Adjusting room sound using DSP (not damping)

[kostiska]

I am wandering about using dsp room correction. There are two options as far as I know. TacT and Deqx. Deqx performs a speaker correction via active crossovers. Are there any views / experience in using deqx with harbeths (SHL5) ? It might sound a bit cynical to intrude in speaker design , but I feel that the room acoustics issue must be dealt drastically for the many of us that cannot afford a special listening room. It might influence what we here even more than the speaker itself.
Thank you

Kostis

[Jeff Stake]

I agree completely. Yesterday, using a TacT S2150XDM, I measured the following speakers: Harbeth C7ES-2, Harbeth M30, Spendor SP1/2E, Spendor LS3/5A 8 ohm, and Rogers LS3/5A 15 ohm. From 100 Hz to about 300 or 400 Hz, and especially between 100 and 200 Hz, the frequency responses of these excellent speakers were similar. The height of the mic and the height of the speaker [together creating floor bounce, etc.] and the room itself were having more effect on the sound in that range than the choice of which speaker to use. And the effects of the floor bounce and room were indeed terrible to behold. The tests showed huge swings, more than 10 dB from top to bottom, through that range. Depending on speaker and mic height, there were sometimes two peaks and two valleys between 100 and 200 Hz, often so steep that they would be impossible to detect with 1/3 octave pink noise and impossible to correct with 1/3 octave equalization. To reiterate, in that critical range, the room and placement were making more difference than the choice of speakers. It was interesting to see the changes in results obtained by adjusting the height of the speakers. It was also interesting to see that changing the height of the mic relative to the tweeter. By moving the mic just a few inches, I was able to smooth the highs considerably. Note, the TacT software does not seem to measure room effects above about 1KHz. From that point up, the measurements are probably similar to what would be achieved in an anechoic chamber.

[kostiska]

Thank you very much Jeff for the information. This is a quantifiable justification of the importance of room correction (as expected if I may add). Following logical assumptions , the use of properly designed speakers would be greatly enhanced in a given faulty listening space if someone could apply dsp for room correction. In deqx technology in order to perform this properly, one should have direct access in the drivers bypassing the internal crossover. As a prospect Harbeth buyer I wonder how I could do this without voiding warranty.
Has anyone tried this (especially with deqx where you can have active crossovers as another parameter to calibrate performance).

Thanks

[Jeff Stake]

You are welcome, Kostiska.
My measurements would indicate that the Harbeth and Spendor and Rogers speakers I tested have excellent crossovers. I did not notice any significant frequency response aberrations between 2KHz and 4KHz. For that reason, I would try digital correction on the whole speaker rather than trying to replace the internal crossover. Digital correction is needed most for the region below 200Hz. It could be useful up to 400Hz, and maybe up to 1K. Whichever it is, you do not need to correct anything in the crossover region. Trying to optimize both the bass and the crossover will make your job more difficult and make it far less likely to succeed. If you bypass the crossover you will void the warranty and increase the risk of damage to the drivers. Why void the warranty when you do not need to do so to get fabulous sound? Since you do not own Harbeths yet, you do not know how wonderful they can sound without any correction. After you apply room correction to the whole speaker, or just to the woofer through the woofer terminals if you wish, I predict that you will see that you do not need the headaches and dangers of bypassing the crossover. Fix only what needs fixing.

[kostiska]

Jeff, don't get me wrong. I am not trying to fix something. Mostly I try to make a buying decision to get best results (based mostly on textual information as I have limited access on actual listening experience). So my initial question is to consider dsp room correction in general as technology (which seems very plausible to me since the room is last important link of the chain). So I searched the net for dsp implementations and I found TacT and Deqx. Focusing on Deqx they offer speaker and room correction as a bundle (DEQH they call it). But they need to replace the crossover in order to manage off axis response (possibly needed in a mic managed calibration process) and phase management (time domain speaker performance). I have no data to prioritise these objectives because I gather the deqx software uses them as phases for the final outcome. The claim is I guess that off axis speaker performance optimisation of an even excellent on axis performant speaker will greatly enhance in room performance
So I got to the Harbeth group to ask for help for any user familiar with this (Harbeths are the speakers of choice, I intend to use ). You mentioned the woofer terminals, does this mean that there is access from the terminal in the back side of the speaker directly to the woofer (i guess there is a low pass filter somewhere) ?
Going back to the initial dsp room correction question it would be interesting to note that if you get into the digital paradigm the audio equipment has a new emerging branch (digital amplifiers etc versus the tubes and analog branch) which has quite a different pricing (more consumer friendly). Harbeths being an optimised solution for the unreplaceable last analog stage could possibly be the best solution for the dsp case as well.

