SDA: Dimensional driver's band width?

sheathensemble

Hi guys, I am just curious as to what band width the dimensional driver's cover. I know that different models of sda's use different bandwidths for the dimensional drivers. I am wondering what the ideal bandwidth is...

Thanks in advance!!!

heiney9

For your 1C's read this review, it explains it all.

http://www.polksda.com/sda1creview.shtml

H9

sheathensemble

Thanks, I knew I had seen it around somewhere before! So do you think that 200-1000 hz is the ideal dimensional bandwidth? 1c's are old and I am just curious as to if the bandwidth used would be subject to change... Has anyone heard anything different about what the bandwidth should be?

I am going to try to make "3d" recordings, and part of implementing my plan is to include dimensional processing at the mixing stage...

sheathensemble

also, I wouldn't mind knowing how many db the dimensional driver's are attenuated.

Schurkey

sheathensemble wrote: »

I am going to try to make "3d" recordings, and part of implementing my plan is to include dimensional processing at the mixing stage...

How will a pre-processed recording play on SDA speakers?

Kind of like other forms of pre-processing--compression, out-of-phase issues, improper panning, etc... Once the processing is in the recording, it'll be tough to get it back OUT if needed.

sheathensemble

Schurkey wrote: »

How will a pre-processed recording play on SDA speakers?

Kind of like other forms of pre-processing--compression, out-of-phase issues, improper panning, etc... Once the processing is in the recording, it'll be tough to get it back OUT if needed.

I am trying to use the SDA approach to removing cross talk in the mixing stage with out the use of an array or any analog processing. If I am successful I would imagine playing the recordings on sda's would be similar to using a carver dimensional processor with your sda's...
So I really don't think these recordings will be for use will sda's. But rather for getting a similar effect on an ordinary stereo setup.

I figure the dimensional driver's in sda's have three aspect's that would need to be replicated... the polarity (reversed), the bandwidth (very narrow, from what I understand 200-1000 hz) and the volume (attenuated below the stereo driver's.) These are all very easy to replicate in a standard audio program.

My main goal is to use this processing along with 3 dimensional recording techniques to achieve a true 3 dimensional recording. But the processing on its own would also have other uses, for example I could take any audio file and make a stereo dimensional mix.

I'm really just experimenting at this point, and could use any input about bandwidth and attenuation. I feel like I am talking out of my a$$ right now... but if I have any success i will post it.

So if anyone has any info in regards to SDA's or dimensional processing units please chime in.

Cheers

DarqueKnight

sheathensemble wrote: »

also, I wouldn't mind knowing how many db the dimensional driver's are attenuated.

The SDA signal level is 1/2 of the stereo signal level.

sheathensemble wrote: »

So I really don't think these recordings will be for use will sda's. But rather for getting a similar effect on an ordinary stereo setup.

You will still have interaural crosstalk due to one ear hearing sound from the opposite side speaker. This cannot be dealt with in a recording because it is an artifact of the listener's hearing caused by each ear hearing two speakers.

The only way to get rid of interaural crosstalk without some type of active or passive cancellation circuitry is to use headphones.

TennMan

That's interesting DK. Based on the amount of attenuation would it be possible to adjust the SDA effect. Have you ever tried tinkering with it?

Face

Here's an binaural recording you can download: https://www.hdtracks.com/index.php?file=catalogdetail&valbum_code=HD090368035264

FTGV

sheathensemble wrote: »

...1c's are old and I am just curious as to if the bandwidth used would be subject to change...

Assuming the paralleled pair of dimensional drivers are half the impedance of the stereo pair they should have the same upper (2.5k?)limit, at least according to the crossover values in the posted 1c schem.

TennMan wrote: »

Based on the amount of attenuation would it be possible to adjust the SDA effect.

Short of separating and driving the Dimensional and Stereo arrays from different amps you would need increased driver sensitivity for an increase in the SDA level.Adding resistance to the left to right speaker SDA link would reduce SDA level.

FTGV

FTGV wrote: »

Short of separating and driving the Dimensional and Stereo arrays from different amps you would need increased dimensional driver sensitivity for an increase in the SDA level.Adding resistance to the left to right speaker SDA link would reduce SDA level.

Clarified.
Ofcourse experimentation may not prove fruitful since Polk have already set the level to where their R&D dept. felt ideal.

Schurkey

First Guess: Easiest way to do this is to buy a Carver Sonic Hologram generator, and record the output.

sheathensemble

DarqueKnight wrote: »

The SDA signal level is 1/2 of the stereo signal level.



