Here's another good article worthy of everyones attention, this time it's a FAQ on Ambisonics:
Ambisonics FAQ
Have fun
Ambisonics FAQ
Have fun
derf said:
The circuits shown seem to attemp to do such, but if you further calculate the different mic positions and the the phase differences among the different mic directions, not really. At least not shown in the so called decoding circuits posted here or earlier.
derf said:
From a point of view for cancellation, the so called decoder only cancels the effects that are 180 out of phase. So the real decoder must do something else, but explanation is only good enough for an average user, not for a DIYer.
derf said:
I would think the Ambisonics people would if they did actual analysis prior to applying the technique.
derf said:
2+2+2 is just the speakers that I mentioned. Since they have explained the speaker setup in such a way that technically it is possible to have good replay potential, it is worth a try to listen. With the Ambisonics, they just explain it like it's magic and you can do whatever, even without a decoder, not explanation on speaker positioning. I wonder what speaker setup you use to make this technique sound so convincing?
derf said:
Where is the correction? There are not even any research reports showing the concept.
derf said:
I don't know. Most explanations are good for people with not much technical knowlegde, I think most people here have some technical background. I think it's not just Ambisonics or not, it's a question which out of so many techniques is technically better. The figure 8 mic is the explanation, but in reality the recording mic only has 4 mics located in directions not as the figure 8 directions. Nothing explains that.
If any explanation had been published in Wireless World before, please point out the year and issue, and I will dig it up.
Hmm, this isn't done in the Ambisonic decoders. It's done in the mic stage in the same box as the preamp. No-one has posted anything regarding this yet...
TBH, I don't think any cancellation goes on at all. The Soundfield just uses patterns(i.e figure 8) that diminish in such a way that they can use the phase of the arriving signals at the microphones(in phase/out of phase, x degrees in/out) to determine the angle/point of arrival.
I must admit, I'm not Ambisonics person, yet. I've yet to find/build myself a suitable decoder, but I'm sure once I do, I'll be able to further explain the whole shebang.
OK, 2+2+2 is the speakers, alright. So what signal from what mic goes where, are you duplicating a signal from one stereo pair with the 4 front speakers and 2 rear?. It's all pretty confusing...
Hmm, it might be possible to have good replay, better than Stereo even. But technically, it's still got the same problems.
Ambisonics isn't magic, but it's as close as you'll get to recreating a soundstage in 3 dimensions and is inherently very flexible. Of course, certain speaker configurations are preferable to others(i.e some kind of square, rectangle etc), but if the speaker is off out the way somewhere, the decoder can compensate. No-one has mentioned not using a decoder to get Ambisonics, although while we're on the subject, you can get an enhanced Stereo effect by replaying UHJ through just the two channels.
The correction is done inside the Soundfield microphone, the capsules are arranged in a tetrahedral pattern to make it easier for the mathematicians to do their job.
Have you read these texts?
My guess would be you haven't. They may not be incredibly technical, but they provide some very good information regarding Ambisonics and sound capture/reproduction in general. I highly reccomend you read them both, before you get back to me, it'll stop us needlessly going over the same ground.
Nothing explains that, was you listening when I posted the reply from a Soundfield representative earlier?
A figure 8 pattern(they're all hypercardoid capsules)will receive in phase information in the front part of the 8 and out of phase information in the rear, this helps the mic dertermine the location of the sound arriving at the capsule.
In most people who've encountered/worked with Ambisonics, using a Soundfield microphone to capture B-Format is technically the best capture device at this present time, even if it never gets decoded, it still provides superior Stereo reproduction. Which really is quite a feat, because it's not what it's really designed around, but solid engineering has made it like this.
If you dig about Ambisonic.net, there is plenty of articles, I'm sure some may even be pretty technical. The Soundfield website does a good job of explaining their microphones and seem to answer questions swiftly via email. There's also the University of York, which has quite a bit of information pertaining to Ambisonics.
The information is almost all out there, whether you read it/absorb it or not, is up to you
.
derf said:
Dragonmaster posted some DIY encoders(at least that is what they called it) in post #30. Looking at that and the mic positions, there seems not realtion with the figure 8.
I have found some reference articles in published in 1977 Wireless World, I will get those and have a look at them (wonder why they never got to part three though). I like information published by wireless world because they provide just enough information to make it clear.
I think you are taking this too personal. This is just a discussion of the recording technology itself, and has nothing to do with whomever personally.
The information provided by different members up to now probably is good for consumers. Fun reading.
Where does it tell you where to put the speakers to get accurate reproduction?
I'm not taking it personally, honest
. Although, sometimes I do tend to sound like I am 
This is very true, the information provided by me and others can only provide some of the story, that's why I place so much emphasis on texts written elsewhere, as they cover the topic more thoroughly than me or probably anyone else could in any forum post.
