DIY Waveguide loudspeaker kit - Page 62 - diyAudio
Go Back   Home > Forums > Commercial Sector > Manufacturers > GedLee

GedLee Home of the renown Geddes Loudspeakers

Reply
 
Thread Tools Search this Thread
Old 31st July 2008, 10:16 PM   #611
gedlee is offline gedlee  United States
diyAudio Member
 
Join Date: Dec 2004
Location: Novi, Michigan
Quote:
Originally posted by noah katz
I'm delighted to hear of the availability of these kits.

Dr. Geddes, do I understand correctly from the Nathan 10 thread that it's too late to get in on the the first batch of 12" systems?

I have a question about the earlier discussion on using a non-ideally placed center channel speaker vs. none.

When the suggestion was made to create a phantom center image with a pair of speakers to the side of the screen (the left and rights being more widely spaced), it was pointed out that this would create problems with comb filtering.

Why isn't the same criticism apropos for a phantom center using the L/R's?

Dr. Geddes, have you done any tests on the audibility of comb filtering?

I thought that anything less than 1/6 (or is it 1/3) octave wide inaudible; wouldn't this apply to comb filtering?

I can't say for sure that you wouldn't make the first build of the Abbey12s, but its close. There are too many variables. Clearly, once I am setup to run parts I run as many as I can or need to. So basically, if your on the list there is a good chance, unless I run out of material or something. I was able to run all the Nathans that I had when I took down the setup, but more have come in since then.

I think that 'comb filtering" in the way that you are describing it is a misnomer. It's a spatial comb filter, rather than one in time. The sources are seperated in space, but not in time. Hence, as long as you are near the centerline there is no temporal comb filtering, and what you allude to does not apply. As you get further off axis then both the polar response and the temporal response will degrade. But this would take some twenty degrees or more. With a true center channel this off-axis effect will not happen, but thats all.

The test that I did was when I was listening to a DVD that was in dolby digital 5.1. For some reason the center channel was not working, but yet there was still a clear center image. I checked this on several players and they all did the same thing - there was no center channel in the mix! But there certainly was no apparent flaws with the center "phantom" image. In fact I had listened to this disk several times before and never noticed this lack of center channel. But since there was no center channel there was no way to actually A-B any differences.
  Reply With Quote
Old 31st July 2008, 10:57 PM   #612
diyAudio Member
 
Join Date: Jul 2004
Location: Mountain View, CA
"I think that 'comb filtering" in the way that you are describing it is a misnomer. "

I don't think I described it; I referred to others' objections that multiple center channel speakers was undesirable because of it.

"But this would take some twenty degrees or more."

Can I conclude then that you don't agree with said objections?

Thanks
__________________
-----------------------------------------
Noah
  Reply With Quote
Old 31st July 2008, 11:11 PM   #613
gedlee is offline gedlee  United States
diyAudio Member
 
Join Date: Dec 2004
Location: Novi, Michigan
Quote:
Originally posted by noah katz
"I think that 'comb filtering" in the way that you are describing it is a misnomer. "

I don't think I described it; I referred to others' objections that multiple center channel speakers was undesirable because of it.

"But this would take some twenty degrees or more."

Can I conclude then that you don't agree with said objections?

Thanks

No I don't agree. I said that before.

I would prefer to use a phantom center than any center that is not behind the screen and identical to the LR. If you can have an identical center and its directly behind the image, then this is hard to beat - thats what I do - but if you can't do that, then go with a phantom center.
  Reply With Quote
Old 31st July 2008, 11:25 PM   #614
diyAudio Member
 
Join Date: Jul 2004
Location: Mountain View, CA
"No I don't agree. I said that before. "

Sorry, I missed that, thanks.
__________________
-----------------------------------------
Noah
  Reply With Quote
Old 1st August 2008, 12:14 AM   #615
gedlee is offline gedlee  United States
diyAudio Member
 
Join Date: Dec 2004
Location: Novi, Michigan
I meant in another thread, somewhere, we were talking about center channels and phantom images. I don't think that it was in the thread. I'm not sure where your comment was from. I think that some people are concerned with putting an MTM on its side for the center, and that concerns me too.

The phantom center image is what we have always had with stereo. I have always wondered about the lateral polar response from a center panned image as it would have to be very narrow with lots of lobes. Perhaps this is why I don't usually like the speakers too far appart and maybe why I, and many other, have found the large toe-in that I recommend to be so effective.

I have always thought that a better solution would be to seperate out the common signal from the left and right and feed it to a center speaker and not the left and right. The signal processing for this would be tough in analog but certainly doable in DSP. But this IS exactly what 5.1 does with the front signal so it is done aleady in that situation. Again, maybe, this is why I always seem to prefer the DVD audio to CD as this processing has already been done.