You will still have interaural crosstalk due to one ear hearing sound from the opposite side speaker. This cannot be dealt with in a recording because it is an artifact of the listener's hearing caused by each ear hearing two speakers.

The only way to get rid of interaural crosstalk without some type of active or passive cancellation circuitry is to use headphones.

This cross talk cancellation is what I am trying to impart into a recording...

For example let's say you have any old audio file...

-Import it into a DAW.

-Clone it so you have a duplicate of the original file.

-Take the duplicate and reverse its polarity and swap its stereo.

-Place a filter or EQ of some kind on the duplicate.

-Set this filter/EQ up to cut out all frequencies below 200 hz and above 1000hz.

-Adjust the volume of the duplicate track to half of the original files volume.

-Export new file that contains the new processed duplicate file.



That's the theory anyways...


And yes recording the output of a carver unit may be a simple solution... If you own a carver unit. I am trying to accomplish similar processing using only host vst plugins in a basic DAW.

I am still experimenting with the bandwidth and amount of attenuation on the duplicate track. So far I would say I have heard some results, nothing yet that I would publish though.

Thanks for the Info DK, Very appreciated.

Cheers:cheesygrin:

sheathensemble

Face wrote: »

Here's an binaural recording you can download: https://www.hdtracks.com/index.php?file=catalogdetail&valbum_code=HD090368035264

Binaural recording is the 3 dimensional recording technique I was talking about. So hopefully using this technique I will be able to make a recording which captures the timing cues and attenuation cues of the binaural recording technique and use digital processing based off of SDA technology to incorporate intramural cross talk cancellation.

Schurkey

sheathensemble wrote: »

This cross talk cancellation is what I am trying to impart into a recording...

For example let's say you have any old audio file...

-Import it into a DAW.

-Clone it so you have a duplicate of the original file.

-Take the duplicate and reverse its polarity and swap its stereo.

-Place a filter or EQ of some kind on the duplicate.

-Set this filter/EQ up to cut out all frequencies below 200 hz and above 1000hz.

-Adjust the volume of the duplicate track to half of the original files volume.

-Export new file that contains the new processed duplicate file.



That's the theory anyways...


And yes recording the output of a carver unit may be a simple solution... If you own a carver unit. I am trying to accomplish similar processing using only host vst plugins in a basic DAW.

I am still experimenting with the bandwidth and amount of attenuation on the duplicate track. So far I would say I have heard some results, nothing yet that I would publish though.

Thanks for the Info DK, Very appreciated.

Cheers:cheesygrin:

1. Wouldn't you also need a time-delay on the "correction" signal to mimic the Stereo-to-SDA driver distance? (which itself mimics the distance between the ears of a typical person.)

2. Just a guess--you need to cut the volume to 1/2 BEFORE you scrape away the frequency extremes, not after.

3. I don't understand where the idea of the SDA signal being 1/2 volume comes from. What in the crossover reduces the SDA volume? I'd have said the SDA signal is full-volume, at least in my 1Bs, where the SDA cable carries a full-voltage, full-bandwidth SDA signal to the opposite speaker cabinet. Once inside the opposite cabinet, there's a bigass capacitor and inductor to filter the bandwidth...but no resistor to cut the volume. Different generations of SDA maybe have different amounts of SDA volume???

ambiophonics

To quote my SN - what you are looking to accomplish is very similar to "ambiophonics". Basic crosstalk cancellation, if you google ambiophonics you will find several downloadable java apps that can help you accomplish what you are looking for. Move your speakers closer together (about 20 degrees apart rather than the normal 60) and ambiophonics will blow your mind.

DarqueKnight

TennMan wrote: »

Based on the amount of attenuation would it be possible to adjust the SDA effect.

Of course.

TennMan wrote: »

Have you ever tried tinkering with it?

No. This is why:

FTGV wrote: »

Of course experimentation may not prove fruitful since Polk have already set the level to where their R&D dept. felt ideal.

The article authored by Matthew Polk and published in the June 1984 issue of Audio magazine explains in mathematical detail the theoretical foundations of SDA technology ("Polk's SDA Speakers - Designed-In Stereo").

The level of the crosstalk cancellation signal must be set at a level where the crosstalk signal is effectively eliminated, yet there is no interference with the direct stereo signal and the localization signals. The optimum level for the cancellation signal was found to be 1/2 that of the stereo signal.

sheathensemble wrote: »

This cross talk cancellation is what I am trying to impart into a recording...

sheathensemble wrote: »

Binaural recording is the 3 dimensional recording technique I was talking about. So hopefully using this technique I will be able to make a recording which captures the timing cues and attenuation cues of the binaural recording technique and use digital processing based off of SDA technology to incorporate intramural cross talk cancellation.