[Jeff Stake]

Yes, all of the current Harbeth speakers except the C7 ES-3 allow you, on the rear panel, to connect one amp to the tweeter and another amp to the woofer, but the crossover remains in the circuit for both. I do not understand how the electronic crossover could change the off-axis response, other than by changing the crossover frequency or slope, which I think would be a mistake. Maybe Alan can comment on whether that is possible. If you are trying to contain your costs, note that the TacT S2150XDM integrated amp offers room correction for a lower total price than buying the room correction pre-amp plus an expensive amplifier.

[Groovetracer]

Sorry but this message was corrupted. Groovetracer commented that he was considering bypassing the M40s intenal crossover and using an extrenal DSP box ....

[A.S.]

As I said, I strongly recomment that you do not bypass the internal crossover in the M40 - or indeed in any Harbeth. The crossover adjust the levels, phase, frequency division and slopes of the drive units. Without access to my original design objectives, measurements and techniques there is really no chance of replicating (hence at least matching let alone improving upon) the ex-factory situation. If there was, I would give up the painstaking design process and buy a programmable DSP box and go on a long holiday!

"By the way ... word of caution: This bi-amping arrangement I have not tried. As I've said before, the M40s bass crossover expects to be loaded by the midrange crossover (and vice versa) from the same amp source. You are in uncharted territory!"

[Groovetracer]

As I said, I strongly recomment that you do not bypass the internal crossover in the M40 - or indeed in any Harbeth. The crossover adjust the levels, phase, frequency division and slopes of the drive units. Without access to my original design objectives, measurements and techniques there is really no chance of replicating (hence at least matching let alone improving upon) the ex-factory situation. If there was, I would give up the painstaking design process and buy a programmable DSP box and go on a long holiday!

"By the way ... word of caution: This bi-amping arrangement I have not tried. As I've said before, the M40s bass crossover expects to be loaded by the midrange crossover (and vice versa) from the same amp source. You are in uncharted territory!"

Alan, I guess I'm not fully understanding your explanation regarding "bypassing the internal crossovers" I never said I was going to use DSP. What I asked you originally was whether I could remove the jumper from the lower and mid driver and use a horizontal biamp procedure with a solid state amp powering the bottom end and a tube amp powering the 8" driver. Why would I bypass the internal crossover? I apologize for my ignorance Alan but could you elaborate on this for the final time?

[A.S.]

OK some slight misunderstanding on my side. But it doesn't change my overall recommendations. If you remove the biwire links and attempt to drive the bass unit (via its crossover section) independently from the mid and top, you will not get the result you are hoping to achieve.

To drive the individual sections of the passive M40 (or indeed any multi-way speaker) with amplifiers of different types, different brands, different models you need electronic and acoustic measuring equipment. At the very, very least you need to determine the phase of the output of the amp compared with its input - you can not assume that it is in phase: many are not. You must also be absolutely certain about the gains in the amplifiers and you have to turn the loudest one down until it is exactly the same as the quietest one, and then fix the gain - glueing the voloume controls in fixed positions would be essential.

Second, the M40's crossover needs to sense at its input the presence of all three sections .... bass, mid and HF connected together, in parallel, as with the links in place. If you isolate the sections and then drive them with separate amplifiers (whether they are identical or not) the crossover's input loading will be totally unpredictable, and this will screw-up both the audio levels arriving at the drive units, their inter-phase, their acoustic Q and the filter shapes. Guaranteed.

