Hi DBMandrake, all
I should have prefaced the recordings by saying I am not in the recording business and don’t plan to record any more than informally. This is a project I have been working on for a good while now but is not on the high priority list, more like when there is time to work on it I can. My goal if it works well enough is to eventually offer this as a product at work but right now it is work in progress. What I have done so far is demonstrated that with the rest of the channels, one can make a very convincing capture / reproduction of a live event (so far limited to exciting stuff like s neighbor’s Harley, trains, birds and bugs, fireworks and a friends son’s Irish folk band playing at his birthday party).
The idea or approach is based on a realization I had while developing the Unity / Synergy horns about how we hear totally differently than a microphone measures. AS here, the idea being that if one can create an apparent phantom image strongly and anywhere between the two sources, that “all you need to do” at the receiving end is capture it with that extra vector information and make the system as many channels wide as desired.
Mandrake your right, there is no head transform involved, that part happens when the speakers in front of you create the sound and it arrives from a given height. With headphones, your ears tell you what they think the direction is based on a life time of hearing. We don’t hear reflections and comb filtering etc like they appear on a measurement but rather these are these sounds as clues that we use to hear direction and height and distance etc.. Without knowing it, our hearing systems takes two inputs, interprets what we detect against past experience and determines the likely direction and such.
I had suggested that these be played through headphones at least first as they do provide a high signal to noise (noise including reflections etc) and complete isolation from one ear to the other. In the latest stage, these goes to the right and left channels with a rear and two side/rear channels to complete the circle.
I have some of the parts for a six channel version which will include “up” information as a lifelong dream has been to be able to record something and then on playback, have it sound like your really there. To me, the biggest obstacle is capturing that complex signal intact hence this slowly moving invention.
I do have a couple other recordings I could ask they put up but I know they are building a new web site and am not sure how keen they will be about adding to our old one.
Try the fireworks if you haven’t but be careful there is no compression.
Best,
Tom
I should have prefaced the recordings by saying I am not in the recording business and don’t plan to record any more than informally. This is a project I have been working on for a good while now but is not on the high priority list, more like when there is time to work on it I can. My goal if it works well enough is to eventually offer this as a product at work but right now it is work in progress. What I have done so far is demonstrated that with the rest of the channels, one can make a very convincing capture / reproduction of a live event (so far limited to exciting stuff like s neighbor’s Harley, trains, birds and bugs, fireworks and a friends son’s Irish folk band playing at his birthday party).
The idea or approach is based on a realization I had while developing the Unity / Synergy horns about how we hear totally differently than a microphone measures. AS here, the idea being that if one can create an apparent phantom image strongly and anywhere between the two sources, that “all you need to do” at the receiving end is capture it with that extra vector information and make the system as many channels wide as desired.
Mandrake your right, there is no head transform involved, that part happens when the speakers in front of you create the sound and it arrives from a given height. With headphones, your ears tell you what they think the direction is based on a life time of hearing. We don’t hear reflections and comb filtering etc like they appear on a measurement but rather these are these sounds as clues that we use to hear direction and height and distance etc.. Without knowing it, our hearing systems takes two inputs, interprets what we detect against past experience and determines the likely direction and such.
I had suggested that these be played through headphones at least first as they do provide a high signal to noise (noise including reflections etc) and complete isolation from one ear to the other. In the latest stage, these goes to the right and left channels with a rear and two side/rear channels to complete the circle.
I have some of the parts for a six channel version which will include “up” information as a lifelong dream has been to be able to record something and then on playback, have it sound like your really there. To me, the biggest obstacle is capturing that complex signal intact hence this slowly moving invention.
I do have a couple other recordings I could ask they put up but I know they are building a new web site and am not sure how keen they will be about adding to our old one.
Try the fireworks if you haven’t but be careful there is no compression.
Best,
Tom
Our hearing groups reflections in time but an ETC doesn't
Psychoacoustic Bark wavelet groups reflections in time but an ETC doesn't
- Elias
Thanks for the translation !
According to google translator Blauert's directional bands Blauertsche Bänder = Czech blue ribbons
My goodness !
Time for conclusions !