Again, you cannot impart crosstalk cancellation in a recording because you will still have each ear hearing the sound from two speakers.

Binaural recordings must be listened to on headphones in order to achieve the desired crosstalk cancellation effect. While the binaural recording and headphone playback process can result in a wide sound stage, the process is not adequate for reproducing the sound localization signals. The process is also not representative of the way we hear. When we listen to a live concert, we hear sounds from multiple locations and each ear does not work in isolation. The brain computes intensity, phase and arrival time to determine the location of a sound.

From Matthew Polk's 1984 article:

"Although the elimination of interaural crosstalk seemed to give significant advantages to headphones, the phones still failed to produce a convincing sonic illusion." (p. 34, col. 3.)

"The width of sound stage presented by loudspeakers is limited by interaural crosstalk. The stability of the sonic image of headphones is limited by the lack of realistic directional cues. This meant that we would have to do more than eliminate interaural crosstalk. In addition, we would have to find a way for the directional cues contained in the recordings, such as they are, to reach the listener's ears in a manner acceptable to the hearing process." (p.37, col. 1)

Therefore, headphones (binaural listening) can remove interaural crosstalk and expand the sound stage. However, the binaural recording and playback process cannot adequately reproduce a three dimensional sound stage with sharply defined image localization.

Read the above referenced paper by Matthew Polk for details. From there you can do a more intensive investigation into the properties and limitations of binaural and stereo reproduction.

sheathensemble

Schurkey wrote: »

1. Wouldn't you also need a time-delay on the "correction" signal to mimic the Stereo-to-SDA driver distance? (which itself mimics the distance between the ears of a typical person.)

2. Just a guess--you need to cut the volume to 1/2 BEFORE you scrape away the frequency extremes, not after.

3. I don't understand where the idea of the SDA signal being 1/2 volume comes from. What in the crossover reduces the SDA volume? I'd have said the SDA signal is full-volume, at least in my 1Bs, where the SDA cable carries a full-voltage, full-bandwidth SDA signal to the opposite speaker cabinet. Once inside the opposite cabinet, there's a bigass capacitor and inductor to filter the bandwidth...but no resistor to cut the volume. Different generations of SDA maybe have different amounts of SDA volume???

1- I see what your saying. I can't comprehend what the reasoning for that would be. TO me it seems that while having the processed signal coming from a driver of it's own would be beneficial; I also think that having everything come form one source location would have it own benefits In regards to timing... Perhaps this is even a draw back to using separate driver's to carry the dimensional signal. I don't know, If anyone has some theory on this... Chime on in.

2- You are correctamondo, but just for the purpose of following a formula

3- It comes from DK... I asked, he answered LOL. I remember hearing in a previous thread that the dimensional driver's were attenuated to some extent. Too what extent exactly... I am unsure.


So far I am just experimenting by ear... I have just posted a sample that I think is showing some progress in the right direction. I will state that I did not use the 200-1000 hz bandwidth described in the sda1c review Posted earlier and also the attenuation was done to an amount I found worked (by ear). I've only played with this since my last post so I am sure I will be making lots of changes in regards to both the bandwidth and amount of attenuation of the duplicate track. But anyways here is a link to the first sample... feel free to check it out and let me know what you think.

http://www.soundcloud.com/dimensional-mixes/burial-un-true-dimensional-mix

To anyone else who may be checking it out...
Listen to it on a normal stereo set up... or if you only have SDA's I guess you could pull the plug on them and use them as a normal stereo rig...

Schurkey

DarqueKnight wrote: »

Again, you cannot impart crosstalk cancellation in a recording because you will still have each ear hearing the sound from two speakers.

I understand that the Carver Sonic Hologram generator and the SDA produce similar but not identical sonic results.

That said, if an ordinary pair of stereo speakers can produce interaural crosstalk cancellation based on manipulation of the stereo signal by the Carver unit--rather than using separate speaker drivers having time-and-phase compensation as in the SDA approach...

...why can't that manipulated stereo signal of the Carver unit be recorded, or a similar process applied to the recording? The stereo speakers don't know if the signal is freshly processed by the Sonic Hologram generator, or if the manipulated signal is coming directly off the recording. All they react to is instantaneous voltage and current--not whether the processing took place prior to the recording, or after it's been replayed. If the recording has the crosstalk cancellation time-and-phase information encoded, it couldn't matter that the processing was done prior to the recording.

