You aren't boring nobody. Its your thing. You may paint it n times. It was just humor to remind that you 'have work to do' as Walt reminded Locke in "Lost''. And the most interesting part is keeping us in suspense. We sonic Romans give thumbs up to the preparations and we are eager for the main festivities to begin. If you may.
It's inspiring So this is the kind of music you will be playing thru the speakers....?....... 😀 😀
ShinOBIWAN said:
Fine tune! Don´t think it will display all the qualities of your speakers but a damn fine tune.
Btw. If I may derail the thread once more I just got the Armin Van Buren: A State Of Trance 2007 compilation... That's one very very nice trance-CD. Go buy! ( no, I don't know Armin, and no a don't make a profit if you decide to buy a CD.)
Kind Regards and just keep them pictures coming.
TroelsM
Btw. If I may derail the thread once more I just got the Armin Van Buren: A State Of Trance 2007 compilation... That's one very very nice trance-CD. Go buy! ( no, I don't know Armin, and no a don't make a profit if you decide to buy a CD.)
Kind Regards and just keep them pictures coming.
TroelsM
salas said:
AT is an excellent choice, these cones sound not as muddy as you expect. Transients are more like the Audax HDA cones, but with more resolution.
In my world its an oldfashioned way of thinking, from the days when they(we) didnt know how to make any xo work, and they made up that conclusion - i would think that each frequency has its own ressonanse no matter what driver is used
Can you explain that a little bit further?
Lindell said:
Absolutely.
I have audiophile recordings too but I generally only listen to those when I want to remind myself how good the equipment can sound. When it comes to sitting back to enjoy music rather than the equipment then you'll find my tastes are very much non-audiophile ie. mp3's, internet radio, gaming, TV and movies.
salas said:
Salas, I mentioned it earlier but just to briefly go over it again; I've already had the speakers running but in an TMW config rather than the final WMTMW. I was fairly pleased, no concerns about the RAAL and AT working together either. I do have some problems with the system. I wouldn't call the AT 5" the greatest midrange in the world its OK I guess but a little too soft and lacking presence and projection. But the bass is possibly the best I've heard in my room - smooth and distinctly unimpressive until you realise thats just how it should be and the RAAL I love to bits.
Much of the XO talk will be fairly fruitless since my methods are far from traditional.
Here's an excerpt from a conversation between me and another forum member that goes into a little detail about the process and methods I use to linearise and voice a loudspeaker:
Linkwitz is really talking about units such as the Behringer DCX. I had something more powerful in mind. Such as 30k+ taps on an FIR filter giving effectively 30,000 points of frequency and phase shaping. These points can be manipulated and shaped with target curves that are created in software and the curves can either be drawn or taken from presets.
So in effect you measure your drivers in semi-anechoic conditions and mounted in the final cabinet. This data is imported into the filter generation software such as "Acourate". From here the software takes a look at the response and you decide just what you want to do with that response. Do you want to flatten the entire passband of the driver to within +/-0.5dB and/or do you wish to extend the passband of the driver another octave either side of the drivers original passband in order to facilitate more ideal rolloff characteristics? There's an infinite number of options and all this is quite easily achieved for someone who is familiar with what makes up a technically correct speaker - subjective tuning comes later on.
After you've done this you then apply the crossover filter transfer functions to your 'corrected' driver response. Then apply the filter, re-measure in semi anechoic conditions to confirm filter performance.
The upshot of all this is your filters almost perfectly mimic the required acoustic rolloffs(ie. the transfer function of your filters is correct in the acoustic domain rather than just the electric) and the pass bands of the drivers are flat and level matched. You can then look into things such as symetrically matching the slopes on both the high and low pass around the XO point for excellent phase and frequency response characteristics etc. Out of all this a linear system(or making the best out of a bad situation) is born.
Now once you take your loudspeaker out of the semi anechoic environment in which it was measured, setup and, as a result, subsequently performed as a near linear system you'll find all has gone to pot as you now have room influences affecting the performance. Whats important here though is that you can determine that any deviation from the semi anechoic response is room related.
You then proceed to add the DRC components(FIR filters similar to the driver correction types used above) to help smooth the worse of the room problems lower down. Extra care must be taken at this stage though because excessive DRC can sound worse than no DRC. But because we're working from a solid grounding after the all that semi anechoic filter generation work you can confidently move forward knowing that the speakers are 'correct' and the rest is a balance between room correction and what sounds subjectively right to you. Where people using DRC often go wrong is trying to mash the in-room loudspeaker performance back to that of the near perfect semi anechoic measured performance - it can very rarely be done with just digital methods. My suggestion is only correct upto 400 or 500hz then use physical treatments above 800-1Khz. This allows the treatments to discrete and rather more minimal because your only dealing with upper frequencies and the bass, which needs huge and impractical treatments, is dealt with digitally.
There's obviously more to it than just this basic outline but I'm sure you can see its has power and flexibility like nothing else. Which brings about the question of just what will it do, if at all, to enhance the sound over the Linkwitz' already excellent implementation? Nothing is for nothing so I have no definite answer for that but I'm confident you could build upon Linkwitz' work to further enhance the performance and better tailor the speaker to the listening environment.
BTW If your source is digital then you'll be keeping the signal chain entirely digital right up to the amps. Further performance gain can be got by looking at some very nice external master clocks/DAC combos.
Its a shame I didn't do the presentation over the weekend during the meet. It was intended to be an indepth and practical demonstration of the points I've mentioned here. Much of this is more easily explained with graphs and visual examples.