Trust me: even as the designer I wouldn't undertake such an experiment unless I had all the equipment I've mentioned, a calibrated microphone and at least 3 months of free time. There is absolutely no chance that you can improve upon the design as it stands by trial and error. If you could then we should swap jobs!

[Groovetracer]

OK some slight misunderstanding on my side. But it doesn't change my overall recommendations. If you remove the biwire links and attempt to drive the bass unit (via its crossover section) independently from the mid and top, you will not get the result you are hoping to achieve.

To drive the individual sections of the passive M40 (or indeed any multi-way speaker) with amplifiers of different types, different brands, different models you need electronic and acoustic measuring equipment. At the very, very least you need to determine the phase of the output of the amp compared with its input - you can not assume that it is in phase: many are not. You must also be absolutely certain about the gains in the amplifiers and you have to turn the loudest one down until it is exactly the same as the quietest one, and then fix the gain - glueing the voloume controls in fixed positions would be essential.

Second, the M40's crossover needs to sense at its input the presence of all three sections .... bass, mid and HF connected together, in parallel, as with the links in place. If you isolate the sections and then drive them with separate amplifiers (whether they are identical or not) the crossover's input loading will be totally unpredictable, and this will screw-up both the audio levels arriving at the drive units, their inter-phase, their acoustic Q and the filter shapes. Guaranteed.

Trust me: even as the designer I wouldn't undertake such an experiment unless I had all the equipment I've mentioned, a calibrated microphone and at least 3 months of free time. There is absolutely no chance that you can improve upon the design as it stands by trial and error. If you could then we should swap jobs!

Thank you again Alan for taking the time to respond. I completely understand the challenges and drawbacks in my application of bi-amping now. I wasn't aware of the possible phase issues, let alone the crossover design (obviously) so thank you for clearing those important issues up for me.
I thoroughly enjoy the performance of both M30 and M40 and applaud your openess to answer questions regarding your products.

[A.S.]

My pleasure. I have gathered my working knowledge over some 25 years the hard way, by trial and observation of results. I'm a little more conscious of the time available to man now and the last thing I want is to see any of you chasing elusive butterflies. I've done that for you and have learned the lessons which I've built in to Harbeth designs.

You get on with enjoying the music - I'll worry about the back room stuff!

[T.W.]

I found another possible solution: LYNGDORF AUDIO RP-1.
I wonder if somebody owns it and if it is recommended.
(I would us it with M40 and Marantz PM-14 MKII KI)

Any comments welcome...

T.W.

Link:
http://www.lyngdorf.com/index.php?op...d=21&Itemid=45

[Ned Mast]

I'm currently using a TacT 2.2XP with an analogue amp and M40s. I think the technology is basically equivalent to the Lyngdorf - at least in the way in which it "corrects" room anomalies. I recommend it highly. I think the extent to which it improves the sound of your system will depend to a large degree on the amount of room treatment you have installed and where you have your speakers placed. With extensive room treatment - especially dealing with bass frequencies - and ideal placement away from reflective wall surfaces, the improvement will be subtle. But, I believe, still noticeable. With my unit I am able to place the M40s with their backs only about 2 feet from the wall behing them. Without room correction, this results in a lack of clarity in the bass; with the correction, absolute clarity. Of course, the advantage of putting the speakers close to a rear wall is the added reinforcement of the bass. But this can lead to a somewhat 'boomy' bass. The TacT - or the Lyngdorf - will allow you to have this bass reinforcement but with clarity. With my M40s either closer to the rear walls, or quite a bit farther out, the advantage of room correction was still noticeable in the treble. Again, clairy. Without correction, it is as if there is a veil between me and the speakers. Individual musical lines are not as clearly delineated; there is not as much brightness to the sound, and I DON'T meant edginess. Transients seem to be greatly softened. A harpsichord recording, for example, looses the sparkle to it's sound without correction. In short, then, having acquired Harbeth speakers, I firmly believe the next best purchase to make is digital room correction of some sort. MUCH more rewarding than expensive amplifiers or wires.

