ears. And are a testament of how well adapted our brain is to make something out of that, somthing that still makes sense to us!
One more question, how accurate is that mic position at the left ear and the size of dummy head? I could probably "tune" something for your setup to pré convolve some test tracks for you to listen too. If anyone has the room for it to hear a difference, you have.
Last edited:
Not that accurate. I actually redid the measurement more accurately after my initial post, but didnt share it as to not clutter the thread. The general tendency was there in the first measurement.
I also am not sure if a compressed pink fluffy dummy head adds or subtracts from what is really going on.
Let me think about this a little more now that I am confident I understand the issue better.
On another note, changing the default save image width from 1607 to 3000 makes the graphs a lot clearer.
Attachments
Last edited:
Wesayso,
I respect what you are doing and accept that you are attempting to improve your system. I watch every post by you and Pano and Bob and others. I just don't know how well this transfers to a typical two channel speaker system that is not a line array like you have created. You have basically eliminated the floor bounce and ceiling reflections and this must have a significant effect on the center phantom image you would have from a typical stand mounted or floor standing speaker. I am not criticizing your attempts in the least, it is just a very extreme outlier you have created here with your speakers and room. I understand only some of what you do as I am not up on all the DSP processing yet, I will get here but it will take time to catch up. I understand in theory what you are doing but don't have the practical experience yet to process in my little brain! This is something I could comprehend having in my speaker system, I am just not sure yet how you would implement that and make it consumer friendly, this is not something I would think you would want to allow the end user to play with, this is not just giving them a simple tone control which almost always ends up with a smilely face curve, now we are playing with timing which I know the majority if not all consumers have no concept of. Auto room correction would be the real solution but I imagine you know that has limitations still, there are few who are working at this level, you indeed are a world apart from the average music listener, let alone the audiophiles.
I respect what you are doing and accept that you are attempting to improve your system. I watch every post by you and Pano and Bob and others. I just don't know how well this transfers to a typical two channel speaker system that is not a line array like you have created. You have basically eliminated the floor bounce and ceiling reflections and this must have a significant effect on the center phantom image you would have from a typical stand mounted or floor standing speaker. I am not criticizing your attempts in the least, it is just a very extreme outlier you have created here with your speakers and room. I understand only some of what you do as I am not up on all the DSP processing yet, I will get here but it will take time to catch up. I understand in theory what you are doing but don't have the practical experience yet to process in my little brain! This is something I could comprehend having in my speaker system, I am just not sure yet how you would implement that and make it consumer friendly, this is not something I would think you would want to allow the end user to play with, this is not just giving them a simple tone control which almost always ends up with a smilely face curve, now we are playing with timing which I know the majority if not all consumers have no concept of. Auto room correction would be the real solution but I imagine you know that has limitations still, there are few who are working at this level, you indeed are a world apart from the average music listener, let alone the audiophiles.
The trick with room treatment is to realize that typical foam or drape treatments work great in the higher frequencies but often not very well below about 400HZ.
One of the nice things about my OB dipoles is that the null at the sides means less reflection off side walls in my room, so imaging embedded in the program content is less confused. Sidewall reflections add a sense of spaciousness, but it's fake, and consistent with every recording, so could get boring after a while.
A waterfall graph of my OB speakers would probably look horrible, yet what you get in return is a more 3-D sound... so again it's about tradeoffs. I think they effectively recreate some of the spaciousness in the lower mid frequencies that got stepped on by inter-aural crosstalk.
One of the nice things about my OB dipoles is that the null at the sides means less reflection off side walls in my room, so imaging embedded in the program content is less confused. Sidewall reflections add a sense of spaciousness, but it's fake, and consistent with every recording, so could get boring after a while.
A waterfall graph of my OB speakers would probably look horrible, yet what you get in return is a more 3-D sound... so again it's about tradeoffs. I think they effectively recreate some of the spaciousness in the lower mid frequencies that got stepped on by inter-aural crosstalk.
Bob,
Room treatments are great at as you say some frequencies and panels for random dispersion can be very useful as long as you can incorporate those. What always bothers me is those who use the term bass traps, that is so incorrect as the actual absorption is not very low in reality, they should really be called mid-bass traps at best. No trap is going to dampen the low frequencies we are capable of producing from most speakers that can reach down to 20hz.