And why not just feed the center with L+R? Because this leaves the common signal at the L & R. It should be removed and this isn't so easy to do. Might make a good thesis project!
  Reply With Quote
Old 1st August 2008, 01:29 AM   #616
KSTR is offline KSTR  Germany
diyAudio Member
 
KSTR's Avatar
 
Join Date: Jul 2007
Location: Central Berlin, Germany
Earl,

maybe you want to have a look at some of the linear rematrixings, available under marketing names like "Trinaural" (by James Bongiorno of Spread Spectrum Technologies) or "Optimum Linear Matrix" (by Michael Miles of MilesTech). There are some threads about this topics here, too.

To my knowledge it is not possible to extract a center signal from a LR signal and getting a null output at the same time from L&R for that center signal with any "stupid" linear matrix (that is, not analysing the LR-content for correlation etc. BTW, Lynn Olson has expertise on such steered mechanisms and knows a lot about the problems that come with "intelligent" analysis of the signal).

While with a basic static rematrixing of L & R into L' & C & R' like
L' = L - 0.5R
R' = R - 0.5L
C = 0.5(L+R)
we don't get silent side channel signals for a mono content, there at least is 6dB less output from each side channel compared to the center (and this is, as referenced in the Miles paper, the maximum seperation one can obtain). This rematrixing, in my personal experience as well as that from others who have tried it, greatly improves the depth and "being there" sensation and the "stereo effect" in general (on like 99% of all recordings, problems are only with HRTF-related imaging tricks like Q-sound that rely on the standard setup by design), making direct (centered) sounds more direct and ambient sounds more ambient. Effectively it swaps roles of phantom and real image sources in a benfical way for most recordings (especially close-miked, panpot-stereo rock/pop/jazz stuff)

Also it widens the sweet spot dramatically in the sideways direction (at the cost of a smaller front-to-back extension), without too much relying on the trading effect (which is what you do with huge toe-in and beaming speakers). But for me the most notably difference is that the stereo image does not collapse that much when you turn your head, compared to the two-speaker setup.

You should really give it a try (the matrix is easily realized with a few opamps), controlled directivity speakers like yours are certainly a premium choice for a trinaural setup. It's different in some sapects (for instance room exitation and some perceived exaggerated focussing on the center for the first few hours or even days of listening), but it's still pretty much compatible to normal stereo recordings and typically produces a better stereo illusion.

I know this all reads like typical marketing bla-bla, but I swear I never experienced such a dramatic improvement for 2-channel stereo reproduction. While better speakers and excellent room acoustics are big steps forward without doubt, they don't have that "quantum leap"-character of improvement. It's a change of first principles, not of subsequent details... and for some it is even too much...

- Klaus
  Reply With Quote
Old 1st August 2008, 02:52 AM   #617
gedlee is offline gedlee  United States
diyAudio Member
 
Join Date: Dec 2004
Location: Novi, Michigan
Quote:
Originally posted by KSTR
Earl,

To my knowledge it is not possible to extract a center signal from a LR signal and getting a null output at the same time from L&R for that center signal with any "stupid" linear matrix (that is, not analysing the LR-content for correlation etc.).

Klaus

Thats what I said, I don't think its possible to do in analog either. And I don't think its possible without cross correlation either. But with cross correaltion in DSP it is certainly possible. I could write out the algorithm myself.
  Reply With Quote
Old 1st August 2008, 03:22 AM   #618
limono is offline limono  United States
diyAudio Member
 
Join Date: Mar 2007
How audible is matching the drivers and elements of crossover if I may ask .
I know that serious high-end manufacturers also A-Note which provides the kits is matching the drivers and elements of crossover claiming that it's crucial for quality of the speakers.
Is the kit "voiced" with particular parts (caps , coils , wire) in mind ?
Regards, L
  Reply With Quote
Old 1st August 2008, 03:29 AM   #619
gedlee is offline gedlee  United States
diyAudio Member
 
Join Date: Dec 2004
Location: Novi, Michigan
Crossover design and matching it to the specific drivers is essential, but "voicing" it through listening, well thats not my approach. I spend a lot of time on the crossover, but its the design that matters not the parts. Small things like the resistance of the inductors have to be taken into account and controlled, a 1/2 ohm or an ohm can make a big difference. But this is all measureable and quantifiable. I design the crosssover for a smooth flat power response not the response along any particular axis. But I do pay close attention to 22.5 because that tends to be on-axis when the speakers are toed in at 45 as I recommend. The axial response tends not to be so flat, but then that's not really the axial response if the speakers are setup right (45 toe in).
  Reply With Quote
Old 1st August 2008, 03:34 AM   #620
gedlee is offline gedlee  United States
diyAudio Member
 
Join Date: Dec 2004
Location: Novi, Michigan
Below is a pdf of the measured response of the Nathans as currently sold. While this is not a Summa, its also not $3500. Its is as good as anything that I've seen in this price range, better than most (but then very few people have the guts to actually publish real data like this).

OOPs sorry too big. Its on my web site at:
Nathan_10 horizontal response
  Reply With Quote

Reply


Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off



New To Site? Need Help?

All times are GMT. The time now is 03:22 AM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio

Content Relevant URLs by vBSEO 3.3.2