There are two frequency bands important for stereo imaging to generate the image in front: 300-600 Hz and 2.5-6 kHz. The lower freq band is very challenging in a small room acoustic space, or a living room. A very directional speaker in the freq range of 300-600 Hz is required. A monopole box will not do ! Dipole line array will do better !
The freq range of 2.5-6 kHz is very interesting, and I believe this is the freq range my experimental stereolithic system was relying !
- Elias
According to google translator Blauert's directional bands Blauertsche Bänder = Czech blue ribbons
My goodness !
Time for conclusions !
There are two frequency bands important for stereo imaging to generate the image in front: 300-600 Hz and 2.5-6 kHz. The lower freq band is very challenging in a small room acoustic space, or a living room. A very directional speaker in the freq range of 300-600 Hz is required. A monopole box will not do ! Dipole line array will do better !
The freq range of 2.5-6 kHz is very interesting, and I believe this is the freq range my experimental stereolithic system was relying !
- Elias
Did I already mention the ignore function?
Out of ignorance the guy will probably still reply to your posts but at least you don't have to look at it anymore.Psychoacoustic Bark wavelet groups reflections in time but an ETC doesn't
- Elias
How relevant is the Bark scale?
Thanks for the translation !
According to google translator Blauert's directional bands Blauertsche Bänder = Czech blue ribbons
My goodness !
Time for conclusions !
There are two frequency bands important for stereo imaging to generate the image in front: 300-600 Hz and 2.5-6 kHz. The lower freq band is very challenging in a small room acoustic space, or a living room. A very directional speaker in the freq range of 300-600 Hz is required. A monopole box will not do ! Dipole line array will do better !
The freq range of 2.5-6 kHz is very interesting, and I believe this is the freq range my experimental stereolithic system was relying !
- Elias
Care to elaborate how you conclude that a dipole line array is desirable when looking at Blauert's research on critical bands?
I just want to thank everybody for helping me get a much more detailed understanding of the imaging game. Designing really good speakers and getting them to sound very good in a typical living room is quite the challenge. Good is usually easy, great and right is much harder (much like playing the guitar). Working with the idiosyncracies of the recording process is another substantial challenge. My experience with pressure zone mics and "binaural" head mics has heightened my awareness of this immensely. I love a project where I know I can do real good, but am unlikely to conquer it entirely, and will learn much along the way.
I believe more insight for room reflections can be gained by the use of modulation analysis.
The key points to follow this line of thought are:
- The information in music is a set of modulations
- A room reflection can be seen as an AM modulation
- A room reflection could be perceivable if it changes the signal in the modulation domain
Here's my take on perception of room reflections:
Elias Pekonen Home Page - Comb Filtering Perception
Why ignore ?
- Elias
The key points to follow this line of thought are:
- The information in music is a set of modulations
- A room reflection can be seen as an AM modulation
- A room reflection could be perceivable if it changes the signal in the modulation domain
Here's my take on perception of room reflections:
Elias Pekonen Home Page - Comb Filtering Perception
Why ignore ?
- Elias
Why ignore ?
In order to get to some kind of general consensus. I don't mean to ignore them permanently
.Member
Joined 2009
Thank you everyone for the software help. I admit I feel a bit stupid running around my room with pillows and comparing Impulse Responses but I feel they presented me with enough information for the simple scope of my test.
We do need a way to keep things standardized in this thread and to help compare results I will try to convert to ETC plots for the future. I'm gave up on trying to figure out the free version of ARTA and I will be looking into REW.
As for wavelets and more complex analysis I don't feel I have enough experience to interpret them beyond the obvious. The topic of this thread suggests that we are not sure what the ideal situation is and looking differently at the data is a great way to discover helpful patterns. If we define better what we are looking for and how can we see it better wavelets will be extremely useful. I like the wavelet study of dipole vs. monopole bass on Elias' page. The modulated bursts very clearly show how monopole bass is dirtier but I doubt that conclusion will be easy to draw from the impulse response graph. My point is that wavelets are probably only marginally helpful for looking at the Impulse Response (to me at least) and much more valuable for evaluation of more specific data.
