The further into the detail of my new project, Im finding that the video side is taking care of itself as I'm comfortable in that area. My audio chops need sharpening.
This brought up the issue of correction and EQ. Is there anyone out there considering the use of Lake processors? I would particularly like to hear from anyone experienced in their use, thinking of Mark Seaton in particular. Would it be possible to give us a Lake 101, as it may be of great interest to this forums readership. I would be very grateful for any time taken

The further into the detail of my new project, Im finding that the video side is taking care of itself as I'm comfortable in that area. My audio chops need sharpening.
This brought up the issue of correction and EQ. Is there anyone out there considering the use of Lake processors? I would particularly like to hear from anyone experienced in their use, thinking of Mark Seaton in particular. Would it be possible to give us a Lake 101, as it may be of great interest to this forums readership. I would be very grateful for any time taken

CM,
As you know I have the Dolby Lake. A few things that it has done for my room has been much improved dialogue intelligibility,better feel of integration of the subs and definitely has made the overall experience much improved from every seat. Recently it has even allowed us to notch out one specific frequency that was causing my projector to move all over the place with very liitle effect to the experience.

One big deal is the ability of the Lake to implement delays really adding to the surround experience being much more enveloping.

Of course Mark set it up and very critically messaged the FR so I hope he will get in here to describe what has been done in a more in depth fashion.

http://www.dolby.com/images/professional/live_sound/lake_processor_3.jpg

http://www.artsonneborn.com//assets/images/db_images/db_Latest_081.jpg
Art

The new DEQX HDP3 and even the NAD T175's version of Audyssey are worth of checking out as well. NAD's version of Audyssey is quite good, even if it doesn't provide the tools I'd like and DEQX, well, I don't think there's a speaker made that couldn't benefit *dramatically* from that.

Art, do you have a technology list of what Lake can do and how it does it? I know DEQX inside/out, but Lake's material, last I checked, was kinda cagey. And I've forgotten everything I think I know about it anyway.

To give you an idea - DEQX applies impulse response correction (FR, phase, time, and, i guess, some distortion correction) with 4096 points of resolution. Adjustable crossovers up to 300dB/octave and some parametric EQ for room correction. Unfortunately, it's a stereo piece. However, it will make a $5000/pr speaker sound better than a $50,000/pr speaker, so it's $4K-$6K price and somewhat DIY/hobbyist aspects are worth overcoming.

The new DEQX HDP3 and even the NAD T175's version of Audyssey are worth of checking out as well. NAD's version of Audyssey is quite good, even if it doesn't provide the tools I'd like and DEQX, well, I don't think there's a speaker made that couldn't benefit *dramatically* from that.

Art, do you have a technology list of what Lake can do and how it does it? I know DEQX inside/out, but Lake's material, last I checked, was kinda cagey. And I've forgotten everything I think I know about it anyway.

To give you an idea - DEQX applies impulse response correction (FR, phase, time, and, i guess, some distortion correction) with 4096 points of resolution. Adjustable crossovers up to 300dB/octave and some parametric EQ for room correction. Unfortunately, it's a stereo piece. However, it will make a $5000/pr speaker sound better than a $50,000/pr speaker, so it's $4K-$6K price and somewhat DIY/hobbyist aspects are worth overcoming.

Thanks for the reply. Tried the Audyssey in the last room.

I was, very specifically, after Lake experiences, I may have reason to consider others later, but not atm. Mark Seaton will be chiming in here soon. I think I will be needing the configurability of the Lake in 4x12 matrices as I will be exclusively 3 way. Looks like i will need 3.

I have some friends in the broadcast media and surround mastering and they ALL swear that the Lake smokes everything else. It does have some limitations but gets updated regularly. The mastering guys said its the only one that they would consider totally reliable in "mission critical" situations.

That's what intrigues me about it, but I just don't really know what it does. Seems possibly like a great add on machine for those willing to shell out $____(?)

Keep in mind that NAD's version of Audyssey sounds much better than even the Audyssey Pro you likely tried (it really does), though it still lacks flexibility for hands on tweaking.

Hi Rob,

John's response was exactly as expected. :rolleyes:

The Dolby Lake processors have a variety of uses. I guess the most important thing to start with is the fact that the various models and configurations are tools, not solutions. The most common configurations are either in 8x8 (input x output) or 4x12 form. Art's system uses the 8x8 configuration. This is used as system/channel EQ, where the EQ can manipulate the signal for each channel with various frequency shaping and delay options. The 4x12 configuration is intended to be used for crossover or other signal splitting duties, while still maintaining the ability to EQ any of the separte input or output channels individually or as a whole. Obviously a 4x12 configuration won't cover a full surround system, additional channel EQ or multiple units would be required. I have recently planned out and quoted a couple custom systems (very high power and/or larger spaces) using a pair of Dolby Lake Processors, both in 4x12 configuration. In the planned system, each main speaker was a 3 way active design, with the Lake handling all crossover duties using external amplification. In this configuration the Lake serves to both execute the custom loudspeaker crossover, as well as the individual channel EQ as is used in Art's system. It has unique crossover functions available, including a variety of linear phase filters as well as more conventional options. Obviously this 4x12 configuration is of no use to most speaker systems which have their crossovers self-contained in either an active or passive form.

Again, there is no microphone that attaches, nor is there any auto-setup button. The Dolby Lake Processor only does what the user tells it to. Many consider it to be one of, if not the best in terms of sound quality and transparency in its ADC and DAC, while its isolation features mean you won't be chasing ground loops. I would say the interface and the types of filtering available are what provide the most unique benefits. Control his handled over a common ethernet network or by direct connection to a PC via CAT5. Multiple units can easily be controlled simultaneously over the same network, even grouping functions between units. This has been common in pro audio for many years now, but the Lake interface does make this much easier.

I'm out of time for the moment, but those curious can check out the overview on Dolby's website (http://www.dolby.com/professional/live_sound/Products/lake_processor.html)for some info on the EQ functions and general capabilities. If anyone has ever poked through the short novel they call a manual, you can understand that even after spending much time on the EQ, delay, and routing side of things in Art's system, I'm still only just getting familiar with the more advanced control functions available. While I currently have too many other financial obligations to keep a ~$6,500 EQ around for experimentation, I plan to add one to my personal setup some time this year.

Oh yeah... The big blue meters are cool! :cool:

OK, I've looked and I can't find it anywhere: what are the processors (how many) in the Lake? Is it one chip, two, four?

The rest of the specs are very self-explanantory, but I can't seem to find that spec anywhere....help.

Dan

That's what intrigues me about it, but I just don't really know what it does. Seems possibly like a great add on machine for those willing to shell out $____(?)

That's the thing John, it's not what it does, but rather what you can do with it. <I tried to spare us the JFK parody> :o


Keep in mind that NAD's version of Audyssey sounds much better than even the Audyssey Pro you likely tried (it really does), though it still lacks flexibility for hands on tweaking.

Audyssey Pro is something I have to spend some more time looking into. I do suspect that for most installers this is a great tool that can better insure the mantra of do-no-harm with respect to EQ. At the same time I haven't yet heard an Audyssey setup I haven't been able to improve significantly with further adjustments, and the setup is far from fool-proof.

I would suggest that where room treatments get much more aggressive, rooms get larger, system/loudspeaker capability increases, and expectations get higher, a manually set-up system is still going to be the way to go. I do hope that eventually I won't have to visit installations to guarantee great sound, and there most certainly are ways to automate much of the process I go through in optimizing a system. Actually packaging all of that into a system will be no small feat, and likely quite a few years down the road.

AMEN!!!






Dan

OK, I've looked and I can't find it anywhere: what are the processors (how many) in the Lake? Is it one chip, two, four?

The rest of the specs are very self-explanantory, but I can't seem to find that spec anywhere....help.

Dan

Hi Dan,

I recall an article in Live Sound International which went through many of these questions a year or two back.

Ahh... a little poking found some articles all linked at the bottom of this page:
Live Sound Documentation (http://www.dolby.com/professional/live_sound/support_download/documentation.html)

Try the first article "Dolby Lake Processor: A New Level of Sound Management”

It talks about the abilities of the lake, talks about configurations of the lake, and what you can do inside with the various I/O cards- but never actually says anything like "2 64 Mhz DSP chips" or "4 128Mhz DSP chips" from______. I guess that's what I'm looking for.

