I like playing with it, no doubt, so it's not a big task.
Also there's a big chance I get to learn something from it, and it does give me something to focus on while waiting for the weather to turn so I can finish my #### subs!
There must be valid reasons why people like certain configurations, as far as speakers go. Many people like Open Baffle. Horns have a large following too!
Electrostats, direct radiators, full range, 2 way, 3 way, 4 way etc...
What I'm trying to do is trying to figure out what part of the perceived wave front (from each of em) triggers that preference, without trying (or having to try) all the different speaker types myself. I've listened to a variety of them trough my life, but don't plan to build a sample set of each of em.
I'm still convinced a clean and correct early wave front followed by a properly timed and shaped "rerun", in the form of decorrelated ambience, can give us a very pleasing overall result with both good imaging and tonal joy.
That concept has brought me this far, now let's see if we can make it even better!
Also there's a big chance I get to learn something from it, and it does give me something to focus on while waiting for the weather to turn so I can finish my #### subs!
There must be valid reasons why people like certain configurations, as far as speakers go. Many people like Open Baffle. Horns have a large following too!
Electrostats, direct radiators, full range, 2 way, 3 way, 4 way etc...
What I'm trying to do is trying to figure out what part of the perceived wave front (from each of em) triggers that preference, without trying (or having to try) all the different speaker types myself
. I've listened to a variety of them trough my life, but don't plan to build a sample set of each of em.I'm still convinced a clean and correct early wave front followed by a properly timed and shaped "rerun", in the form of decorrelated ambience, can give us a very pleasing overall result with both good imaging and tonal joy.
That concept has brought me this far, now let's see if we can make it even better!
I could not restrain myself any longer and used this reverb mixing trick with my Stereo setup. Basically I added an extra 2 channels derived from the mid signal, run that trough a reverb algorithm at 100% wet and add it to left and right after the Stereo reconstruction.
So the reverb gets the mono sum (that's a little different from using pure mono content only, because it's consisting of L + R) We don't have the pure center material available anymore.
First attempt: way too much reverb added. Also sounded way too dark, even though the reverb was acting as a high-pass. Some quick tweaks: move up the high pass frequency and adjust the level way down. Much more promising. First note: tonal balance gets evened out, however I'm still losing perceived details.
Some tweaks later and I'm getting quite a good tonal balance across the stage but details still suffer somewhat. The space sounds big (eyes closed) but I'm losing some definition. Mind you, it also depends on the material being played. I think it has potential. Everything smooth's out, just a bit too much for my taste. However I'm enjoying a brighter sounding center, much more inline with the sides. It gives more of an intimate sound on close vocals, just like a good tweeter can do. Nuances in the voice.
All in all much more work is needed. In this process I opted for the same reverb algorithm for phantom center as well as ambient channels. The Ambience (Room) was used in a large room setting (Live Studio).
The Random Hall I was using on my ambient channels can spin it's tail and be obvious at times on the ambient channels. So I figured using the same algorithm all over instead. Much more things to try of course, but not a bad start. I seem to be able to tweak to taste, there's very little level tweaks needed to get quite huge perceptual changes.
More to follow, I'm sure...
So the reverb gets the mono sum (that's a little different from using pure mono content only, because it's consisting of L + R) We don't have the pure center material available anymore.
First attempt: way too much reverb added. Also sounded way too dark, even though the reverb was acting as a high-pass. Some quick tweaks: move up the high pass frequency and adjust the level way down. Much more promising. First note: tonal balance gets evened out, however I'm still losing perceived details.
Some tweaks later and I'm getting quite a good tonal balance across the stage but details still suffer somewhat. The space sounds big (eyes closed) but I'm losing some definition. Mind you, it also depends on the material being played. I think it has potential. Everything smooth's out, just a bit too much for my taste. However I'm enjoying a brighter sounding center, much more inline with the sides. It gives more of an intimate sound on close vocals, just like a good tweeter can do. Nuances in the voice.
All in all much more work is needed. In this process I opted for the same reverb algorithm for phantom center as well as ambient channels. The Ambience (Room) was used in a large room setting (Live Studio).