I want to be clear: I don't think recording the SDA/Sonic Holography effect is desirable. Once the recording has been tampered with, you'll never get it pristine again. BUT--as an intellectual exercise--I don't understand why the interaural crosstalk cancellation can't be included in the recording.

sheathensemble

Schurkey wrote: »

I understand that the Carver Sonic Hologram generator and the SDA produce similar but not identical sonic results.

That said, if an ordinary pair of stereo speakers can produce interaural crosstalk cancellation based on manipulation of the stereo signal by the Carver unit--rather than using separate speaker drivers having time-and-phase compensation as in the SDA approach...

...why can't that manipulated stereo signal of the Carver unit be recorded, or a similar process applied to the recording? The stereo speakers don't know if the signal is freshly processed by the Sonic Hologram generator, or if the manipulated signal is coming directly off the recording. All they react to is instantaneous voltage and current--not whether the processing took place prior to the recording, or after it's been replayed. If the recording has the crosstalk cancellation time-and-phase information encoded, it couldn't matter that the processing was done prior to the recording.

I want to be clear: I don't think recording the SDA/Sonic Holography effect is desirable. Once the recording has been tampered with, you'll never get it pristine again. BUT--as an intellectual exercise--I don't understand why the interaural crosstalk cancellation can't be included in the recording.

I agree, In terms of SDA speakers, it's like having both the dimensional and normal stereo signals being played through one full range signal...

F1nut

sheathensemble wrote: »

My main goal is to use this processing along with 3 dimensional recording techniques to achieve a true 3 dimensional recording.

Already been done, http://en.wikipedia.org/wiki/QSound

bmbguy

sheathensemble wrote: »

I agree, In terms of SDA speakers, it's like having both the dimensional and normal stereo signals being played through one full range signal...

Which then reach both ears at different times, introducing more crosstalk... Aren't you then right back where you started?

sheathensemble

F1nut wrote: »

Already been done, http://en.wikipedia.org/wiki/QSound

Well I haven't found a tutorial yet... I am trying to accomplish this with the stardard plugins that come with every decent DAW. For the sake of Joe Blow.

bmbguy wrote: »

Which then reach both ears at different times, introducing more crosstalk... Aren't you then right back where you started?

Ok I see what you guys are saying about the need for a delay in the signal. Hey there is always room for improvement! Extremely small timing issue, but it sould make a huge difference.




once again, http://soundcloud.com/dimensional-mixes/burial-un-true-dimensional-mix

sheathensemble

speed of sound... 343.2 metres per second

Distance between ears... 6 inches

I don't think my software is capable of making such compensation.

DarqueKnight

Schurkey wrote: »

I want to be clear: I don't think recording the SDA/Sonic Holography effect is desirable. Once the recording has been tampered with, you'll never get it pristine again. BUT--as an intellectual exercise--I don't understand why the interaural crosstalk cancellation can't be included in the recording.

sheathensemble wrote: »

I agree, In terms of SDA speakers, it's like having both the dimensional and normal stereo signals being played through one full range signal...

I am an avid practitioner of modifying stereo equipment to improve performance. My starting point for all such exercises is gaining a thorough understanding of what the manufacturer's design. From there, I determine what things to change in order to achieve improved performance. In the context of this thread, an excellent starting point is Matthew Polk's 1984 technical article.

To answer your questions:

Consider the stereo recording and playback process: In two channel stereo recording, a pair of microphones are set up which mimic the hearing response of human ears. This pair of microphones records the frequency, phase, intensity and arrival time of multiple sound sources arranged in three dimensional space in front of them.

At playback, we have the difficult task of recreating the frequency, phase, intensity and arrival time of multiple sound sources from just two sound sources, the loudspeakers. Since the two sound sources (loudspeakers) destructively interfere with each other, the sound stage size and sound localization cues are compromised.

The question that needs to be understood is, do crosstalk and diminished localization cues originate with the original sound source or are they artifacts of using two loudspeakers to reproduce multiple sound sources?

The answer, of course, is that sound stage size restriction and image localization problems originate with the loudspeakers. Therefore, the optimum solution would focus on solving the problem at the source of the problem: the loudspeakers. Saying that you want to fix interaural crosstalk and localization errors in the recording does not make sense because interaural crosstalk and localization errors do not originate with the recording. Altering the frequency, phase, intensity and arrival time information in the original recording will result in a sonic mess that is no way representative of the sound "heard" by the stereo microphones during the recording session or by the ears during a live music performance. Ideally, we want to keep the signal "heard" by the recording microphones as pure as possible, all the way out to the loudspeakers.