I've heard the Lake in an install in Austin- it sounded good, I hear it took forever to setup (probably kind of like what Mark was alluding to with Art's sytem), but in the end the results were really nice!

Dan

Oh, here's why I asked: I was wondering how the Lake compares to this:

http://www.symnetaudio.com/index.php?Show=187&Show1=&Show2=188

Pro-guys?

Audyssey Pro is something I have to spend some more time looking into. I do suspect that for most installers this is a great tool that can better insure the mantra of do-no-harm with respect to EQ. At the same time I haven't yet heard an Audyssey setup I haven't been able to improve significantly with further adjustments, and the setup is far from fool-proof.

Audyssey Pro just isn't worth it until they do a major software overhaul. It is quite literally not as well tuned as the NAD implementation with full surround built in. It tightens up the room/speaker impulse response, but does so with a bright, thin sound, whereas the NAD setup sounds great. And you're right, it needs a lot of manual tweaking to volume, distance and crossover settings and that's the thing that worries me. i shouldn't be getting errors when all the speakers are identical. I've suggested a whole raft of improvements to them, but they seem somewhere between reluctant and obstinate about making changes.

I would suggest that where room treatments get much more aggressive, rooms get larger, system/loudspeaker capability increases, and expectations get higher, a manually set-up system is still going to be the way to go. I do hope that eventually I won't have to visit installations to guarantee great sound, and there most certainly are ways to automate much of the process I go through in optimizing a system. Actually packaging all of that into a system will be no small feat, and likely quite a few years down the road.

Audyssey is great when you won't treat the room and especially if you use inwalls. It is more advanced than the Lake system in that it's impulse response correction, except it's not nearly as flexible. i call it the 'trained monkey system'. Certainly not a true pro system as it assumes you're an idiot with a mic stand. BUT, the concept is powerful and potentially amazing in the future. The Audyssey Pro looks to be similar to the base 8x8 Lake processor, one being totally automated for chimps, the other being totally manual for experts.

As for the Lake info, i think I have it figured out. I wouldn't quite put it on par power-wise with the DEQX system for what it does and now that DEQX has fixed their sonic issues and made a higher fidelity piece, it may well equal the Lake in the transparency domain. But it would take 4 $6K DEQX units to do what two maxed out Lake 4x12 processors can do. So it depends if the Lake costs more than $12K each and how much work you're willing to do. DEQX still pretty well requires quasi to full anechoic measurements of the speakers to work but it can automatically fix details that the Lake would take tons of programming to do.

Interesting product though. What is the price range of these?

THe Lake is Pro gear. on that side of the house, ppl are (on the most part) far less concerned with the guts as opposed to what it puts out.

It is a big (sonically) DSP box with lots of cycles. you tune it, not the other way around. as such, the amount of folks who can actually use it are far and few between. especially in this world. Tzucc has one as well to manage all his "subbage"

Dizz, seems to me it's like the swiss army knife of processors. here's a review -

http://mixonline.com/gear/reviews/audio_dolby_lake_processor/

looks like at $7G, it's not a bad price at all for what it does. The DEQX seems to be more focused on two tasks whereas the Lake is less focused, but more flexible.

In the end, it seems like you would choose the product to the task and your own talents. I can make a DEQX sound good, even great, but I'm not sure I have the raw talent to get everything out of a Lake, nor that I'd want to develop it. It would be nice to see the Lake's flexibility and hardware with the DEQX's impulse response capabilities for sure.

Actually the question about processing power has nothing to do with caring about what's inside so much as it's about knowing where the limitations might be. With one configured in a 4x12 setup, how many high-frequency channels can you run before you start losing the speed necessary to sound good? The system I heard with the Lake(s) had either 3 or 4 in it, I really don't remember- so obviously IT wasn't running out of processing juice!

I was wondering because as systems grow to the level of Jeff's or Art's or heck- even Steve's, would you have to add more Lakes, or would the 8x8 and 4x12 have the juice to handle it?

Just curious, plus I was curious how it compared to the Symetrix piece.

Dan

when you take a product like a lake, and tie it to a product like Smaart or Pro Tools, you have an extremely powerful diagnostic, (Smaart) and then you tie it to a very powerful EQ on steroids, and add a very knowledgeable operator in the middle and the combo cannot be beat.

So like it talks about in the article, create a master setup file that flattens the room to the nth degree then allow user EQ's that compensate for all types of music, or just allow people to mess around.

THe power of a lake is not really enjoyed though until you get into a really big/complex system where its power it properly utilized. but when it is used to full potential, a deqx or tact or others like that do not compare, but as has been identified, they are different.

This is not the kind of knowledge you typically find in the CE world.

How does the Lake stack up against the BSS Soundweb product (SDEC Processor if you are a JBL Synth fan)?

Looks to me like they are similarly featured devices that have a lot to offer any high end home cinema.

Neil

THe power of a lake is not really enjoyed though until you get into a really big/complex system where its power it properly utilized. but when it is used to full potential, a deqx or tact or others like that do not compare, but as has been identified, they are different.

Depends on the app. Correcting impulse response as DEQX does is a far more sophisticated thing, even in the hands of pros. Lake is a bigger collection of less sophisticated tools, but studio guys want to adjust everything by ear to their tastes, so i understand why it is what it is, having worked around some of these guys. Making a room flat at console is might be desirable for doing a mix in essentially an anechoic chamber, but not necessarily in a home. Keep in mind that DEQX can fix hundreds of ridiculously tiny audible problems in a speaker that no EQ can. Everyone who has a high-end 3-way speaker system and a desire to have the best possible sound should get a DEQX and triamp their speakers. If you were to gut the crossovers on a set of Wilson X2s, for instance and replace with DEQX, the result would be in a completely different realm that what it can do stock. Completely different.

This is not the kind of knowledge you typically find in the CE world.

Though, to be fair, it's not necessary in the CE world. Studio and home are totally different environments, one for work, one for pleasure. One for finding, fixing flaws and producing a product, the other for faking reality and providing an immersive, natural sound. We just replaced 3 Rane EQs in a system that provided tons of control, but didn't really address the system issues like NAD's preamp did.

The Lake processor would be great for an active system in real theaters. If i had the money and the time, i'd get one just to play with. But I don't have that much of either. i honestly think CM should get one and spend a year becoming and expert in the system and post about what he's doing with it. For $7k, it's not a bad investment even if you can only get 50% of what it can do, assuming you're willing to put the time in. i'm not, but someone around here should ;)

The competing products to the Lake are products like the BSS Soundweb London, SymNet, Rane RPM, MediaMatrix variants, Bi-Amp AudiaFlex, etc., etc.

I have used the SymNet 8x8, and had a day here or there working with the Soundweb, Rane RPM, and MediaMatrix products. The SymNet products had been my pick for good value and tons of capability. Most all of these are now able to be configured over a network, some more easily than others. Interestingly, most of these other products are more flexible in signal routing, and how the signal is handled. They are open-architecture where you have inputs and outputs on your screen, and can drag 'n drop any processing block you like, and mix(combine) signals from and to at any point in the chain. These are basically very flexible EQs for rather reasonable prices ($3000-5000). The Rane RPM, Soundweb and SymNet products are the ones I have seen make it into home theater use the most. Do note that ALL of the above mentioned processors are packed into tiny cases, usually 1RU, and have very noisy fans. These can't go in a listening room. The only DSP units I've found having more than 2 channels and not using fans were the Shure P4800 (4x8), and the lowly Behringer DCX2496 (3x6). There are a few others with 2x6 or 2x2 capability, but these all have to be configured separately, making setup a real pain or less flexible.

The Dolby Lake Processor is a bit different in its goals and target use. Unfortunately those goals make it a little less flexible in architecture (signal routing). The Lake is very much targeted at live sound use. The signal processing they use allows equal or lower latency per processing power than most any other similarly capable DSP. For the peanut gallery, latency is basically the time it takes for throughput from input to output. With classic processors, the latency increases with every block of EQ, limiting, or other processing added to the chain. Others now have similar goals or modes you can lock them in, but the Lake is lower latency than most. In HT use and for any recorded playback system, latency is a non-issue, so this benefits us none, but it is the reason the Lake has a more fixed architecture than some others. Over the past few years they have continued to "open up" more configuration options through firmware and software updates, so I am hopeful they will continue to allow more options.