The Random Hall I was using on my ambient channels can spin it's tail and be obvious at times on the ambient channels. So I figured using the same algorithm all over instead. Much more things to try of course, but not a bad start. I seem to be able to tweak to taste, there's very little level tweaks needed to get quite huge perceptual changes.
More to follow, I'm sure...
I can pretty much shape it anyway I like it. But how it should be is a big question mark. That doesn't matter too much though, it's meant for enjoyment, not production .
Even if the high passed reverb is used on the phantom center, there is enough L and R info in there to make a large perceptual difference.
Use a simple way to play with. It's quite interesting how "we" react to these tweaks.
It's a lot of fun to play with. To start I like to exaggerate any setting to get a quick feel of the direction it takes me. But even subtle tweaks matter once you get closer to something you like.
Meanwhile I found that the "Ambience" preset has it's "Spin" factor at zero to begin with. That's probably why the "Random Hall" can have that obvious spin in it's tail sound. At default it was at 2.5 I believe. So I'll play with that too, it's quite fun if it's subtle, one more thing to dial in.
.Even if the high passed reverb is used on the phantom center, there is enough L and R info in there to make a large perceptual difference.
Use a simple way to play with. It's quite interesting how "we" react to these tweaks.
It's a lot of fun to play with. To start I like to exaggerate any setting to get a quick feel of the direction it takes me. But even subtle tweaks matter once you get closer to something you like.
Meanwhile I found that the "Ambience" preset has it's "Spin" factor at zero to begin with. That's probably why the "Random Hall" can have that obvious spin in it's tail sound. At default it was at 2.5 I believe. So I'll play with that too, it's quite fun if it's subtle, one more thing to dial in.
Last edited:
Another free hour or two of listening, at 17 ms .
Better! Good enough to end with an hour of Led Zeppelin (that's saying a lot in my case).
It's still a test setup, I'll tweak it for a few weeks more. It's fun though, the "Spin" on the ambience does help. A real atmosphere tweak, if used with some restraint.
Time's up though. Why stop tweaking if it is this much fun!
.Better! Good enough to end with an hour of Led Zeppelin (that's saying a lot in my case).
It's still a test setup, I'll tweak it for a few weeks more. It's fun though, the "Spin" on the ambience does help. A real atmosphere tweak, if used with some restraint.
Time's up though. Why stop tweaking if it is this much fun!
Keep on going with that processing; maybe I will catch up some day. Odds increase dramatically if you show us how
(I know this is slightly off topic but the thread is quiet)
Today I am still in analysis phase. I entered my trial CBT design into Vituix in such detail that it was able to show me vertical directivity based on driver piston size. I tried the frequency dependent weighting that was mentioned earlier in the thread. I simply put a PEQ that boosted centered on 80 Hz by the negative of the group's weight. It worked until I looked at the directivity and saw that it had resulted in vertical lobes at 200 Hz - about where the boost from the PEQs tailed off. So that is not such a good idea; probably stereo subs &/or more robust driver.
Thinking of using the top weighting group of drivers in the CBT for late ceiling splash content as well as the main content. Those drivers point up and the primary content is weighted down 12 db or so...so they have the headroom for it.
(I know this is slightly off topic but the thread is quiet)
Today I am still in analysis phase. I entered my trial CBT design into Vituix in such detail that it was able to show me vertical directivity based on driver piston size. I tried the frequency dependent weighting that was mentioned earlier in the thread. I simply put a PEQ that boosted centered on 80 Hz by the negative of the group's weight. It worked until I looked at the directivity and saw that it had resulted in vertical lobes at 200 Hz - about where the boost from the PEQs tailed off. So that is not such a good idea; probably stereo subs &/or more robust driver.
Thinking of using the top weighting group of drivers in the CBT for late ceiling splash content as well as the main content. Those drivers point up and the primary content is weighted down 12 db or so...so they have the headroom for it.