One of the reasons I end up sounding like I'm taking it personally, is because I really believe in Ambisonics, it's based on some very concrete grounding and I think it's the nearest we've got to accurate soundfield capture/reproduction.
I think 3D sound is the way of the future. It's just how it's implemented that will make it acceptable to the recording engineer, and the listener. Other techniques gain popularity not because they are technically better, but that they are flexible enough to accomodate variety of recording, processing, and playback conditions.
There are many recording and playback techniques. Binaural recording can be good if it were expanded in scope. Digital process is to a point that you can take two channels and expand it to surround selecting the type of room you what(not 3D). As digital processing power increases, there will be no need to record the ambient signals, but can be recreated just by modeling the sonic characteristics of specific location.
All this 3D stuff also require right speakers. I think that with the popular 5.1 setup, I would add one more above the main speakers, and one more below. This would also be compatible with the 7.1 system which is probably already available. Although technically you really only need 4 layed out like a sleeping pyramid, very high quality and specific speakers would be necessary, but not commonly available. So taking into consideration the common home situation, 7.1 is probably the most realistic approach.
As for Ambisonics, I think the recording technique makes it convenient for certain specific recording situation in movie making, I don't think it will become popular enough in the home situation to become wide spread, but if it can be cheaply bundled into existing products, there may be a chance. The same goes to the MDG 2+2+2.
I'm working on getting things 3D ready.
There are many recording and playback techniques. Binaural recording can be good if it were expanded in scope. Digital process is to a point that you can take two channels and expand it to surround selecting the type of room you what(not 3D). As digital processing power increases, there will be no need to record the ambient signals, but can be recreated just by modeling the sonic characteristics of specific location.
All this 3D stuff also require right speakers. I think that with the popular 5.1 setup, I would add one more above the main speakers, and one more below. This would also be compatible with the 7.1 system which is probably already available. Although technically you really only need 4 layed out like a sleeping pyramid, very high quality and specific speakers would be necessary, but not commonly available. So taking into consideration the common home situation, 7.1 is probably the most realistic approach.
As for Ambisonics, I think the recording technique makes it convenient for certain specific recording situation in movie making, I don't think it will become popular enough in the home situation to become wide spread, but if it can be cheaply bundled into existing products, there may be a chance. The same goes to the MDG 2+2+2.
I'm working on getting things 3D ready.
Actually, I have posted one encoder which was a page with a micrphone+encoder and 4 "de"coder schemes on the same page. (Was that link: http://www.geocities.com/ambinutter/AMBISONIC_HOME_PAGE.html )
The Integrex "de"coder(in the previous link) is an article by Wireless World. There is a lot of side information of how it works. The disadvantage to this and most older decoders is that there's no Z input because consumers only had stereo storage mediums. But since at least 3 channel mediums and the maths are there, it would be possible to add this possibility to a decoder, but it would only work on B-format or one of the 3 other unreleased UHJ formats(3-4 channel).
More technical explanation:
http://members.tripod.com/martin_leese/Ambisonic/diy.html
Actually, the only thing that seems to be repeated and insisted on in every FAQs is that the advatage of Ambisonic over quad, 5.1... is that there's no particular sweet spot or position(eg. have to look to the front) to have.
I think that the advantages with Ambisonic is that they don't use pair-wise mixing like x.1, there's height information and you can have as much speakers as you want.
Yes, Ambisonic is easy to add. Feed in the stereo input to the decoder, and feed the output to a receiver. But, a problem that I think is Integrex specific is that there are no channel specific outputs(It uses 4 amps and a complex speaker wiring). So if your system is going to be non-Ambisonic specific, you will have to replicate the arrangement you want by placing some opamps in the scheme to make separate outputs. Also, Ambisonic seems to be easy to adapt to a x.1 system: There's an hexagonal arrangement in the Integrex decoder documentation page that is similar to 6.1(Center front, Center rear, rear left, rear right, front left, front right).
The Integrex "de"coder(in the previous link) is an article by Wireless World. There is a lot of side information of how it works. The disadvantage to this and most older decoders is that there's no Z input because consumers only had stereo storage mediums. But since at least 3 channel mediums and the maths are there, it would be possible to add this possibility to a decoder, but it would only work on B-format or one of the 3 other unreleased UHJ formats(3-4 channel).
More technical explanation:
http://members.tripod.com/martin_leese/Ambisonic/diy.html
Actually, the only thing that seems to be repeated and insisted on in every FAQs is that the advatage of Ambisonic over quad, 5.1... is that there's no particular sweet spot or position(eg. have to look to the front) to have.
I think that the advantages with Ambisonic is that they don't use pair-wise mixing like x.1, there's height information and you can have as much speakers as you want.
Yes, Ambisonic is easy to add. Feed in the stereo input to the decoder, and feed the output to a receiver. But, a problem that I think is Integrex specific is that there are no channel specific outputs(It uses 4 amps and a complex speaker wiring). So if your system is going to be non-Ambisonic specific, you will have to replicate the arrangement you want by placing some opamps in the scheme to make separate outputs. Also, Ambisonic seems to be easy to adapt to a x.1 system: There's an hexagonal arrangement in the Integrex decoder documentation page that is similar to 6.1(Center front, Center rear, rear left, rear right, front left, front right).