On the matter of processing power that Dan asked about, I think this is where our audiophile egos get a bit ahead of us. Quite frankly, our systems are very tiny compared to most pro audio installations and traveling setups. The pro audio speakers and line-sources these are designed to accomodate require much heavier processing than we are likely to ever need. For those interested, page 15 of the Dolby Lake System Manual has a good bit of the specs, which are equal or better than most any other DSP I've seen. The internal sample rate is 96kHz and data path is 32 bit.

A few things do really set the Lake appart in operation and performance from those listed above. While not really a performance issue, most consumers won't be too cozy about the phoenix style connectors, where the Lake has all XLR connectors. The Lake also has their IsoFloat technology which you can read about on the website, but basically insures you won't have to worry about ground loops; the most required is a click of the mouse. :) Those features are nice, but what really sets the Lake appart are the signal processing functions of EQ and crossovers, and the GUI that is used to adjust them.

The lake has unique filter types, and I would argue a much better way to adjust them graphically. The most unique of these is what they appropriately call the MesaEQ. This filter is enabled by the way they handle parametric EQ. They call this Raised Cosine Equalization (http://www.dolby.com/professional/live_sound/live_sound_technologies/). Click that link and be sure to read the white paper (http://www.dolby.com/assets/pdf/professional/live_sound/DLP_RaisedCosineEQWhitePaper.pdf) which gives a very good explanation of the benefits. Long story short, it is much easier to get the EQ to do what you expect, and more importantly, it is much easier to generate filters which match the acoustic problems we encounter in the real world.

As I told Art after working with the Lake in his system, I found myself actually taking the time to address problems I otherwise might not have as it would have taken at least 3 interacting filters to get a comparable result, and adjusting such a hodge-podge of filters is a very time intensive PITA. Here we can more directly set the filter to do exactly what is needed. The other products mentioned are very flexible with many great capabilities, but they are still digital executions of classic filtering. It's expensive in terms of EQ, but far from outrageous in the world of hi-end electronics.

While I agree that not everyone will know what to do with a good EQ, I would dissagree with Dizzman that it takes immense knowledge to get beneficial results. Many are currently using EQ's like the Soundweb, SymNet and Rane products, and I would argue that any of them worth hiring could do a much better job with a Lake once they took the time to read the manual and learned how operate it. That isn't to say that any random guy with an RTA will do the job I do, but that's the case with any capable product, just as you would expect Ken Whitcomb to be able to get more from a top projector than the random guy who just got his ISF certification.

and now back to work...

Keep in mind that many, if not most really great studio guys that will use this tool can hear problems and tell you at exactly what frequency they are occurring without needing to measure. They're really hands on guys that don't need mics and computers for the most part to tell them what to fix and how. i've had some of them by the shop at times and they're like 'this speaker has a resonance at 450hz' or 'your room has a 63hz mode that you need to fix' and i'm saying, 'yeah, you're right, it does' and wishing I was that adept at interpreting what my ears are hearing. i'm pretty good, but some of these guys are truly golden-eared, though probably most would tell you that a lot of this audiophile crap doesn't much matter in the scheme of things because even they can't hear that stuff. Audiophiles are great at ignoring obvious flaws while paying attention to microscopic or even non-existent flaws. Studio guys are the opposite, or, often, both.

It's frustrating for me having all the so called audiophiles completely disinterested in something like DEQX or Audyssey, let alone something like the Lake, but highly interested in going from a $5k amp to a $10k amp to a $20k amp to solve what is obviously a speaker and/or room problem. I only ever did sell 1 DEQX locally, the rest were all out of state DIY guys, so i largely gave up on it, even though I love the product (I did sell 3 dozen NHT systems with the technology built in though). Couldn't sell Audyssey Pro either, though I have to say, I didn't like the result either, so my heart wasn't in it. But everyone with the processing built in can't wait to hear the result, so the other option is wait for the technology to trickle into a consumer piece with more automated and/or easier to use controls. I wouldn't be surprised if Dolby doesn't make some sort of "Lake for idiots" processing available for surround preamps.

John,

There's no reason a home audio professional shouldn't strive to develop the listening and frequency identification capabilities most professionals do. Of course throw someone in front of a high power stage monitor and force feedback into the system and it's amazing how powerful a learning motivator/accellorater it is to have a 3kHz ring drilling you in the forehead. :rolleyes:

I see you are still hung up on the concept of impulse response correction. I thought the DEQX didn't directly do this, but I could be wrong. There are plenty of share-ware computer programs which let you experiment with this. I do think this has application at the loudspeaker level, but it's highly problematic to attempt at a single listening position, let alone a wide seating area. Most of the tools available use all sorts of tricks to keep the correction from being too agressive, as it will end up causing more problems than it solves, or makes for a listening window that requires a cradle for your head.

I like many things about the Audyssey solution, and DEQX is a cool concept & tool that they never scaled up to multi-channel. Contrary to your endless posting on the matter, there are much more elegant, and better performing means to design loudspeakers covering a listening area than brick wall FIR filters. Sometimes we need that overlap and behavior, but there certainly are good uses for linear phase filters, and they can even be had with shallower slopes. The use of a 4x12 configuration of the Lake is really requires a loudspeaker designer to make sure you aren't destroying what the speakers were designed to do, although so many home speaker designs are ignorant of 3D response it's easily overlooked by many. This brings us back to the Audyssey system. While I have sometimes observed extremely good results in voice matching of speakers, the more problems that exist seat to seat, the less beneficial the results tend to be.

The ultimate problem comes down to recognizing when a seat or area is bad, and no amount of EQ can do anything about it without significantly compromising performance elsewhere in the room. This can only be addressed through acoustic treatment, loudspeaker/subwoofer placement, and loudspeaker design. Sometimes the loudspeaker design part can be addressed in adjustments of an active crossover, but I hold those functions separate from channel EQ in the system, with an occasional exception.

Most I have talked with in the pro, studio, and ultra high end HT market find the Dolby Lake, the Soundweb London, and to a lesser extent the SymNet processors to be in higher regard for sound quality, although mostly this comes down to the s/n of the system and the quality of the ADC/DACs used. Personally I wouldn't hesitate to put a Lake in the highest quality of systems.

There's no reason a home audio professional shouldn't strive to develop the listening and frequency identification capabilities most professionals do. Of course throw someone in front of a high power stage monitor and force feedback into the system and it's amazing how powerful a learning motivator/accellorater it is to have a 3kHz ring drilling you in the forehead. :rolleyes:

Well, I certainly do, but i went from being 'tin-earred' to at least bronze or silver eared. But I haven't the pitch perfect capabilities that some of these guys do, nor do many 'serious audiophiles'. i am good enough to find flaws in speakers and report them to the designer or hear obvious room issues and that's good enough for me, i suppose, i'll rely on more advanced electronics for the rest ;)

I see you are still hung up on the concept of impulse response correction. I thought the DEQX didn't directly do this, but I could be wrong. There are plenty of share-ware computer programs which let you experiment with this. I do think this has application at the loudspeaker level, but it's highly problematic to attempt at a single listening position, let alone a wide seating area. Most of the tools available use all sorts of tricks to keep the correction from being too agressive, as it will end up causing more problems than it solves, or makes for a listening window that requires a cradle for your head.

Speaker impulse response, not room impulse, though they have been working on that. Audyssey does room impulse. Together, they can be quite formidable. I think you're hung up a bit on DEQX *necessarily* being brick wall filters. That's not the case. You can do anything from 24dB/octave and up, last I checked, so you choose the best crossover, not the steepest one and with time/phase correction, there's no need for less than 24dB. Plus, the FR/time correction and ability to perfectly blend the drivers without months and months of fiddling is the key, let alone simply pulling out all the capacitors, resistors, inductors.

The new HDP3 should satisfy the more demanding audiophiles for lower noise electronics too. I highly suggest playing more with one. When you tri-amp a speaker, any speaker (though preferably pistonic monopoles with low diffraction cabinets and proportional drivers), with DEQX you get an instant and amazing improvement in clarity, dispersion, coherency, transparency. It's not being hung up for me, I've used it a lot on different speakers and I've put up Xd against a lot of $10K-$20K systems. It's really that good. Why virtually everyone on this forum doesn't have one is a head scratcher to me. Keep in mind, I don't even sell them any more so it's not trying to attract business. Nor am I dissuading anyone from buying a Lake or Audyssey or anything else. All of these tools can be useful.