Forgive me Ronald if this is all old hat to you, but I thought you might find it interesting StereoTimes -- Ear Pinna and Realism in Music Reproduction
Keep on going with that processing; maybe I will catch up some day. Odds increase dramatically if you show us how
(I know this is slightly off topic but the thread is quiet)
Today I am still in analysis phase. I entered my trial CBT design into Vituix in such detail that it was able to show me vertical directivity based on driver piston size. I tried the frequency dependent weighting that was mentioned earlier in the thread. I simply put a PEQ that boosted centered on 80 Hz by the negative of the group's weight. It worked until I looked at the directivity and saw that it had resulted in vertical lobes at 200 Hz - about where the boost from the PEQs tailed off. So that is not such a good idea; probably stereo subs &/or more robust driver.
Thinking of using the top weighting group of drivers in the CBT for late ceiling splash content as well as the main content. Those drivers point up and the primary content is weighted down 12 db or so...so they have the headroom for it.
Too bad that the frequency dependent shading does not seem to work out. One of the mayor advantages of why I chose the array topology. Nearly unlimited dynamics while not stressing the individual drivers. Aside from the needed floor space that was one of the reasons for me to go with a straight array.
I think using the top drivers to double act as late ceiling splash would be a fun option to play with!
Thanks for posting it.
Yes, I've certainly read more work/text from Ralph Glasgal and I agree with many points being made in that article.
If you read the part about Arnold Klayman you'll notice the similarity to what I do with my ambient channels.
A trick I picked up from werewolf from his 'teachings' over on the DIYMobileaudio forums. The same werewolf that gave us an excellent lecture about infinite arrays here: Infinite Line Source: analysis
My mid/side EQ trickery is loosely based on my experiments with Ambiophonics. I've tried a couple of those race algorithms and cooked up my own FIR variant. While I wasn't successful in finding something I wanted to keep, I'm sure I'll revisit it sometime soon. My biggest problem is that I can't follow the advise of moving the speakers closer together. Even though it does make a lot of sense.
I do think there's enough logics in that article, as to how we, as humans, perceive audio. I do believe it's the specific combing patterns (or what it does to the perceived FR at the ear) and HRTF functions that allow us to make sense of it all. At least for the part from ~1 KHz and up. One of the videos on the Beolab 90 got into that same theory, a video done by the head engineer, The "Tonmeister" Geof Martin. I think this is the one: Inside the process of BeoLab 90 - Bang & Olufsen Merchant City Glasgow.
What I've perceived in my (many) experiments alludes to a wonderful way of having audio in your room. If only I can find a way to combine all of the things I've heard so far! I do believe I'm getting closer. My first post has a couple of reviews linked. The version DIYaudio member Boden got to hear was with a cross talk compensation algorithm active (my own brew with FIR files). In all honesty, in the long run I got head aches from longer listening sessions. But that might have had nothing to do with the cross talk compensation I was running but with my FIR EQ scheme at that time.
My current experiments are head ache free
. But can't get it to do everything yet that I've heard before. I'm getting great tonality right now from left/center/right but lack a bit of depth (to my taste) in the center. With depth I mean from the perceived front of the stage out to the back or farthest instrument. It's there (as in layered sound), but I have had better (with a slightly worse tonal balance).
If I can find it again, I'm sure it will be very satisfying to listen to.
The beauty of the current settings is that in the next room, the connected kitchen space, you still get the idea there's a 'live' concert going on in the living room. Pleasing sound and fun dynamics.
Last edited:
You've stirred up something here
. I've always wanted to try using cross talk cancellation on phantom sounds only, without affecting the sides. It sounds (no pun intended) easier than it is. One can simply split the signal, like in my mid/side EQ. But in reality, there will always be L and R info in the sum part, that needs to be compensated for in the sides.It will be worth pursuing, but it needs a complete restructure to fit it all within the JRiver/Metaplugin setup and I'm running out of processing channels fast!
(reserving 2 channels for sub duty, I have only 6 channels to play with)I have some ideas how to go with it, but I need to validate if it is possible.