Finally don't get me wrong, its definitely not all a bed of roses but the potential upshots are an accuracy of sound that you'd struggle hard to find in other filter technologies. Whether that is your 'thing' is another less black and white topic
Salas these are specially for you mate 😉 😛 😀
Better shots of the colour which more accurately depict the overall look the eye sees:
Speakons for the loudspeakers:
Better shots of the colour which more accurately depict the overall look the eye sees:
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
Speakons for the loudspeakers:
An externally hosted image should be here but it was not working when we last tested it.
m0tion said:
Short to medium term - yes.
I want the DEQX but I'm stuck between a rock and a hard place. On one hand they've annouced a new product, the PDC3.0, but its still not available and on the other you have the PDC2.6 which is available but will be made obsolete with the introduction of the new unit. Since there's no trade in for a PDC2.6 against the upcoming PDC3 and your looking at about £2000 for either, then it makes for a compelling financial reason to hold off.
ReTrO said:
Richard these are the codes I used to order:
NLT FC
Rubber sealing gasket for NLT4MP
NLT 4 MP-BAG
Male 4 pole panel mount (colour:black)
NL 4 FRX
Female 4 pole right angle plug
Hi all,
Using exceptionally long FIR filters in speaker crossovers seems to be gaining in popularity of late (probably due to modern hardware finally having enough processing power to run them), and are treated more and more as some kind of panacea. But the approach has its problems and I can only think it's a terrible idea. I'll put my points forth:
- FIR filters give rise to time-domain problems. Longer FIR filters have longer time-domain effects. You really want to keep these effects as short as possible because they're just as audible as anything else. While the extra filter length does give you extra control in the frequency domain this can easily be more than offset by the 350ms(!) or more of each pre- and post-echo (assuming 32k taps at 441.kHz)
- A drive unit in a box is broadly-speaking a minimum-phase creature if you ignore its nonlinearities. Minimum-phase correction, such as a structure of biquad filters or similar, can be used to correct all of the linear response deviations of a drive unit in both the amplitude AND the phase domains. It will do this without adding any problems of its own. Of course, such a structure cannot be used to correct nonlinearities in the drive unit (cone breakup, motor issues, surround resonance, etc) - but then nor can any electronic system, FIR or otherwise, so the point is moot.
- It is very difficult to measure anything accurately enough that doing a "hard inversion" will give you good results. Even doing it with expensive kit in an anechoic chamber isn't perfect, so basing it on quasi-anechoic measurements in a room seems to be heck of a leap of faith at best. You're back down to the long-protested issues of "Well, this one measurement from this one point in my room tells me the frequency response is flat, so therefore my drive units now perform perfectly!"
An approach at least worth considering is to stay far away from very long FIR filters in the crossover. Equalization of individual drive units toward flatness could be better achieved with minimum-phase filters. Once the drivers are flat, the crossover itself can then be implemented with relatively short FIR filters to give you an overall linear-phase response with minimal time-domain problems. Worth noting is that it's probably more beneficial to use linear-phase filters for the mid/tweet crossover than it is for the mid/woofer - the room is having a much bigger effect on the phase of your sound at the lower frequencies.
It is easy to think that because the maths says certain things are invertible, that they really ARE practically invertible in the room. I think though that while digital correction can be very useful, too much of it sounds like exactly that. All things in moderation!
I'm also not a fan of FIR-based room correction but that's a whole 'nuther kettle of fish.
As always, YMMV. These are just my musings on the subject.
Using exceptionally long FIR filters in speaker crossovers seems to be gaining in popularity of late (probably due to modern hardware finally having enough processing power to run them), and are treated more and more as some kind of panacea. But the approach has its problems and I can only think it's a terrible idea. I'll put my points forth:
- FIR filters give rise to time-domain problems. Longer FIR filters have longer time-domain effects. You really want to keep these effects as short as possible because they're just as audible as anything else. While the extra filter length does give you extra control in the frequency domain this can easily be more than offset by the 350ms(!) or more of each pre- and post-echo (assuming 32k taps at 441.kHz)
- A drive unit in a box is broadly-speaking a minimum-phase creature if you ignore its nonlinearities. Minimum-phase correction, such as a structure of biquad filters or similar, can be used to correct all of the linear response deviations of a drive unit in both the amplitude AND the phase domains. It will do this without adding any problems of its own. Of course, such a structure cannot be used to correct nonlinearities in the drive unit (cone breakup, motor issues, surround resonance, etc) - but then nor can any electronic system, FIR or otherwise, so the point is moot.
- It is very difficult to measure anything accurately enough that doing a "hard inversion" will give you good results. Even doing it with expensive kit in an anechoic chamber isn't perfect, so basing it on quasi-anechoic measurements in a room seems to be heck of a leap of faith at best. You're back down to the long-protested issues of "Well, this one measurement from this one point in my room tells me the frequency response is flat, so therefore my drive units now perform perfectly!"
An approach at least worth considering is to stay far away from very long FIR filters in the crossover. Equalization of individual drive units toward flatness could be better achieved with minimum-phase filters. Once the drivers are flat, the crossover itself can then be implemented with relatively short FIR filters to give you an overall linear-phase response with minimal time-domain problems. Worth noting is that it's probably more beneficial to use linear-phase filters for the mid/tweet crossover than it is for the mid/woofer - the room is having a much bigger effect on the phase of your sound at the lower frequencies.
It is easy to think that because the maths says certain things are invertible, that they really ARE practically invertible in the room. I think though that while digital correction can be very useful, too much of it sounds like exactly that. All things in moderation!
I'm also not a fan of FIR-based room correction but that's a whole 'nuther kettle of fish.
As always, YMMV. These are just my musings on the subject.
Wing:
I remember reading a thread before where you were exploring some different digital crossover and correction techniques. Did you come to any conclusions about how to implement a good system?
I remember reading a thread before where you were exploring some different digital crossover and correction techniques. Did you come to any conclusions about how to implement a good system?
- Status
- Not open for further replies.