John,

I'm already designing and selling 3-way, active loudspeakers with digital crossovers, so no need to sell me on the concept. ;)

I have played with a DEQX a few years back and found it quite useful and powerful as an active crossover. The limitation of 2ch/box was what was always a stopper for my applications. IMO, a loudspeaker has to be designed with this in mind, or the designer of the speaker should be the one implementing the settings, but that's also why I like fully active systems with the amplifiers accounted for in the design phase of the speaker, not after. Jim Salk has a speaker he offers with the option of the DEQX crossover instead of the conventional passive option, and the designer of the crossover set up the system to maintain the strengths he was after with the design. This is where this sort of system is beneficial, and much less so to the majority of enthusiasts trying to mix it with a random speaker.

In the pro market EAW is doing some nice work with some of their new products using FIR filtering to greatly improve the behavior of their compression drivers and horns. It actually makes them sound comparable to other very well designed speakers. :p My point is that FIR filtering isn't a requirement to get good sound, especially with problems often encountered below 500Hz in-room, which tend to approximate the same minimum-phase behavior of the many forms of non-FIR based filtering. For most of the problems we can actually fix, the advantage of FIR/zero-phase filtering is not as big as many make it out to be.

True, true, though i honestly haven't heard a speaker that wouldn't benefit from DEQX substantially. The Revel Studio2s just pulled into town today (4th order, wide dispersion, superbly flat response), so maybe these will be the first :)

i'm not saying that you can't build a great speaker without it, but there aren't very many great speakers out there and you can take a generations old $20K+ speaker and have it outperform the current version quite dramatically, as i did with some older Thiels. So, DEQX and a 6-channel amp easily trumps a passive crossover with $100K monoblocks in all but a very few anomalous circumstances. Although, I guess you'd have to *want* your speakers to be as transparent and accurate as possible and clearly that isn't the case with everyone.

Are there any other end users other than me yet ?

Art

What kind of signal is being fed to the Lake?

If analog from the pre/pro I guess you are doing another A/D and D/A.

Digital from a Casablanca (AES fronts), Meridian (S/PDIF, 48/24), or Tact (either) would be very cool, but only the Tact can accept HDMI.

Are there any other end users other than me yet ?

Art, you know you're unique :)

I've been interested, but how much does it help speakers with passive crossovers ie. MBL 101e's. Also, without HDMI inputs from a source, how can it deal with the new audio codecs, for HT. Obviously for music that's not important, at this point. I have a TACT 2.2 xp that I plan to try out when my music server is ready, in a week or so.
When they get "Lake for dummies", I'll be all over it.

David

Contrary to your endless posting on the matter, there are much more elegant, and better performing means to design loudspeakers covering a listening area than brick wall FIR filters. Sometimes we need that overlap and behavior, but there certainly are good uses for linear phase filters, and they can even be had with shallower slopes.
Mark,

I heartily agree with the above. The sharp "brickwall" filters make the designer's
job easier - since it "uncouples" the bandpasses involved - a single driver for a single
bandpass.

However, the discontinuity in the filter will give rise to a discontinuity in the angular
response. At slightly less than the crossover frequency, you have the larger driver
reproducing that frequency, and at slightly greater than the crossover frequency you
have the smaller driver reproducing that frequency.

You can't beat the physics that the angular radiation pattern of the larger driver is
going to be different than the angular radiation pattern of the smaller driver - that's
just simple "antenna theory". The larger driver is closer to its "beaming" limit than
the smaller driver.

So in transitioning from slightly below the crossover frequency to slightly above the
crossover frequency - you have a sudden shift in radiation pattern - because you had
a sudden shift in which driver was being used.

It's much better to make that a smoother process. Discontinuities - like sharp brickwall
filters should be avoided to the maximal degree possible. Nature doesn't divide the
frequency space into bandpasses as we have to do. We should not call attention to
the fact that we have to do that by having crossovers that are discontinuous and
sharp instead of being smooth. It would be much better to blend the transition.

It's analogous to why a Class A/B amp is better than a pure Class B. The Class B
discontinuously hands off from the positive half-cycle amp chain to the negative
half-cycle amp chain at an instant in time - the zero crossing. If that isn't done perfectly,
i.e. unless the active elements be they tubes or transistors; are perfectly matched;
you get a crossing distortion.

The Class A/B amp does this hand-over gradually; not at an instant in time. Any
distortion due to components that are not perfectly matched [ and they never are ];
is mitigated by a smoother transition.

However, the discontinuity in the filter will give rise to a discontinuity in the angular
response. At slightly less than the crossover frequency, you have the larger driver
reproducing that frequency, and at slightly greater than the crossover frequency you
have the smaller driver reproducing that frequency.

Greg, let's not get back into this. The beauty of DEQX is that you can choose relatively mild 24dB/octave which is a long way from the 300dB/octave. Just because DEQX can do it, doesn't mean you have to as a designer. It's a tool, nothing more, it gives the capability *in case* you need it.

You can't beat the physics that the angular radiation pattern of the larger driver is
going to be different than the angular radiation pattern of the smaller driver - that's
just simple "antenna theory". The larger driver is closer to its "beaming" limit than
the smaller driver.

That's why you select the proper crossover frequency and proportionally sized drovers, so the dispersion matches up.

So in transitioning from slightly below the crossover frequency to slightly above the
crossover frequency - you have a sudden shift in radiation pattern - because you had
a sudden shift in which driver was being used.

Not unless you're a really bad designer and pick an unnecessarily high crossover frequency.

It's much better to make that a smoother process. Discontinuities - like sharp brickwall
filters should be avoided to the maximal degree possible. Nature doesn't divide the
frequency space into bandpasses as we have to do. We should not call attention to
the fact that we have to do that by having crossovers that are discontinuous and
sharp instead of being smooth. It would be much better to blend the transition.

Over what slope? We know that low order designs create far greater cone resonances, motor distortions, acoustical distortions, dynamic range issues, dispersion problems and other things which *also* don't happen in nature. Pick your poison. Five problems? Or one? Or choose a blend that minimizes all of them below noticeability, such as 24-48dB/octave? DEQX allows you to do whatever you want. No one is holding a gun to anyone's head.

I'll tell you one thing, when you have a speaker that sounds as coherent 6" in front of the speaker or anywhere else for that matter as it does at the listening position, it's a sweet thing. A violin doesn't start sounding like crap the moment you stand up or move three seats over. Having "created" my own speakers using DEQX and choosing from any variety of slopes, I can tell you that the ability to create steep slopes is a capability, not a benefit, and it's a relatively minor part of what the system does. it's how it blends the driver seamlessly *even with* steep slopes that makes it such an elegant system, by time aligning and idealizing the performance of the drivers before integrating the sound. I've heard plenty of low-order speakers that sound far less seamless than a high-order DEQXed speaker.

Art, you know you're unique :)

Yea ,but that has never counted for points.:o

Art

Tzucc has one.

Mark, my comments were more directed along the lines of "there are not as many folks in CE that take this stuff really seriously as guys in the Pro AV world" When was the last time a CE guy was at Pat's classes? And i know from ten years of experience at Extron that when we started talking about test gear, REAL test gear, there were not that many folks that knew what i was talking about, when i talked about it at CEDIA classes... crickets!

It is a tool, a very powerful one for a knowledgable sound designer.

My Bad, i just recalled that Tzucc has the Lake Contour.

http://www.dolby.com/professional/live_sound/products/lake_contour.html
What is cool about this stuff is that for a dynamic environment, you can slave the unit to something like Smaart, and then determine your end result, but have smaart control it real time. the possibilities are endless. I recall seeing the Sim system (different SIM) years ago and what they were doing was amazing. Even as the room (65000) filled up, there were no major changes in the sound. pretty awesome.

Home theater could turn this all around. Audiophilia practically killed common sense, but home theater has brought it back as it's brought people that care more about the results than the gear.

Hi Mark,

Enjoy reading your comments here and at AV123.

Upcoming Sherwood Newcastle R-972 will have integrated Trinnov Optimizer. (I have no connection to either company). Just wondering if you are familiar with Trinnov and if so, would comment on it?

Best Rgds

The competing products to the Lake are products like the BSS Soundweb London, SymNet, Rane RPM, MediaMatrix variants, Bi-Amp AudiaFlex, etc., etc.