Room treatments are great at as you say some frequencies and panels for random dispersion can be very useful as long as you can incorporate those. What always bothers me is those who use the term bass traps, that is so incorrect as the actual absorption is not very low in reality, they should really be called mid-bass traps at best. No trap is going to dampen the low frequencies we are capable of producing from most speakers that can reach down to 20hz.
Thank you, I think...
Personally I like a more firm grip on the matter. That's why I try to measure and theorize my way to better sound.
My personal panels are 7 cm thick fiberglass insulation with 2 layers of wool felt spaced off the wall a bit. I find them very effective for what I wanted them to do. But here we are talking room again
... really, shouldn't we get another thread going for that kind of thing? It's not that it gets a lot of attention in average on this forum.To me, the most intriguing part of audio is the way our brain picks up the right queues. I still think time is one of the big parts that gets solved in our head.
Based on our primal survival instincts we are equipped with a brain that does know how to process the sound it hears very effectively. If we help our brain by putting up that puzzle more clearly we can perceive more of it.
Some day I'd like to see a full set of measurements of OB speakers. I'll admit I haven't looked for it but that energy from the back side of the cones has to go somewhere... I get the nulls at the sides. It is that energy firing at the front wall that doesn't simply dissolve... OB has many followers and if I would have had more room, who knows. Linkwitz's Orion speakers were actually the first thing I ever seriously considered, a long time ago. I studied what he did and why. But the room requirement didn't fit my needs. The study was interesting enough to keep an eye on what he did over the years.
But there are a lot more interesting people, even on this forum. I don't value the opinion of one specific guru. I'm convinced every one of them is right, one way or another
. In that puddle of madness I search for my own answers. One more try, Bob, did you see/read jim1961's room thread? For me it did two things:
1) awesome respect for what a single dedicated person can do to improve his listening environment! Not taking the DSP shortcuts like I am (are they really that short?
) but solving it acoustically and mechanically. It was a very interesting study for me and a primary example of what to look/aim for in my own DSP experiments.2) after seeing measurements taken in that room of Troels Gravesen's speaker design one can only admire what he (Troels) can come up with. I bet Troels would love a room like that to design speakers in. I know I would!
I did check out the Jim1961 room (gearsluts). Something like that seems ideal for mastering a recording. Playback might be nice too in a room like that. I think I prefer a little more liveness in a playback/room system, as long as it's done right. The worse case scenario is when you have only one reflection. It's cancellations may well be 10dB+. When you've got many random reflective paths, they all have different cancellation frequencies, so largely fill in each others cancellations. That may show a funky waterfall graph, but it can be much more pleasant to listen to. If you have a coffee table or a ceiling, you've got at least one significant reflective surface in an otherwise well damped room. Room reflections are not always our best friend, but we can work with them to some extent. Vertical line arrays get rid of the floor and ceiling. OB's get rid of the side walls to some extent, and use the front wall for enhancement. Rear woofers timed and phased properly can largely eliminate the rear wall (or so I've read). I found that damping corners of a room is one of the most effective room treatments. The hand clap ring test showed a huge improvement with that.
As far as room acoustics seeming off subject, I think it's a pretty important part of the process of getting good imaging in the middle from just two speakers. If you clean up the first 20mS too much, you might even make the inter-aural comb in the center image more audible...
As far as room acoustics seeming off subject, I think it's a pretty important part of the process of getting good imaging in the middle from just two speakers. If you clean up the first 20mS too much, you might even make the inter-aural comb in the center image more audible...
Maybe you haven't looked recently, but with the addition of my ambient channels, my room has many and varied timing late reflections, not just one
As far as cleaning up the first 20ms too much, I haven't noticed unpleasant side effects, especially when you have enough late energy (>25ms) to make up for the absence of the early kind (<20ms).
I will stop here as to not derail things too much. If you want to talk more, visit my page.
Not sure how to respond to this... it definitely describes a different path from what I have chosen. I do like reflections, late (scattered) ones. But removing the early reflections got me more and more 'into' the recording so far. No regrets there.
I still believe there's a solution that can 'mend' that tiny problem. As that's how I see it. Even though my experiments today weren't a complete success, I got lost in the music and had the hairs on my body standing up multiple times. It wasn't because it didn't sound right.
The main queue it wasn't quite right yet was I couldn't follow "some" vocal lines clear enough. Getting it right will automatically fix that problem as a PEQ bump of 2 dB at 1850 Hz and 5500 Hz changed it completely. So I'm still a bit off, but have a new experiment lined up, ready to go. On paper it's better than my previous one.
I'm not taking down damping panels anytime soon
.