Hopefully I'll get better understanding after I read the WW articles, which seems to be the beginning of all this.
All these information do not address the following:
1. Difference in speaker location will yield difference in playback accurracy. If someone can explain one speaker combination and how a decoder would be adjusted to accomodate that, then it would be clear. I think what is technically correct is the issue, not just broad statements saying you can use as many as you want. If the location of speakers do not matter, does this mean I can line all speakers against a single wall?
2. Transient and phase characteristics when recording the source is very improtant to preserve the original imaging. When I read an article in WW about binaural recording and playback, the author clearly addressed how correct souce allocation is preserved throughout the recording and playback chain mathematically, taking into acount human hearing FR differences when speakers are located at different locations. Nowadays people use a more complicated Head Related Transfer Function. Up to now, still there is no information that explains this kind of information. If just adding, subtracting, gain adjustment is all to it, those are far behind in technology and complexity of the real situation.
Now I think things may sound remarkable even if they are not technically correct, but then not all kinds of music sources will sound as pleasing.
Additionally, if you are emotionally involved, then it is difficult to make an objective assessment.
We have too many sales people we meet every day here, so we just have to be more technically involved to sort out sales talk and the facts. It's just becomming a habit.
Generally, if I don't get a direct answer to a question, just side stepping to a different issue. Then I know that the answer is not available.
All these information do not address the following:
1. Difference in speaker location will yield difference in playback accurracy. If someone can explain one speaker combination and how a decoder would be adjusted to accomodate that, then it would be clear. I think what is technically correct is the issue, not just broad statements saying you can use as many as you want. If the location of speakers do not matter, does this mean I can line all speakers against a single wall?
2. Transient and phase characteristics when recording the source is very improtant to preserve the original imaging. When I read an article in WW about binaural recording and playback, the author clearly addressed how correct souce allocation is preserved throughout the recording and playback chain mathematically, taking into acount human hearing FR differences when speakers are located at different locations. Nowadays people use a more complicated Head Related Transfer Function. Up to now, still there is no information that explains this kind of information. If just adding, subtracting, gain adjustment is all to it, those are far behind in technology and complexity of the real situation.
Now I think things may sound remarkable even if they are not technically correct, but then not all kinds of music sources will sound as pleasing.
Additionally, if you are emotionally involved, then it is difficult to make an objective assessment.
We have too many sales people we meet every day here, so we just have to be more technically involved to sort out sales talk and the facts. It's just becomming a habit.
Generally, if I don't get a direct answer to a question, just side stepping to a different issue. Then I know that the answer is not available.
Yes, you can put all the speakers on the same wall. Ex: 4 speakers to the front wall that are placed eg. like this: Front left, mid-left, mid-right, right.Then, you tell the decoder/build the circuit to retreive the information from these places. You will not get the rear sound tho.
There are no limitations to the position, all you need is the good speaker position.
What I don't understand tho is when you record with multiple figure-of-8 microphones, how are they placed and mixed. I don't think you can just mix the X capsule with an other X looking at a different place. Maybe there's some kind of circuit similar to the decoders?
There are no limitations to the position, all you need is the good speaker position.
What I don't understand tho is when you record with multiple figure-of-8 microphones, how are they placed and mixed. I don't think you can just mix the X capsule with an other X looking at a different place. Maybe there's some kind of circuit similar to the decoders?
DragonMaster said:
We can use any system and put as many speaker how ever we want even without Ambisonics, so it really isn't that special. I just read the first part of WW, but first glance is that:
1. I does not specify speaker location. Some of the ways speakers are connected gives no regard to impedance effect on amp performance.
2. There is gain compensation for speaker distance, but no time compensation (delays)
3. There is no explanation as to how localization of the original performer is preserved.
4. It seems to just use low frequency phase shifring to create an illusion of large space by partially out of phase playback.
5. It doesn't talk about recording yet.
So still lots of understanding is necessary to see if this is a valid method or not.
I agree with you for this point.
This decoder is not perfect, as there are some ProLogic II decoders without angle,delay and distance adjustments.
Note that one sentence in the article is telling that they talked about it in the past and an other one is talking about future articles on recording. The problem is to find these.
DragonMaster said:
The Ambisonic mic setup looks elegant, but the quality of the recording would rely on matched calibrated mics to even get acceptable imaging because of the close distance between each capsule. Similar technology would be a phase array radar where each antenna need to meet spedific toleranes in order to get accurate performance. This means each capsule must pass through a calibration circuit before summing signals in the encoder. This is not done it the published circuits.
derf said:
Looks like a religious figure.
I'm still going through all these Bibles, but nothings ticking yet.- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.