I have used the SymNet 8x8, and had a day here or there working with the Soundweb, Rane RPM, and MediaMatrix products. The SymNet products had been my pick for good value and tons of capability. Most all of these are now able to be configured over a network, some more easily than others. Interestingly, most of these other products are more flexible in signal routing, and how the signal is handled. They are open-architecture where you have inputs and outputs on your screen, and can drag 'n drop any processing block you like, and mix(combine) signals from and to at any point in the chain. These are basically very flexible EQs for rather reasonable prices ($3000-5000). The Rane RPM, Soundweb and SymNet products are the ones I have seen make it into home theater use the most. Do note that ALL of the above mentioned processors are packed into tiny cases, usually 1RU, and have very noisy fans. These can't go in a listening room. The only DSP units I've found having more than 2 channels and not using fans were the Shure P4800 (4x8), and the lowly Behringer DCX2496 (3x6). There are a few others with 2x6 or 2x2 capability, but these all have to be configured separately, making setup a real pain or less flexible.

The Dolby Lake Processor is a bit different in its goals and target use. Unfortunately those goals make it a little less flexible in architecture (signal routing). The Lake is very much targeted at live sound use. The signal processing they use allows equal or lower latency per processing power than most any other similarly capable DSP. For the peanut gallery, latency is basically the time it takes for throughput from input to output. With classic processors, the latency increases with every block of EQ, limiting, or other processing added to the chain. Others now have similar goals or modes you can lock them in, but the Lake is lower latency than most. In HT use and for any recorded playback system, latency is a non-issue, so this benefits us none, but it is the reason the Lake has a more fixed architecture than some others. Over the past few years they have continued to "open up" more configuration options through firmware and software updates, so I am hopeful they will continue to allow more options.

On the matter of processing power that Dan asked about, I think this is where our audiophile egos get a bit ahead of us. Quite frankly, our systems are very tiny compared to most pro audio installations and traveling setups. The pro audio speakers and line-sources these are designed to accomodate require much heavier processing than we are likely to ever need. For those interested, page 15 of the Dolby Lake System Manual has a good bit of the specs, which are equal or better than most any other DSP I've seen. The internal sample rate is 96kHz and data path is 32 bit.

A few things do really set the Lake appart in operation and performance from those listed above. While not really a performance issue, most consumers won't be too cozy about the phoenix style connectors, where the Lake has all XLR connectors. The Lake also has their IsoFloat technology which you can read about on the website, but basically insures you won't have to worry about ground loops; the most required is a click of the mouse. :) Those features are nice, but what really sets the Lake appart are the signal processing functions of EQ and crossovers, and the GUI that is used to adjust them.

The lake has unique filter types, and I would argue a much better way to adjust them graphically. The most unique of these is what they appropriately call the MesaEQ. This filter is enabled by the way they handle parametric EQ. They call this Raised Cosine Equalization (http://www.dolby.com/professional/live_sound/live_sound_technologies/). Click that link and be sure to read the white paper (http://www.dolby.com/assets/pdf/professional/live_sound/DLP_RaisedCosineEQWhitePaper.pdf) which gives a very good explanation of the benefits. Long story short, it is much easier to get the EQ to do what you expect, and more importantly, it is much easier to generate filters which match the acoustic problems we encounter in the real world.

As I told Art after working with the Lake in his system, I found myself actually taking the time to address problems I otherwise might not have as it would have taken at least 3 interacting filters to get a comparable result, and adjusting such a hodge-podge of filters is a very time intensive PITA. Here we can more directly set the filter to do exactly what is needed. The other products mentioned are very flexible with many great capabilities, but they are still digital executions of classic filtering. It's expensive in terms of EQ, but far from outrageous in the world of hi-end electronics.

While I agree that not everyone will know what to do with a good EQ, I would dissagree with Dizzman that it takes immense knowledge to get beneficial results. Many are currently using EQ's like the Soundweb, SymNet and Rane products, and I would argue that any of them worth hiring could do a much better job with a Lake once they took the time to read the manual and learned how operate it. That isn't to say that any random guy with an RTA will do the job I do, but that's the case with any capable product, just as you would expect Ken Whitcomb to be able to get more from a top projector than the random guy who just got his ISF certification.

and now back to work...

Hi Mark,

Enjoy reading your comments here and at AV123.
Just wanted to chime in on that...Mark's post are always a pleasure to read and learn from.

That's why you select the proper crossover frequency and proportionally sized drovers, so the dispersion matches up.
John,

As long as the drivers are different sizes - they will NEVER "match up" in terms of
radiation pattern. That's just physics!!!

John & Morbius,

As usual, most statements of absolutes regarding loudspeaker design are going to be wrong in various situations or for particular applications. The DEQX product is actually quite useful if you want to design a speaker around it, or effectively re-design the crossover function and voicing of an existing loudspeaker with it. I'm not really interested in getting into a discussion of loudspeaker design philosphies in detail, as my priorities will differ greatly from others, and frankly I've done enough teaching of the competition. :rolleyes:

Bringing this back to the topic of the Lake Processor, if the Lake is chosen to also provide crossover functions as well as more traditional channel EQ, it is a highly flexible unit in this regard, typically allowing generation of the desired response both more directly and more quickly. How a designer decides to use this flexibility will differ with application, goals for the loudspeaker, and the components chosen.

John,

As long as the drivers are different sizes - they will NEVER "match up" in terms of
radiation pattern. That's just physics!!!

Greg, 'matching up' is somewhat relative, but there are a dozen variables going on aside from this. Designing a speaker is somewhat like herding cats. You chase one and all the others go off on tangents. You have to think in 3-dimensions. What DEQX does is solves a lot of the issues so you really only have to think i in one or two dimensions. i would think that you'd love DEQX for this reason alone ;)

Mark doesn't worship DEQX as i do because clearly he's on another level in his experience and doesn't *need* DEQX. I certainly couldn't create or improve on a good speaker without it. He could.

Greg, 'matching up' is somewhat relative, but there are a dozen variables going on aside from this. Designing a speaker is somewhat like herding cats. You chase one and all the others go off on tangents. You have to think in 3-dimensions. What DEQX does is solves a lot of the issues so you really only have to think i in one or two dimensions. i would think that you'd love DEQX for this reason alone ;)

John,

It is precise BECAUSE I'm thinking in 3-dimensions that I say they won't match up!!!

I AM talking about the 3-D radiation pattern - and at a given frequency; the 3-D
radiation pattern of a 5" driver is going to be different than the radiation pattern
of a 1" driver.

That's the problem with the sharp crossover at frequency fc. At a frequency of
fc - 1 Hz; the 5 inch driver is reproducing that frequency with nothing from the
1" driver. Therefore the 3-D radiation pattern is that of the 5" driver.

At a frequency only 2 Hz higher - fc + 1Hz; the 1" driver is reproducing with nothing
from the 5" driver. So the 3-D radiation pattern looks like that of the 1" driver.

Mathematically, there is a discontinuity in the 3-D radiation pattern. Do I have to
expand this out in Spherical Harmonics for you? The 3-D radiation pattern is
discontinuous - and Nature doesn't do that!!

I've been interested, but how much does it help speakers with passive crossovers ie. MBL 101e's. Also, without HDMI inputs from a source, how can it deal with the new audio codecs, for HT. Obviously for music that's not important, at this point. I have a TACT 2.2 xp that I plan to try out when my music server is ready, in a week or so.
When they get "Lake for dummies", I'll be all over it.

David


Hi David,

The much more common useage of the Lake in home audio systems is as a multi-channel EQ, just as I used it in Art Sonneborn's system. It is a very powerful crossover if configured for that, but coldmachine is intending a rather different scale system than most, where even the largest Genelecs are not well suited (more so coverage than output). Art's main speakers don't really function any differently in connection/use than passive speakers like your MBLs. The output from the surround processor goes to the Lake, the modified signal outputs from the Lake to the amplifiers. In his case the amplifiers are just internal to the speakers and have active crossovers. The internal operation of the speaker is different with its own benefits, but the basic function of the whole is no different.

Using the lake in a conventional system, I would start with no EQ and measure the response of the speakers at various locations across the desired listening area, as well as near field after a quick initial listening. We are looking to compensate problems that are more common between the listening positions. The most significant work to be done and benefit to be had is usually below 500Hz, with more gradual filters shaping the response at higher frequencies with an occasional notch or two used to attenuate an offensive frequency range. Understanding the physics of how a loudspeaker operates and radiates sound in all directions, as well as understanding the performance limitations is where the type of knowledge I have allows some less obvious settings to be made to the benefit of the listeners.