Ned

[T.W.]

Thanks Ned - that's what I wanted to hear :-)

My problem id that I never had and still don't have a perfect room. It's big with some funiture and carpets. The main problem is that the M40's sound too bassy. But that's not the problem of the M40's. It's the room. I moved the speakers around. There are of course some places where the "stage" is much better. But for the bass it doesn't seem to change anything at all. I tried all the "rules" where to put the speakers. But most of them are nonsense.

I will now try to find some info about the TacT you mentioned and then I will decide what to do ... Probably spend some money...

Thanks again,
Thomas

[Ned Mast]

Thomas,
There are various DSP components available - TacT and Lyngdorf are a couple of the better known ones. So I would do a lot of research. I don't know of any that aren't fairly expensive; I bought my unit used to save a bit.

Room/speaker interaction seems to be a funny thing. My M40s sounded at least good everyplace I set them up (even with the over-emphasised bass in a couple of places). Yet when I auditioned them in the dealer's room, they were disappointing. After all I had read, I knew that the room had to be the problem. Fortunately, he let me take them home on a trial basis, and I knew pretty much immediately that they would stay.

One thing that helped in my room to get the best sound from them was to get them well off the floor. My stands are 25". If you have them much lower than that, you might want to try getting them up higher to see if that makes a difference.

Good luck, Ned

[T.W.]

Hi Ned,

I finally found my solution: Lyngdorf TDAI 2200 with RoomPerfect.
It's just great! On the Lynkdorf page you'll find some reviews. There
is also one from Peter Greene in "the absolut sound". The funny thing
is, that Peter used the Lynkdorf with M40's.

Ok, it's not cheap but probably the best investment that I ever made -
sorry the second best - the best are the M40's.

Thomas

[Ned Mast]

Hi Thomas,

Congratulations on getting the Lyngdorf. With that and the M40s you should really have a wonderful sounding system! Let us know how you do with it.

Ned

[T.W.]

After 2 weeks with the Lyngdorf driving my M40s I'm
even more sure that this is an ideal combination.
In my opinion the Lyngdorf and the M40s love each other.

The sound is still "Harbeth". Roomperfect doesn't seem to
add or remove anything from this sound. It just removes the
very bad influence of my room. The sound is very neutral.
There is still a very deep bass but humfree. The stage is
brilliant and voices are nailed on in their position. (This
is of course the feature of the M40s but a bad room messes
this up. )

I never received a sound like this before in any room,
at any speaker position with any speaker.

With Roomperfect you also have to forget what you learned
about speaker position and stand height. Normally you would
put the speakers more far from the back walls. Especially
with the M40s you would also choose some higher stands to
remove some of the bass. But you would also like to have
the tweeter around ear height. This requires high listening
chairs :-)

Lyngdorf recommends to put speakers closer to the back walls
to have a better impulse characteristic and to let
Roomperfect do the job of reducing the bass. That works perfect!
This could also help to reduce the stands height for an optimal
tweeter height. (Has also a higher wife acceptance factor)

Another thing to mention is that the TDAI 2200 is a fully digital amp.
It has anolog and digital inputs. Fortunatelly my CD player has
a digital output and the result seems to be better than using
anolog input. Another good thing: I'm now out of the discussion
about the best NF cables ... Good digital cables are cheap.

I would really like to know what you think about the Lyngdorf
compared to your DSP solution. You could walk into a hifi shop,
borrow the thing and give it a try. But be warned - it will stay!

Thomas

[Ned Mast]

Hi Thomas,

I'm really happy to hear that your experience with the Lyngdorf has turned out to be so positive! With my TacT, I've been able to do exactly what you've done - put the M40s closer to the wall behind them and then let DSP deal with the peaks in the lower frequencies. The bass response is absolutely solid and clean. But also in the upper frequency range, there is a clarity to the sound that isn't there when the DSP isn't engaged. I can only think that in time anyone seriously concerned about achieving good sound from his/her system will put DSP at the top of the list (just after Harbeth speakers!).

Ned