---------
As for progress with my setup, I am experimenting with the speakers toed out, firing towards the close sidewall. I had to move them further from the walls and closer to the center of the room. I feel the sound has improved but I'd like to listen for at least a week before deciding where they will stay. I need to make some adjustments for the new position to compensate for room disturbances in the bass. It was pointed out earlier, that the directivity of the tweeter narrows in the last octave off-axis and a drop in the Frequency Response registers at the sweet spot. I'm not sure if compensation is necessary because then the on-axis signal that fires into the wall will be uneven. I will have to evaluate this further.
We do need a way to keep things standardized in this thread and to help compare results I will try to convert to ETC plots for the future. I'm gave up on trying to figure out the free version of ARTA and I will be looking into REW.
As for wavelets and more complex analysis I don't feel I have enough experience to interpret them beyond the obvious. The topic of this thread suggests that we are not sure what the ideal situation is and looking differently at the data is a great way to discover helpful patterns. If we define better what we are looking for and how can we see it better wavelets will be extremely useful. I like the wavelet study of dipole vs. monopole bass on Elias' page. The modulated bursts very clearly show how monopole bass is dirtier but I doubt that conclusion will be easy to draw from the impulse response graph. My point is that wavelets are probably only marginally helpful for looking at the Impulse Response (to me at least) and much more valuable for evaluation of more specific data.
---------
As for progress with my setup, I am experimenting with the speakers toed out, firing towards the close sidewall. I had to move them further from the walls and closer to the center of the room. I feel the sound has improved but I'd like to listen for at least a week before deciding where they will stay. I need to make some adjustments for the new position to compensate for room disturbances in the bass. It was pointed out earlier, that the directivity of the tweeter narrows in the last octave off-axis and a drop in the Frequency Response registers at the sweet spot. I'm not sure if compensation is necessary because then the on-axis signal that fires into the wall will be uneven. I will have to evaluate this further.
You should cheer up a bit ! Afterall, Zwicker was a German, right
So what? Doesn't answer how relevant bark scale is.
Thanks Elias,
but now that I have (somewhat) learned what the pretty pictures are trying to explain, I need 5 ms (and up) reflections at a decent and realistic level - not 0 dB. Otherwise it lacks relevance to the real world of hearing IMHO.
Rudolf
Your welcome!
Yes, a binaural "check" is a *MUCH* better method to see if your designs and placements are getting better or not. VERY few use this method in my experience - but it's just a lot easier to remind you what compelling reproduction should be like with the "check" (spatially and tonally). SL is one of the few more notable designers that uses this to check his own designs.
The "toe-out" and more "near-field" method of placement has a serious limitation in the midrange with conventional loudspeakers. Most "hit" the bottom of their baffle-step loss somewhere between 600-300 Hz. At that point they are radial (or horizontally "omni"). Their non-directional behavior at these lower freq.s does little for maintaining imaging character and tonal behavior. (..and it gets particularly bad with loudspeakers employing significant baffle step correction.) So even if you "eq" flat at your position, the result is often still "lacking". A dipole or cardioid in this range substantially improves this.
Freq. response deviations will always pose difficulties. Yeah, for that REW (or some other equalization method).
Just make sure you are not *too* far off axis for the direct sound. What angle off axis will you be listening? What sort of speakers are you using?
To resurrect an old post, I just stumbled across a new Linkwitz page that has recently been added discussing directivity and constant directivity in particular, which is interesting as Linkwitz is a well known dipole proponent. He provides links to a lot of the designs and approaches that have been discussed in this thread, (as well as others) which people might find interesting:
Constant directivity loudspeaker designs
One thing that caught my eye though is the section "Horbach-Keele linear-phase digital crossover filters", which shows a design approach that looks very similar to your Snell XA Reference tower design:
Snell Acoustics XA Reference Tower loudspeaker | Stereophile.com
From the description in the stereophile article this more recent (?) Keele design seems to be using a very similar approach of using spaced pairs of drivers matched with their crossover frequencies to create a uniform lobe, albeit using DSP based linear phase filters. Any comments ?
Good thinking ! However, the effect of the reflection amplitude bares only a slight effect on the modulation domain, for example:
5 ms 0 dB reflection:
An externally hosted image should be here but it was not working when we last tested it.
5 ms -6 dB reflection:
An externally hosted image should be here but it was not working when we last tested it.
Only the notch at the modulation freq of 100Hz is affected noticeable.
See more here:
Elias Pekonen Home Page - Comb Filtering Perception
- Elias
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.