I wouldn't hold by breath on a "Lake for dummies" product, unless someone else steps up to license the technology, as a few have in digital mixing consoles and amplifiers as with Lab Gruppen's PLM (for those who want 4ch of Lake and 10kW in 2RU spaces :D). Others with more close contact with Dolby have relayed to me that they have no interest in directly supporting the home audio market with this product, but of course we are welcome to use them. I don't blame them at all, but it is a shame. Who knows, I still have my fingers crossed that some day we can have a digital processing loop available in surround pre-amps which don't require all the extra conversions (the Lake and most other processors mentioned here can handle multiples channels of S/PDIF, digital I/O). Many in the pro audio world are already operating fully digital from mic input to speaker terminals, I suspect it is only a matter of time for home audio to get there. One can hope...

John,

It is precise BECAUSE I'm thinking in 3-dimensions that I say they won't match up!!!

I AM talking about the 3-D radiation pattern - and at a given frequency; the 3-D
radiation pattern of a 5" driver is going to be different than the radiation pattern
of a 1" driver.

That's the problem with the sharp crossover at frequency fc. At a frequency of
fc - 1 Hz; the 5 inch driver is reproducing that frequency with nothing from the
1" driver. Therefore the 3-D radiation pattern is that of the 5" driver.

At a frequency only 2 Hz higher - fc + 1Hz; the 1" driver is reproducing with nothing
from the 5" driver. So the 3-D radiation pattern looks like that of the 1" driver.

Mathematically, there is a discontinuity in the 3-D radiation pattern. Do I have to
expand this out in Spherical Harmonics for you? The 3-D radiation pattern is
discontinuous - and Nature doesn't do that!!

Okay, i'm going to leave this alone out of respect for everyone else on AVS. There's so much wrong here and Greg needs to learn so much about the realities of speaker design that it would cause a melee of LOTR proportions. Besides, i meant actually *think* in 3-dimensions, not think *about* 3-dimensions since you just showed how you think in theory, not in reality. Go ahead, say something else boisterous and authoritative, it will be wrong and I'm going to say absolutely nothing in response.

O There's so much wrong here and Greg needs to learn so much about the realities of speaker design that it would cause a melee of LOTR proportions.

John,

A speaker is an "antenna". There's physics that applies not just to speakers but to
ANYTHING that radiates waves in 3-D. That's a physics that I have studied; it
applies to speakers and applies to other antennas.


Besides, i meant actually *think* in 3-dimensions, not think *about* 3-dimensions since you just showed how you think in theory, not in reality.

You're going to have to explain what you mean by that one! ????

There's a difference in thinking "in" 3-D as opposed to thinking "about" 3-D?


Go ahead, say something else boisterous and authoritative, it will be wrong and I'm going to say absolutely nothing in response.

NO - I'll say something that is perfectly correct - that you will "think" is wrong.

Greg, 'matching up' is somewhat relative, but there are a dozen variables going on aside from this. Designing a speaker is somewhat like herding cats. You chase one and all the others go off on tangents. You have to think in 3-dimensions. What DEQX does is solves a lot of the issues so you really only have to think i in one or two dimensions. i would think that you'd love DEQX for this reason alone ;)

Mark doesn't worship DEQX as i do because clearly he's on another level in his experience and doesn't *need* DEQX. I certainly couldn't create or improve on a good speaker without it. He could.

John,

It is precise BECAUSE I'm thinking in 3-dimensions that I say they won't match up!!!

I AM talking about the 3-D radiation pattern - and at a given frequency; the 3-D
radiation pattern of a 5" driver is going to be different than the radiation pattern
of a 1" driver.

That's the problem with the sharp crossover at frequency fc. At a frequency of
fc - 1 Hz; the 5 inch driver is reproducing that frequency with nothing from the
1" driver. Therefore the 3-D radiation pattern is that of the 5" driver.

At a frequency only 2 Hz higher - fc + 1Hz; the 1" driver is reproducing with nothing
from the 5" driver. So the 3-D radiation pattern looks like that of the 1" driver.

Mathematically, there is a discontinuity in the 3-D radiation pattern. Do I have to
expand this out in Spherical Harmonics for you? The 3-D radiation pattern is
discontinuous - and Nature doesn't do that!!


Ladies please... :rolleyes:

The DEQX system isn't a bad option toward minimizing the problems inherently designed into many speakers with its own set of trade offs, some of which Morbius has highlighted. At the same time, pushing a crossover an octave or so lower can easily have two drivers behaving very similar in 3D radiation (mostly dependent on baffle size at that point). A simple example would be a 5" midrange crossed to a 15" woofer at say 800Hz vs. 200Hz. Both the radiation of the drivers around crossover and the required spacing between drivers will make for problems with an 800Hz crossover that won't be a problem when crossed 2 octaves lower.

Arguements can be made both for and against steep vs. shallower crossovers. Both are compromises, and your own prioritization of acoustic behavior will leave you leaning toward one solution or the other... at least until you realize it's better to not create the problem in the first place rather than waste time on a better band-aid. :p

Mark,

Keep in mind that I've stated several times that I don't think that steeper is automatically better, but having the *ability* to choose the best crossover for the situation, *including* amazingly steep ones, is better. Greg makes it seem like all I care about is steep crossovers, but that is only one trick in the DEQX arsenal, as you know. i consider 4th order to be plenty steep for most applications.

Besides, i think you know what I mean about herding cats and thinking in 3 dimensions. DEQX is like having a herd dog for those of us that aren't good at herding cats ;)

Besides, i think you know what I mean about herding cats and thinking in 3 dimensions. DEQX is like having a herd dog for those of us that aren't good at herding cats
John,

I DO think in 3-D; I have to - the devices that I write software to model are
3-D devices. ALL my physics and software is done in 3-D. I have no idea what
your distintion between thinking "in" 3-D vs. thinking "about" 3-D means.

One can herd cats - if one knows how.

Remember the 2000 Super Bowl advertisement by EDS?

"Anyone can herd cattle; but when you drive 10,000 half-wild shorthairs...."

Mark, could the Lake perform as a preamp (input source switching for three sources) as well as a crossover/PEQ/

Could it also function as a DAC?

Comments on the DEQX "debate" ...

I've used a heavily modified DEQX for several months now. It seriously trounced well designed high end passive crossovers in two different systems (Salk HT2s MTMs and Selah line arrays)

That being said, it took a lot of experimentation to get those crossovers dialed in. A nice feature is that in a few hours time you can A/B dozens of alternative crossover configurations and let your ears be the judge.

While I understand the mathematics, I trust the end results more: it may not work in theory, but it works in practice.

Message deleted.


John, I'd suggest just deleting this post (as I will with this quote). Greg's understanding of antennas is quite relevant to loudspeakers, although this tangent is quite off-topc from the discussion.

At some point I can only hope you will keep to debating issues in such threads rather than repeatedly calling to question the intelligence, logic or sanity of someone who has a different position, priorities or understanding than yourself. It may serve to make you feel better for the moment, and manage to spur on more heated posting, but it's never going to communicate your idea or change anyone else's mind on any matter. So are we done with the thread crapping yet? :(

Sure, I just get frustrated with tangential argumentation style which some people here perpetually use to try to get around valid points. Antenna theory is fine theory, but has little to do with the actual discussion of flexible technology.

Mark, could the Lake perform as a preamp (input source switching for three sources) as well as a crossover/PEQ/

Could it also function as a DAC?

Hi Bruce,

It could do this to some extent, but the limit is the number of inputs vs outputs and the means of control. If you only wanted to use it for 2ch output, yes, it could do what you describe, although I'm not certain about the ramifications for volume adjustment, and you would need an appropriate connected interface to do so (PC most likely). I'm actually looking into something not too different from this, but it's pretty far down the to-do list.

It can directly take S/PDIF or AES digital input, and could easily be configured as 12 inputs and 4 outputs, which could serve as 6 stereo inputs + 1 or 2 digital inputs with pair of stereo or bi-amp outputs. The control would require someone rather familiar with the Lake and more common control methods of such processors.

Mark- would I run this thru my Meridian 861 and is there a list of people in different areas of thec ountry that are capable of setting up the lake system?- I want to get the best from my system but obviously need someone local to set it up-gary

How does this device differ from the Tact Audio Theater Correction System or the Lyngdorf system?