For a test I just applied a simple cross talk signal to the (L+R) phantom part... Hmm, interesting... but the stereo mix is entirely different (not in a good way). The phantom part does sound quite good though, more definition/detail. No way of knowing that this detail is right though. More often than not "different" can be fooling you to think it's better. At least for a short while.
So will it be worth several hours of "reprogramming" to try? I'd never know without trying, would I?
If I hadn't run cross talk compensation before I'd probably not consider this. However there were some things that really startled me in the experiments some time ago. So it deserves another shot. Here's to hoping it will be worth it. First some (real) work to finish up on and then on to the task at hand.
Thanks @scottjoplin
You're welcome Agreed, different is often perceived as better and it's whether you're more comfortable with it over time I find is the better test of any change, of course you adapt, so perhaps continual "upgrades" are easily achieved by switching back and forth every now and again
That article is from a link here Online LEDR™ Sound Test | Listening Environment Diagnostic Recording Test which I've been encouraging people to have a go at on this thread Question regarding phase differences of amps.
Audiocheck is quite a useful site with some tests and links
Agreed, different is often perceived as better and it's whether you're more comfortable with it over time I find is the better test of any change, of course you adapt, so perhaps continual "upgrades" are easily achieved by switching back and forth every now and again That article is from a link here Online LEDR™ Sound Test | Listening Environment Diagnostic Recording Test which I've been encouraging people to have a go at on this thread Question regarding phase differences of amps.
Audiocheck is quite a useful site with some tests and links
Right now I just send a reduced SPL level and inverted signal of frequencies, high passed at ~1 KHz back into the front mains. The base of that signal is derived from the (L+R) part of the total signal. (taken after first splitting up into mid and side). A very crude way of achieving some cross talk cancellation in a standard Stereo triangle.
In previous attempts I've played with finding the right delay for my specific listening spot (and head size). Just by listening to a lot of delay variations by first making a crude sketch of my setup and measuring the difference in path lengths between left speaker and left ear and left speaker and right ear etc. I'll admit, I even used a string or thin rope from the actual speaker to my head/ears to confirm.
My magic number is ~0.272 ms of delay.
It totally messes up the sides as is. The challenge is to keep the signal of the sides unaltered, as I'm very pleased with their sense of reality.
If I can get that all to work together at once I can refine it.
Other tricks I have tried are to use the back channels for compensation, but that hasn't been successful.
I know the delay is working, as I wiggle my head in the sweet spot with this engaged I snap in and out of the sound enhancement.
A little goes a long way here. Just setup in a crude way first, to get an idea of what it does. I've done it with FIR filters as well as just mixing with what JRiver has to offer (less exact delay timing with that).
I haven't started my reworking of the processing chain just yet, I'm still enjoying going back trough the old Western Electric 1928 - How far have we come in the last 100 years? thread. I've read it all before and reading these threads often spike new and interesting experiments. I especially like the thoughtful posts by Tom Danley among a few others. You were quite active in that thread, Pano. I can relate to how you "feel" about the difference between great sound as being a metric (getting as close to the input signal as possible and proving it with measurements) as opposed to great sound by hearing it sound so real and convincing.
While most posters with a science background on that thread were promoting the march of progress, I'm more interested in what it was that makes/made the sound of these old horns so enchanting.
Our two ears and that brain in the middle makes this hobby overly complicated. Yet most readers seem to fly right over or past that fact.
I bet a single speaker has a way better chance of sounding spooky real. Yet if you find a combination that works as well in Stereo and fools you it really is something else. I've had or better yet heard flashes of greatness. So far I've not been completely successful of combining it all in one go.
When you do get close enough, it's like being in a time capsule. That's why I won't stop hunting for it.
I guess you've got to have a few things right for it to happen. Experiments with some parts right can work on a limited number of material though. But I am convinced that if we can get most of it close enough to right, it will last.
In previous attempts I've played with finding the right delay for my specific listening spot (and head size). Just by listening to a lot of delay variations by first making a crude sketch of my setup and measuring the difference in path lengths between left speaker and left ear and left speaker and right ear etc. I'll admit, I even used a string or thin rope from the actual speaker to my head/ears to confirm.