The Sonic Holography solution offered by Carver was not entirely satisfactory because it did not address both interaural crosstalk and image localization, only interaural crosstalk.

The passive acoustic solution integrated within the crossovers of SDA loudspeakers was a more efficient and natural sounding process because it allowed the loudspeakers to fix the problems that were created by them. Moreover, the dual issues of interaural crosstalk and image localization were addressed interactively.

sheathensemble

DarqueKnight wrote: »

The Sonic Holography solution offered by Carver was not entirely satisfactory because it did not address both interaural crosstalk and image localization, only interaural crosstalk.

The passive acoustic solution integrated within the crossovers of SDA loudspeakers was a more efficient and natural sounding process because it allowed the loudspeakers to fix the problems that were created by them. Moreover, the dual issues of interaural crosstalk and image localization were addressed interactively.

"Image Localization" seems to be a pretty self explanatory term. But what you have said here is making me wonder what you're referring to. Can you further explain what you mean when you say image localization? But what device is deployed in SDA's that targets this? I always thought that it was the cross talk cancellation which was responsible for the localization effect through out the sound stage. To me, it seems like all of the parameters of SDA technology can be replicated by software.

I didn't notice any loss of imaging in the mix I posted here. Just a wider sound stage. Which is what I was trying to achieve and will continue to try and improve. Once again here's the sample I made earlier: http://soundcloud.com/dimensional-mixes/burial-un-true-dimensional-mix

If you have not already please have a listen, It seems to be working to me... but then again maybe I am just hearing what I want to and belong in a mad house.

sheathensemble

Well I would hate to say it... BUT THIS REALLY WORKS! I think I have a winning formula already. I will upload some more tracks very soon! probably tomorrow...:cheesygrin:

Drenis

Given that I've never heard SDA's, I listened to your recording and I'm not sure what I'm listening for. I heard traits that I've heard used in other mixes or similar music genres as far as a wider sound stage. I guess it didn't really wow me. Perhaps it's the speakers I'm using?

heiney9

sheathensemble wrote: »

"Image Localization" seems to be a pretty self explanatory term. But what you have said here is making me wonder what you're referring to. Can you further explain what you mean when you say image localization? But what device is deployed in SDA's that targets this? I always thought that it was the cross talk cancellation which was responsible for the localization effect through out the sound stage. To me, it seems like all of the parameters of SDA technology can be replicated by software.

I didn't notice any loss of imaging in the mix I posted here. Just a wider sound stage. Which is what I was trying to achieve and will continue to try and improve. Once again here's the sample I made earlier: http://soundcloud.com/dimensional-mixes/burial-un-true-dimensional-mix

If you have not already please have a listen, It seems to be working to me... but then again maybe I am just hearing what I want to and belong in a mad house.

In talking with Matt Polk at the last Polkfest held at HQ he said it took he and another engineer 7 years to write the DSP software for what was then the Surround Bar 360. He said he even had to wait for technology to catch up, so yes, today the software is available, but just as little as 5 years ago it wasn't.

He envisioned an new line of SDA's (non-surround bar type) having both active and passive processing, if he were to do a new SDA line-up of floor stander/bookshelf speakers. This is coming from the guy that worked with this process for 25+ years.

heiney9

Q-sound is a similar manipulation done at the mixing phase and it's very exaggerated and unrealistic and unnatural compared to passive SDA's. In fact I have passed on buying cd's mixed with Q-sound because it bugs me so much because it's so processed and fake sounding.

Software is definitely available but in the end I'm not sure the same results are possible, unless you are talking about a Polk produced product like the Surround Bar, which I was very skeptical of until I experience it. It was very convincing.

H9

Joe08867

You need to do this new method you tried on something listenable. Sorry had to say it. And also something that isn't a processed electronica sound. You are just cutting out certain things inherent to the music itself. Lets hear this on a rock tune or something with actual instruments. Honestly I couldn't tell a difference because I do not know the material. And listening to a Youtube version showed very little change.

Also: for some reason the SDA effected sound gave my ears a high pitched ring that I did not hear on the original youtube version. It was quite uncomfortable. I literally had a whining sound in my left ear that hurt. I moved to four different spots of the recording to see if it was just a section but it is throughout the recording.