Mathematically, there is a discontinuity in the 3-D radiation pattern. Do I have to
expand this out in Spherical Harmonics for you? The 3-D radiation pattern is
discontinuous - and Nature doesn't do that!!

Here's the opinion of a very well respected speaker designer; the legendary
Leo Spiegel, the designer of the original Apogee speakers:

http://www.apogeespeakers.com/leo_spiegel_interview.htm

The electrical crossover point is about 350Hz, but the effective acoustic
crossover point is around 600 to 700Hz as the electrical network is only one
element of the transfer function and the characteristics of the ribbon have
an effect here too. The blending of the woofer into the midrange/treble ribbon
is very gentle. Every time we try a very sharp cutoff, you can hear the
sound of the two units as two quite separate entities. It has been the same
with everything we have worked with. You can always say 'here's the woofer'
and 'here's the mid/treble tweeter'. But gentle crossover slopes demand good
overlap in the frequency response, or the technique doesn't work, the units
must be well behaved if such slopes are to work successfully".

Not doing this with you Greg. You want to focus on one small aspect of speaker design and it's not that simple.

What kind of signal is being fed to the Lake?

If analog from the pre/pro I guess you are doing another A/D and D/A.

Digital from a Casablanca (AES fronts), Meridian (S/PDIF, 48/24), or Tact (either) would be very cool, but only the Tact can accept HDMI.

That's my only issue, doing it analog. I would like to see some EQ that can do the whole thing via HDMI in, EQ and then HDMI out to the processor to avoid converting to analog. High quality analog EQ can be very transparent. Still, I'd prefer the whole thing be done in the digital domain esp. at this price point. The reality is that you have to work with what is availible. The only current solutions for doing the whole thing in the digital realm is to use on-board processing like availible in Tact, Lexicon, Meridian processors and soon to be Mcintosh.

bebop86 -- on using the unit with an 861 (what I too am contemplating) ...

The only real option is to put the Lake after the 861, feeding it with the 861s analog outputs. The reason for that is two fold: (1) the digital speaker outputs on the 861 are full volume (because they are designed to be used with Ms digital speakers which apply their volume sync after the DAC), and, (2) the Lake doesn't have Dolby (etc.) surround processing, so all your multichannel sources can't be fed to it directly and PEQd.

I get around these same constraints in a very clumsey way: I feed pure two channel digitally to the DEQX, which has a volume control, and simultaneoulsy feed the 861 so it can synthesize surround ambience channels (Linkwitz's approach), and, I feed multichannel to the DEQX via the 861s analog outputs.

Not doing this with you Greg. You want to focus on one small aspect of speaker design and it's not that simple.
John,

I'm not trying to be simplistic. In contrast to the devices I help design - speakers
are extremely simple. Do you think I would get anywhere in my own profession if
I were only concentrating on one small aspect?

I'm just pointing out that the Lake is a very powerful tool; but with that power comes
dangers. Whether you understand it or not [ and you evidently don't ] there are real
mathematical and physical reasons for the results that Leo Spiegel noted in his
experiments at Apogee. The Lake processor in the hands of a real pro like a
Mark Seaton and / or a Bruce Thigpen can be a very powerful and useful tool; like a
scalpel in the hands of an adept surgeon. Like the surgeon; they know what they can
and can not do with this powerful tool.

The more amateurish installer should probably stick with something like the DEQX.

The lesser professional installer should probably steer clear of the Lake; for the same
reason, continuing my previous analogy; that we don't let children play with knives.

How does this device differ from the Tact Audio Theater Correction System or the Lyngdorf system?

Bump

The Tact/Lyndorf systems are automated impulse response correction system. The Lake is a fancy EQ. A good one, but completely different goal and operation.

The Tact/Lyndorf systems are automated impulse response correction system. The Lake is a fancy EQ. A good one, but completely different goal and operation.

If I am understanding you, the TACT et al automate the EQ process whereas the Lake is completey user driven. Fair?

yes. TacT, DEQX, Audyssey and Lyndorf allow *some* user input, but the actual EQ curves would be done through system measurement, rather than using your ears and user input as an engineer might do with Lake.

I was very fortunate to have Alan from deqx come out to my place recently and demo the new pdc (or whatever it's now called) 3 to me (it is a substantial improvement BTW).

We had a bit of fun, and he demoed the 'new measurement' procedure he'd developed (which simply allows a much longer window before the first reflection appears) along with a new bass calibration technique, which worked absolute wonders to an already impressive bass sound.

Anyway pertinant to some of the conversation here, he also changed my crossover points AND reduced my steep slopes I had previously used, from 150 db/octave down to about 60 db/octave.

Admittedly there were quite a few simultaneous changes so it's hard to seperate out the contribution of each, but the net effect was, to put it very mildly, dramatic in the extreme.

So I can vouch for the statement that steep x-over slopes are not automatically the best and point out that as Alimentall states they are available with the deqx, certainly not mandatory.

yes. TacT, DEQX, Audyssey and Lyndorf allow *some* user input, but the actual EQ curves would be done through system measurement, rather than using your ears and user input as an engineer might do with Lake.
John,

The TacT [ and I use the TacT RCS 2.2XP ] equipment allows one to set the
"target curve" - that is what response do you want the system to have.

As Kalman Rubinson explains in his Sept 2001 review in Stereophile of the
TacT RCS 2.0; you do NOT want the target curve to be "flat":

http://stereophile.com/roomtreatments/437/

The next step was to choose a target response curve so that the RCS could calculate a
real-time correction for the speakers. To answer the obvious question of "Why not
automatically target a flat response?": First, a flat integrated response implies a tipped-up
on-axis response, because room responses are rolled off in the HF (see sidebar). Second,
many speakers and some amps simply cannot handle the power needed for a fully
equalized flat response down to the extreme bass. Third, almost all commercial recordings
are balanced for typical domestic systems, which, whatever their provenance, are almost
never flat. And fourth (this might be controversial), a completely flat response would
almost certainly sap your system of some of the very characteristics that make you prefer
it to other systems. But don't take my word for it—the RCS 2.0 has nine program
memories; you should include a flat compensation for your own evaluation.


Once the user has determined what the target response curve should be; and the
system has measured what the response curve actually is - then the TacT computes
the filter that will convert the actual response into the desired response.

Anyway pertinant to some of the conversation here, he also changed my crossover points AND reduced my steep slopes I had previously used, from 150 db/octave down to about 60 db/octave.

So I can vouch for the statement that steep x-over slopes are not automatically the best and point out that as Alimentall states they are available with the deqx, certainly not mandatory.
terry,

The steep x-overs make the designer's task easier because it "decouples" the
passbands - you can make alterations to one without affecting the others. With a
system with gentler slopes, the designer has to adjust the response of the neighboring
bands if one band is altered; since the bands are "blended" in the vicinity of the
crossovers. So it is more challenging to the designer to use the gentler x-overs.

However, if you are the end-user; you're not interested in the fact that the designer
has to solve a coupled system [ which can be done if you know the mathematics ];
you're interested in the results.

Steep cross-overs can have problems with ringing, and with acausal "pre-echoes";
and that is just the nature of the beast if you know the mathematics.

Keith Howard wrote a very nice article entitled "Ringing False" for the January 2006
issue of Stereophile:

http://www.stereophile.com/reference/106ringing/

Morbius stated:

So in transitioning from slightly below the crossover frequency to slightly above the
crossover frequency - you have a sudden shift in radiation pattern - because you had
a sudden shift in which driver was being used.



Not unless you're a really bad designer and pick an unnecessarily high crossover frequency.

John,

You just don't NOT understand. You get a discontinuity from the fact that you have
suddenly switched driver size; and hence radiation patterns. It's not due to the
crossover frequency.

You have this misconception that a change in radiation pattern suddenly happens
at some frequency. This is FALSE - the change in radiation pattern is GRADUAL.
So it is present at ALL frequencies - so there's no magic frequency that if you set
your crossover below that frequency - then you're "safe".

Nature doesn't do much discontinuously; it is usually continuous. That is it is always
present perhaps at low amplitude, then increases as frequency increases, for example.

Too many engineers APPROXIMATE this as a sudden onset - but that's an approximation
and not reality.

You have this misconception that a change in radiation pattern suddenly happens at some frequency. This is FALSE - the change in radiation pattern is GRADUAL. So it is present at ALL frequencies - so there's no magic frequency that if you set your crossover below that frequency - then you're "safe".
As usual, most statements of absolutes regarding loudspeaker design are going to be wrong in various situations or for particular applications...pushing a crossover an octave or so lower can easily have two drivers behaving very similar in 3D radiation (mostly dependent on baffle size at that point). A simple example would be a 5" midrange crossed to a 15" woofer at say 800Hz vs. 200Hz. Both the radiation of the drivers around crossover and the required spacing between drivers will make for problems with an 800Hz crossover that won't be a problem when crossed 2 octaves lower...Arguements can be made both for and against steep vs. shallower crossovers. Both are compromises, and your own prioritization of acoustic behavior will leave you leaning toward one solution or the other

Greg,
You are theoretically correct in the vast majority of statements you make in this forum (even in areas outside of physics), but for some reason, John is able to take you places you are not necessarily wanting to go. The question is not whether radiation pattern changes, but rather what is their experiential effect to the listener (not a measuring instrument or a mathematical equation). As Mark noted, changing the frequency response does have an effect on the experienced "beaming" qualities of a driver. A 12" driver will "beam" far more at 1200 Hz than it will at 200 Hz, thus if the driver you are mating can perform closer to 200 Hz then you reduce the chances of incompatible radiation patterns. I know you know this, but your responses to John are so absolute that you make it seem as if the choice of frequency does not have an effect on how well drivers of differing sizes can mate from a radiation perspective.

You just don't NOT understand. You get a discontinuity from the fact that you have
suddenly switched driver size; and hence radiation patterns. It's not due to the
crossover frequency.

i understand perfectly well, thank you. I like the double negative, BTW, which i guess means i DO understand.

You have this misconception that a change in radiation pattern suddenly happens
at some frequency. This is FALSE - the change in radiation pattern is GRADUAL.
So it is present at ALL frequencies - so there's no magic frequency that if you set
your crossover below that frequency - then you're "safe".

The change is extremely gradual right up until it begins to beam. That's why you work within the ranges where it doesn't. NHT's Xd has a smoother off axis response than most any low order speaker in the horizontal plane and dramatically better in the vertical domain. Therefore, all of your claims fall flat and show your limited understanding of speaker design which is why you and i always end up in an argument. Math is math. Speaker design is more than just math.

Keep in mind that this post is 2 weeks old and you have already responded to it prior. Therefore, I can't see the point of bringing it up again except as a troll.

The problem is to get John to accept the fact that the effect exists at all; and then we
can quantize the effect.

Incorrect. The problem here is that you continuously mischaracterize anything I say, which is what, i think, Raul is trying to explain to you - that you are finding argument where there is none and making a big to do over it.

John thinks that steep slope crossovers are all good, and have
no downsides.

Incorrect. I have said that the benefits generally outweigh the disadvantages, at least with DEQX's design. The downsides are there, but largely cancel acoustically when implemented properly. I've said this to you 100 times.

That wasn't to John's liking -
he thinks steep slope crossovers are a design goal in themselves. .

I have never said such a thing. I have said that many of the benefits of steeper crossovers - low distortion, high output, low cone resonance, wide dispersion - are obvious design goals and steeper crossovers is simply a tool to achieve them.

Sadly, we agree in many ways, but you insist on twisting my POV until you think it is opposite of yours, when it is simply different. If you can't refrain from mischaracterizing my statements, please put me on ignore.

John,

The TacT [ and I use the TacT RCS 2.2XP ] equipment allows one to set the
"target curve" - that is what response do you want the system to have.

As Kalman Rubinson explains in his Sept 2001 review in Stereophile of the
TacT RCS 2.0; you do NOT want the target curve to be "flat":



I agree totally that a flat response at the listening position is undesirable, and the stated reasons in the article apply.

The deqx is a loudspeaker tool and so it aims to create an anechoically flat speaker, in other words an accurate transducer hopefully at last bringing the loudspeaker itself closer to the accuracy of the components preceding it (the loudspeaker has always been the 'weak link' in the chain).

Of course once we put a 'flat' speaker in the listening room then by default the typical response curve is automatically produced. As the Tact measures from the listening position, it requires the introduction of appropriate algorithms to artificially create/recreate the natural frequency response. So in the end a 'natural curve at the listening position' is created by both systems.
The Tact has 9 preset target curves, well the deqx (theoretically) has about 400 possible preset target curves, of course such theoretical abundance would never be utilized.

The remainder of the deqx procedure then targets (as does the Tact) the especially problematic area under 500 hz as a seperate step. This mirrors the longer sweep by the Tact under 200 hz as noted in the article.

I agree with you (for my system) that the steeper curves I first used were not optimum, but just as a Ferrari is capable of doing 200 mph yet would be very rarely ever driven at that speed the fact that the deqx is capable of doing 300 db slopes also does not mean that they would or should be always utilised, as Allimental keeps trying to point out.

EDIT It may (?) be of interest in this discussion of slopes, driver size and beaming that I have an 18 inch woofer crossed to a 6.5 inch mid at 200m hz. Alan himself was curious/sceptical ? about the ability of such disparately sized drivers stitching up well.

He no longer is sceptical.

John,

The TacT [ and I use the TacT RCS 2.2XP ] equipment allows one to set the
"target curve" - that is what response do you want the system to have.

As Kalman Rubinson explains in his Sept 2001 review in Stereophile of the
TacT RCS 2.0; you do NOT want the target curve to be "flat":

The correction curve, however, is automatically derived by the processor. It's not a manual process for the most part. Because otherwise, it would take thousands of points. It's impulse response, not frequency equalization.

EDIT It may (?) be of interest in this discussion of slopes, driver size and beaming that I have an 18 inch woofer crossed to a 6.5 inch mid at 200m hz. Alan himself was curious/sceptical ? about the ability of such disparately sized drivers stitching up well.

He no longer is sceptical.

I have a rule of thumb that each driver, aside from the upper midrange, should be no more than double the size of the next smaller driver, and ideally about 1.75x (4.5" to 6.5" to 10", for instance) , due to blending issues. DEQX largely gets around this and allows some extra play in this area. Matched up a 10" with a 3" on a Thiel and it was amazing how much better the blend in dispersion and seamlessness was than with the stock first order crossovers.

I've never heard a speaker more seamless than a properly DEQX'd speaker, almost regardless of the driver choices.

The correction curve, however, is automatically derived by the processor. It's not a manual process for the most part. Because otherwise, it would take thousands of points. It's impulse response, not frequency equalization.
John,

It's BOTH frequency and impulse response correction - they're mathematically coupled!!!

You really do need to study the mathematics of Fourier Analysis.

The correction curve, however, is automatically derived by the processor. It's not a manual process for the most part. Because otherwise, it would take thousands of points.
John,

Yes - I EXPLAINED that in the part of my previous post that you DELETED from
the quote. Here's the relevant part:

Once the user has determined what the target response curve should be; and the
system has measured what the response curve actually is - then the TacT computes
the filter that will convert the actual response into the desired response.

I say above the "TacT computes the filter". The user doesn't compute or specify
the filter; the user specifies the target response and the TacT processor does the
computation of the filter knowing the "as measured response" and the "desired
target response".

I
Incorrect. I have said that the benefits generally outweigh the disadvantages, at least with DEQX's design. The downsides are there, but largely cancel acoustically when implemented properly. I've said this to you 100 times.

John,

...and you've been 100% WRONG 100 times!!!

I think the best analogy here is the one I originally used in post #29; contrasting
Class A/B and Class B amplifiers.

A crossover with a gentle slope is like a Class A/B amplifier. The "hand-off" from the
positive half-cycle amplifier chain to / from the negative half-cycle amplifier chain is
done gradually over a finite fraction of the cycle interval.

A steep cutoff is analogous to the Class B amplifier; where the "hand-off" is at an
instant in time; the zero-crossing point.

One doesn't celebrate the fact that the Class B amplifier hands off at a instant in time;
while denigrating the Class A/B for a more gradual hand-off.

This discussion really isn't about the technical merits of one approach or the other.

In spite of the technical merits; you defend the high-slope crossover solely because that's
what was implemented in that mid-fi sounding piece of crap called the NHT Xd which
you so shamelessly hawk; the manifest shortcomings of which, nothwithstanding.

Thread closed.