My magic number is ~0.272 ms of delay.
It totally messes up the sides as is. The challenge is to keep the signal of the sides unaltered, as I'm very pleased with their sense of reality.
If I can get that all to work together at once I can refine it.
Other tricks I have tried are to use the back channels for compensation, but that hasn't been successful.
I know the delay is working, as I wiggle my head in the sweet spot with this engaged I snap in and out of the sound enhancement.
A little goes a long way here. Just setup in a crude way first, to get an idea of what it does. I've done it with FIR filters as well as just mixing with what JRiver has to offer (less exact delay timing with that).
I haven't started my reworking of the processing chain just yet, I'm still enjoying going back trough the old Western Electric 1928 - How far have we come in the last 100 years? thread. I've read it all before and reading these threads often spike new and interesting experiments.
I especially like the thoughtful posts by Tom Danley among a few others. You were quite active in that thread, Pano. I can relate to how you "feel" about the difference between great sound as being a metric (getting as close to the input signal as possible and proving it with measurements) as opposed to great sound by hearing it sound so real and convincing. While most posters with a science background on that thread were promoting the march of progress, I'm more interested in what it was that makes/made the sound of these old horns so enchanting.
Our two ears and that brain in the middle makes this hobby overly complicated. Yet most readers seem to fly right over or past that fact.
I bet a single speaker has a way better chance of sounding spooky real. Yet if you find a combination that works as well in Stereo and fools you it really is something else. I've had or better yet heard flashes of greatness. So far I've not been completely successful of combining it all in one go
.When you do get close enough, it's like being in a time capsule. That's why I won't stop hunting for it.
I guess you've got to have a few things right for it to happen. Experiments with some parts right can work on a limited number of material though. But I am convinced that if we can get most of it close enough to right, it will last.
Last edited:
I reread some of that too. I saw that I was a bit rough on a few people in that thread. But it's frustrating and a bit insulting to be told what can't be done, when you've clearly heard. Not just heard it - but sat thru two and a half days of hearing it in the presence of a 100+ other people. And that experience becomes a life changing one, becomes an experience that you chase for a lifetime. To have it dismissed out of hand or be told its not possible just doesn't sit very well.I haven't started my reworking of the processing chain just yet, I'm still enjoying going back trough the old Western Electric 1928 - How far have we come in the last 100 years? thread. I've read it all before and reading these threads often spike new and interesting experiments.
Like you, I use science in the pursuit of a powerful illusion. As you say, it can work like a time capsule. For me the importance is the strength of the illusion, the means used to get there are secondary. Interesting, useful - but secondary.
I do have a genuine interest in how this all works, I guess I've been curious about the inner workings of the world around us since birth . But my goal is the same, as that too can bring huge joy. This hobby is the one thing that brings most of my interests together.
I love music as an emotional experience and basically can't live without it. That's the experience I am aiming for.
'Other people' have already captured that part (the real magic) and locked it into recordings. Our job is to get it out again!
I'm convinced the replay part is all science based, not magic. However I am looking at both camps to find answers, I listen to objective or subjective opinions and experiences alike to try and figure out the why of it all.
I'm pretty sure my little project is a nightmare for purists. Though, with all my digital means I do still hunt for that analog sound experience, as strange as that may be. I'm willing to deviate from the 'norm' if that's what it takes to make my soundscape more real and captivating.
. But my goal is the same, as that too can bring huge joy. This hobby is the one thing that brings most of my interests together.I love music as an emotional experience and basically can't live without it. That's the experience I am aiming for.
'Other people' have already captured that part (the real magic) and locked it into recordings. Our job is to get it out again!
I'm convinced the replay part is all science based, not magic. However I am looking at both camps to find answers, I listen to objective or subjective opinions and experiences alike to try and figure out the why of it all.
I'm pretty sure my little project is a nightmare for purists. Though, with all my digital means I do still hunt for that analog sound experience, as strange as that may be. I'm willing to deviate from the 'norm' if that's what it takes to make my soundscape more real and captivating.
Last edited:
