+1
Even when designed properly we are still left with the shape of the cones we all use that have phase/time issue simply because of shape. Is this correctable above schroeder per topic, not in my opinion. Same is true of baffle diffraction effects, I do not believe so. rePhase and similar could correct for a given position but limited to that position only IMO. In my mind the only correct solution is to start without the major flaws like Earl claims is paramount. Those with greater flaws will show the greatest improvement with, the rest, not so much. If we start with a flaw and either do not consider it, or incorrectly deem it unworthy of further attention, then only a flawed conclusion can result.
Drivers vary in their characteristics so do crossover parts hence there will be interaural differences caused by driver/crossover tolerances. I don't have any data that would show how big those errors are but they are certainly there and individual speaker calibration would be desirable. Only 1dB deviation will lead to an image shift of already 10%: http://www.sengpielaudio.com/InterchannelLevelDifferencesAndInterchannelTimeDifferences1.pdf
Markus
I do have lots of data on my speakers and speaker to speaker variability is pretty small. I'd have to say that it is less than 1 dB. I don't think this is a big issue. Moving in your seat is goi9ng to be greater than that. I should also mention that the curves that you show are for broadband dB shifts. My speakers are never even close to 1 dB broadband. 1 dB maximum difference at some frequency worst case.
Greebster
"Cone" issues are not that great and can be easily corrected in my experience. Basically the pattern of a cone is wider than that of a flat source (this is in my book) and this must be accounted for in the polar design, but that is quite doable.
If you are talking about the cone not moving as a piston, this is, or at least should be, out of band and also not an issue. It's all about doing things right.
The one problem that I had in the past, which has gone away, is the very heavy magnet resonating on the frame. Neodymium magnets solve this problem. Sometimes there is a spider resonance in band, but the good driver designers, like B&C, have gotten all this under control. Today's 12 and 15 inch woofers are excellent products at a fraction of what the older JBL and TAD woofers (which were also very good) cost.
I do have lots of data on my speakers and speaker to speaker variability is pretty small. I'd have to say that it is less than 1 dB. I don't think this is a big issue. Moving in your seat is goi9ng to be greater than that. I should also mention that the curves that you show are for broadband dB shifts. My speakers are never even close to 1 dB broadband. 1 dB maximum difference at some frequency worst case.
Greebster
"Cone" issues are not that great and can be easily corrected in my experience. Basically the pattern of a cone is wider than that of a flat source (this is in my book) and this must be accounted for in the polar design, but that is quite doable.
If you are talking about the cone not moving as a piston, this is, or at least should be, out of band and also not an issue. It's all about doing things right.
The one problem that I had in the past, which has gone away, is the very heavy magnet resonating on the frame. Neodymium magnets solve this problem. Sometimes there is a spider resonance in band, but the good driver designers, like B&C, have gotten all this under control. Today's 12 and 15 inch woofers are excellent products at a fraction of what the older JBL and TAD woofers (which were also very good) cost.
Last edited:
Linearizing the phase (ie undoing the allpass of your crossover) would tighten the impulse even more (cf curves page 29) and would not change the spacial behavior.
I agree.
rePhase is not meant to be a room correction software (hence it is completely OT here), but a *speaker* correction software (and design, as you can do the whole filtering with it if you have a FIR-capable crossover). It is meant to correct the source and not the context.
Of course the lower you go in frequency the more difficult it is to separate the two (when measuring/analysing, but also when listening it seems, so all is well), and that is where approaches like the multiple measurements averaging jlo uses in Align2 really shine.
This is a manual tool, so as with any EQ (as CD compensation or similar EQ and filtering you do when *designing* a loudspeaker) it is up to the designer to choose what should be corrected and how.
For example undoing allpass phase shifts of the crossover is something that will "work" for the whole listening place (as long as the drivers sum correctly, which is always the case anyway: directivity matching, vertical lobs, all known issues...). This is exactly like doing a CD compensation in term of correction "validity".
On the other end of the spectrum doing small high Q phase or amplitude correction based on a single measurement is of course a bad idea as you are more likely to correct highly position-dependent defects (diffraction, specular reflexion, etc.) and make all other position *worse*.
Between those two situation there are numerous of things that can be corrected, and it comes down to measurement techniques (gating, space averaging, etc.) and human judgment...
mitchba,
If I understand your request correctly, my setup and evaluations meet your criteria. I showed some measurements in Post 235 and below show the differences with, and without, the final phase correction filter. As commented above by several others this is actually relative easy and free for anyone to do. Just create a phase correction filter in RePhase and then convolve some music selections or test tones files for comparison using ABX type software.
I would quickly agree with comments regarding XO choices, EQ choices and house curve as being important and relatively obvious contributors to the overall sound quality. However, my ABX and general listening comparisons both using headphones and speakers has found no easily detectable sound quality difference using overall phase linearization.
My only experience is with lots of different of XO/EQ setups with my modest DIY setup. It's a single room and setup but still is enough to satisfy me that the researchers are correct that indicate that phase rotation (GD) within typical levels is not a significant factor. There is no cost or other penalty in my setup however so I still place a RePhase correction filter into FooConvolve for my music listening. There is a certain satisfaction in seeing a more ideal looking transfer function response in the measurements even if I cannot hear a difference.
Acourate and other software use very different EQ approaches than typical PEQ filters. These EQ approaches are more likely the primary contributor to the improvements you hear. I suggest that since you have optimized your setup using Acourate that you can now use RePhase to create a linear filter with a typical speaker's phase rotation and use it to convolve some of your music selections so that you can then ABX test. In that way you should get a good sense of just how much the direct phase rotation correction portion of Acourate is contributing to the sound. I would suspect the sound is similar, but would be interested to know if that is what you find.
The linear phase XO filter available in Acourate and other software are a different subject in my opinion and I am more open to the potential merits of that capability - higher slopes/less overlap without the phase rotation of IIR filters to contend with. I can see that that may be a significant factor.
The charts below use a 12" mic distance for my FR main speaker. The SWs are muted. The comparison is "with" and "without" the RePhase filter applied.
The SPL overlay is identical.
Both phase traces show a minimum phase ripple that results from early dispersion and reflection effects. The convolved (green) measurement shows that the excess phase is effectively removed.
The 2 IRs are plotted separately for clarity. Most of the minor ripples preceding and following the IR peaks are due to the impact of Nyquist filter in my setup and their spacing corresponds to 44.1kHz. These same ripples show up in a simple loopback test of my measuring system.
Mitch,
Looking at the conclusion of the article:
"Page:7 © 2006 by Dr. Ulrich Brüggemann (Acourate), All rights reserved
Conclusion:
Minphase crossovers have disadvantages. Different type of minphase crossovers have different type of disadvantages. With higher filter orders the behaviour gets worse. This has lead to the preference of low order filters. But also low order minimum phase filters result in time distortions. The different frequencies are played at different times. Is it a wonder that we have a never ending discussion about the quality of speakers?"
My conclusion is that the never ending discussion about the quality of speakers has far more to do with their polar and frequency response than their phase response.
As you can see below, my Sony 7506 headphones have a near flat phase response (where their frequency response is flat) without any FIR filters, they are not "playing different frequencies at different times".
Yet on outdoor shows, when I compare them to my three way fourth order minimum phase sound system, they sound quite similar, even though the speakers exhibit over 1000 degrees of (smooth) phase rotation.
I included jtalden's system with and without phase linearization as a comparison.
Like jtalden, my ABX and general listening comparisons both using headphones and speakers has found no easily detectable sound quality difference attributable to phase linearization.
Having done a similar thing when doing a comparison of high frequency drivers
http://www.diyaudio.com/forums/multi-way/212240-high-frequency-compression-driver-evaluation.html
Other than at levels causing quite a bit of distortion I can't tell the difference between the drivers when equalized flat and crossed over to the LF portion of the recording compared to the original recording, regardless of the phase rotation the crossover induced.
That said, just because I can't hear the differences does not make me assume others can't.
I wish I could hear even half the differences I could when I was half my present age.
Like jtalden suggests, you could use RePhase to create a linear filter with a typical speaker's phase rotation and use it to convolve some of your music selections so that you can ABX test how much the direct phase rotation correction portion of Acourate is contributing to the sound.
Art
Attachments
Last edited:
I suppose that my experience is different, pos, Earl. I'm coming from a nearfield type monitor design and applying it to much larger types. Up close and personal even with min phase the aberrations associated with the polar are just that much greater, hence my love of planars. The rule I've been following for nearly 30 years now is proper phase at the crossover point, minimize diffraction, time alignment, acoustically centered with a flat power response in range of operation. To achieve one has to know the limitations of the drivers used, including the shift away from pistonic, breakup, etc. I can concur with you Earl that in your design this appears to be the case. Do not see this as the case for most other designs. Are they all that inferior? I think not and thus is an issue in the polar, regardless of phase tracking. I'm incredibly sensitive to phase and polar inconsistencies, also can hear stereo bass and infrasound. I'm not going to say everyone should or does hear what I do as we are all different to some degree. It was said that Edison due his tone deaf ear could hear distortions that others couldn't pick out and this gave him a competitive advantage. I'm not quoting and don't need a refresher course on specifically was said, the jist of it is enough for me to believe, no double blind studies required. 
pos, yes like an EQ. My desire is to minimize the issue at the source with a broad band AMT someday, even if that means designed from scratch, until then it's an issue for me that cannot be corrected in the broader sense. With the design I'm working on we can look at the speaker without the low end and see it as a simple 2way HP at 100Hz. 6.5" driver crossed ~2.2 -2.3kHz to an AMT. Polar on the mid will start to narrow around 700Hz to the crossover point where we are -6dB at 60°. If I were to not correct for the physical time alignment thinking I could get away with it by adding additional digital delay to the tweeter would be a quick fix if only the on axis was important. Off axis would fall on its face and grossly take the driver further out of alignment affecting the power response phase, impulse. In my example the tweeter is time aligned axisymmetrical and the front baffle is ~7.5"wide flat, is extended by an edgeless elliptical taper to the edge where it rounds off to the back. Total width is ~23" and depth is ~9". The Area immediate to the tweeter is cut back and a channel between drivers with a U shaped chamfered edge to each of the mids. Troels has multiple examples of and something I did with my last short line array built 23 years ago.
With this in a MTM setup adds complexity to the design, but ehhh am having fun, measuring everything and discover what can be further improved as I proceed, one step at a time. Have over a year of thought in this so far, many measurements and no urge to slap it together after a quick couple of calculations and rule of thumb approach. It will be active, but was designed for passive. Considering the various limitations based on driver choices with this design should minimize most issues that I'm picky about, for now
Back to Earl (and OT), in the late 80's finding some of your research in the preinternet days was a breath of fresh air as you had experimented with reticulated foam and multisub approaches. More of a I'd buy that guy a beer (or in your case a nice glass of wine) for finding mathematically what I could see in my measurements. Similar approach used and visualized this like standing waves and diffraction within a waveguide or cavity load in a broadband sense or enclosure within an enclosure. In a effort to put another dot on an i or cross the T, built an MLTL sub as an experiment where it's folded length was enough to be tested as a multipoint sub. If driver faced down, port was nearly mid room height with reverse true. If laid on its side would be similar to flanking and if raised 27" off the floor would load the room from two points with one or the other loading the room from less than 1/3rd way from the ceiling, again ~27" distant. This symmetry balanced (loaded) the room very nicely indeed. The game plan is to have two of similar in each enclosure, perhaps one driver up and the other down with opposing ports loading the room from the opposing drivers position. In this configuration I achieve 4 point symetrical loading per channel and a similar benefit of stacking and flanking subs with only two drivers. Stereo setup equates to 8 points room is loaded from. I may not buy your book, but could see myself buying you a fine bottle of wine for being on the same wavelength.
With that said I need to buy cases of beer and wine to those I give credit to. So you better watch it Bucko's or you might gets some too, I do have a list! (scary I know heh
pos, yes like an EQ. My desire is to minimize the issue at the source with a broad band AMT someday, even if that means designed from scratch, until then it's an issue for me that cannot be corrected in the broader sense. With the design I'm working on we can look at the speaker without the low end and see it as a simple 2way HP at 100Hz. 6.5" driver crossed ~2.2 -2.3kHz to an AMT. Polar on the mid will start to narrow around 700Hz to the crossover point where we are -6dB at 60°. If I were to not correct for the physical time alignment thinking I could get away with it by adding additional digital delay to the tweeter would be a quick fix if only the on axis was important. Off axis would fall on its face and grossly take the driver further out of alignment affecting the power response phase, impulse. In my example the tweeter is time aligned axisymmetrical and the front baffle is ~7.5"wide flat, is extended by an edgeless elliptical taper to the edge where it rounds off to the back. Total width is ~23" and depth is ~9". The Area immediate to the tweeter is cut back and a channel between drivers with a U shaped chamfered edge to each of the mids. Troels has multiple examples of and something I did with my last short line array built 23 years ago.
With this in a MTM setup adds complexity to the design, but ehhh am having fun, measuring everything and discover what can be further improved as I proceed, one step at a time. Have over a year of thought in this so far, many measurements and no urge to slap it together after a quick couple of calculations and rule of thumb approach. It will be active, but was designed for passive. Considering the various limitations based on driver choices with this design should minimize most issues that I'm picky about, for now
Back to Earl (and OT), in the late 80's finding some of your research in the preinternet days was a breath of fresh air as you had experimented with reticulated foam and multisub approaches. More of a I'd buy that guy a beer (or in your case a nice glass of wine) for finding mathematically what I could see in my measurements. Similar approach used and visualized this like standing waves and diffraction within a waveguide or cavity load in a broadband sense or enclosure within an enclosure. In a effort to put another dot on an i or cross the T, built an MLTL sub as an experiment where it's folded length was enough to be tested as a multipoint sub. If driver faced down, port was nearly mid room height with reverse true. If laid on its side would be similar to flanking and if raised 27" off the floor would load the room from two points with one or the other loading the room from less than 1/3rd way from the ceiling, again ~27" distant. This symmetry balanced (loaded) the room very nicely indeed. The game plan is to have two of similar in each enclosure, perhaps one driver up and the other down with opposing ports loading the room from the opposing drivers position. In this configuration I achieve 4 point symetrical loading per channel and a similar benefit of stacking and flanking subs with only two drivers. Stereo setup equates to 8 points room is loaded from. I may not buy your book, but could see myself buying you a fine bottle of wine for being on the same wavelength.
With that said I need to buy cases of beer and wine to those I give credit to. So you better watch it Bucko's or you might gets some too, I do have a list! (scary I know heh
Last edited:
Earl,
Blauert paper used band limited 25 microsecond bursts, this is roughly single sample at 44.1kHz of CD, yet came up with significant results. Most significant is the learning capacity of subjects. All these bursts present primarily as a mere "pip", yet are differentiated by the ordering of information contained in them. This is not "all indications seem to point to this as not being audible."
Your speakers fail to propagate the waveform of source signal.
Crappy speakers are what they are. EQ does much more for good speaker than for a poor speaker.
Your speakers are likely good speakers. Greatest failing is driver separation at crossover frequency, but in well damped room, and at preferred listening distance this isn't such a big liability. Beyond this, there is nothing new, or special about your speakers, and speakers in general.
What is relatively new, is proliferation of DSP knowledge, computational power, and user friendly software. Many are putting this to good use.
Mitch as mixing engineer has exceptionally well trained ears, and his descriptions are consistent with many others on this forum, and many on other audio forums that have taken the time to learn the software and make correction filters.
Good filter is readily demonstrated when speaker produces consistent square wave response from start of pass band up past 4kHz, range of 99% of all formant sounds, and primary range in which timing cues are located.
Playing swept square wave through my speaker drivers raw, or with IIR filters produces waveforms that at some points are relatively square, and in others look much more like triangle waves. The sound characteristics in these differing regions are quite audible, much as when two instruments in same family play the same note. With correction filters, square wave form is consistent from <65Hz to >4kHz, and sound characteristic is uniform.
We herd sheep, we drive cattle, we lead people. Lead me, follow me, or get out of my way. --- George S. Patton.
Instead of treating forum content as potential argument for or against your speakers, and assuming a defensive posture, why not investigate for yourself, and post some results. Cherry picking from ancient papers as only credible information and siting need for full blown statistical experiment is boring and obvious.
Blauert paper used band limited 25 microsecond bursts, this is roughly single sample at 44.1kHz of CD, yet came up with significant results. Most significant is the learning capacity of subjects. All these bursts present primarily as a mere "pip", yet are differentiated by the ordering of information contained in them. This is not "all indications seem to point to this as not being audible."
Your speakers fail to propagate the waveform of source signal.
Crappy speakers are what they are. EQ does much more for good speaker than for a poor speaker.
Your speakers are likely good speakers. Greatest failing is driver separation at crossover frequency, but in well damped room, and at preferred listening distance this isn't such a big liability. Beyond this, there is nothing new, or special about your speakers, and speakers in general.
What is relatively new, is proliferation of DSP knowledge, computational power, and user friendly software. Many are putting this to good use.
Mitch as mixing engineer has exceptionally well trained ears, and his descriptions are consistent with many others on this forum, and many on other audio forums that have taken the time to learn the software and make correction filters.
Good filter is readily demonstrated when speaker produces consistent square wave response from start of pass band up past 4kHz, range of 99% of all formant sounds, and primary range in which timing cues are located.
Playing swept square wave through my speaker drivers raw, or with IIR filters produces waveforms that at some points are relatively square, and in others look much more like triangle waves. The sound characteristics in these differing regions are quite audible, much as when two instruments in same family play the same note. With correction filters, square wave form is consistent from <65Hz to >4kHz, and sound characteristic is uniform.
We herd sheep, we drive cattle, we lead people. Lead me, follow me, or get out of my way. --- George S. Patton.
Instead of treating forum content as potential argument for or against your speakers, and assuming a defensive posture, why not investigate for yourself, and post some results. Cherry picking from ancient papers as only credible information and siting need for full blown statistical experiment is boring and obvious.
The world is not ready yet
Earl, I swear I do hear some benefits on some specific materials, in ABX testing
Joking aside, FIR filtering is not only about linearizing 24dB/oct slopes, or doing brickwall filters (which brings their own problems...). You can do much more with tailored-made slopes crossovers. For example:
- Horbach-Keele for MTM (2007 AES)
- complementary overlapping slopes to have two transducer playing together around the crossover, which can be useful for a midbass crossover
- complementary asymmetrical slopes to reject breakups when the steep slope is used as a LP, or to protect a fragile transducer when the steep slopes is on the HP side.
Many other examples can be found (and will arise in the future). With arbitrary amplitude (and phase) slopes you can tailor your loudspeaker to somewhat control its behavior from the source (directivity, etc.).
- electronic delay: not possible for passive speakers (or, at least, very difficult)
- physically:
-- tilting the speaker backwards => tweeter points somewhere else but not to the listening position
-- let the tweeter step back in the cabinet => nasty edge right next to the tweeter
-- Waveguide/Horn: ask Earl about their problems
Yes, you'll get preringing, but as long as you're using short filters it is not a problem.
But the main point is: two speakers, one with linear phase crossover, one with minimum phase crossover, will have more differences than just the phase. So you can't say it is the phase that makes the different sound.
You were comparing "FIR linear phase vs IIR/FIR minimum phase" so to me it was obvious that we were digital active in both cases, with delays at hand.
Comparing passive crossovers to linear-phase FIR ones carries more differences than just minimum-phase vs linear-phase.
I think he only sees benefits when properly done.
Even with a direct radiator the baffle acts as a finite (and most of the time abruptly so) 180° waveguide. A properly designed waveguide or horn, with smoothed transitions on both sides, is probably better in every respect.
By the way Physical alignment is the best way to get a good summation over a wide angle, whereas delay alignment is only valid on-axis.
You mean shallow slopes?
The problem lies in the filter curves (amplitude and phase), not the length of the FIR. The longer the FIR the better, but the "sharper" the filters the more preringing you will get in case of improper summation (and with non coincident drivers you always ends up with bad summation at some angles).
Shortening the FIR while maintaining the same filter target curves will only make things worse, as depending on the windowing algorithm used you will either end up with shallower curves that do not follow your target, or with giant ripples (or a mix of both).
I agree completely, see post #393.
Last edited:
The preringing is completely predictible based on the filter amplitude and phase curves, there is nothing that can be done about it (for example Accurate preringing reduction algorithm *has* to imply modifications in the target curves).
A given impulse shape correspond to a specific amplitude and phase response, and vice versa (FFT).
The steeper the curves the more ringing you get.
If your target is linear-phase then the impulse will be symmetrical and you will get both pre and post ringing, whereas if your target is minimum-phase you will only get (double amount of) post ringing, which is obviously less noticeable on musical signals.
If those filters form a complementary and aligned crossover, then all the ringing (pre and post, regardless of the phase relation) will cancel, and you will be left with a seamless transition.
Of course if those filters do not have counterparts (eg DAC antialiasing filter, subsonic filter, phase linearization of a box response, ...), you will get the entirety of the preringing.
But even when used in a complementary aligned crossover, pre (or post) ringing will strike back when (where) the summation is not "perfect".
Having a perfectly phase-coherent and amplitude-complementary summation is easy enough on-axis, but becomes difficult to impossible as you move off-axis (non coincident drivers, differences in polar response, etc.).
Having steep slopes only makes things worse as you have less "blend" and summation errors become more apparent and abrupt.
The perfect "bad" example is a brickwall filter in a two-way direct radiator loudspeaker, with a woofer that starts to beam and a tweeter that radiates 180°: measured off-axis you will get a distinct discontinuity, and associated ringing.
So the equation is simple: the steeper the slopes the more pre/post ringing your filters have *and* the more likely your crossover is likely to expose it of axis.
Now the question of the audibility of aforementioned preringings remains in debate (not unlike phase distortion audibility), but for sure it is less "natural" than post ringing, and probably quite unnatural (acausality!).
So it boils down to a compromise between the benefits of sharper slopes (distortion reduction) and the risk of more unnatural artifacts...
Markus, sorry for the OT, this has obviously nothing to do with room EQ...
A given impulse shape correspond to a specific amplitude and phase response, and vice versa (FFT).
The steeper the curves the more ringing you get.
If your target is linear-phase then the impulse will be symmetrical and you will get both pre and post ringing, whereas if your target is minimum-phase you will only get (double amount of) post ringing, which is obviously less noticeable on musical signals.
If those filters form a complementary and aligned crossover, then all the ringing (pre and post, regardless of the phase relation) will cancel, and you will be left with a seamless transition.
Of course if those filters do not have counterparts (eg DAC antialiasing filter, subsonic filter, phase linearization of a box response, ...), you will get the entirety of the preringing.
But even when used in a complementary aligned crossover, pre (or post) ringing will strike back when (where) the summation is not "perfect".
Having a perfectly phase-coherent and amplitude-complementary summation is easy enough on-axis, but becomes difficult to impossible as you move off-axis (non coincident drivers, differences in polar response, etc.).
Having steep slopes only makes things worse as you have less "blend" and summation errors become more apparent and abrupt.
The perfect "bad" example is a brickwall filter in a two-way direct radiator loudspeaker, with a woofer that starts to beam and a tweeter that radiates 180°: measured off-axis you will get a distinct discontinuity, and associated ringing.
So the equation is simple: the steeper the slopes the more pre/post ringing your filters have *and* the more likely your crossover is likely to expose it of axis.
Now the question of the audibility of aforementioned preringings remains in debate (not unlike phase distortion audibility), but for sure it is less "natural" than post ringing, and probably quite unnatural (acausality!).
So it boils down to a compromise between the benefits of sharper slopes (distortion reduction) and the risk of more unnatural artifacts...
Markus, sorry for the OT, this has obviously nothing to do with room EQ...
Hey Barleywater, I am here to share and learn. How I got here is another matter. While I enjoy the dynamics of my miniature PA system, I can do better. That started me researching for state of the art, high efficiency speakers, and Earl’s designs comes up as one of the best. From where I sit, I think you owe Earl an apology.
Hello,
Mitchba :
" That, plus time aligning the drivers, and correcting excess phase, appears to produce the most accurate image that I have heard."
I agree.
Curiosly, abx comparing linéar phase vs minimum phase with convpare / rephase ect does not work well with musical program on headphones, but sound stage definition is very clearly improved with near/ mid field speakers test.
crd
Hi jmbee, sorry for the delay to respond to your post. Looks great! Would you be able to share a bit more on how you achieved your results? Cheers, Mitch
jtalden, art, and pos. Thanks for posting your results and thoughts.
Art, the conclusion you pulled from the paper is the conclusion for "minphase" filters. The conclusion for linphase filters is different:
Art, the conclusion you pulled from the paper is the conclusion for "minphase" filters. The conclusion for linphase filters is different:
"With the linear phase crossovers generated by Acourate we get a perfect behaviour. The crossovers add up to a Dirac pulse. So we have reached a first goal in ourspeaker optimization. Based on the thought experiment with assumed perfect drivers we can concentrate on the behaviour of the crossover filters themselves. And it can be shown that linear phase filters allow us to reach the desired function."
That's part of it as there is more to the conclusion, but folks can read the rest for themselves.
jtalden, "Acourate and other software use very different EQ approaches than typical PEQ filters. These EQ approaches are more likely the primary contributor to the improvements you hear. I suggest that since you have optimized your setup using Acourate that you can now use RePhase to create a linear filter with a typical speaker's phase rotation and use it to convolve some of your music selections so that you can then ABX test. In that way you should get a good sense of just how much the direct phase rotation correction portion of Acourate is contributing to the sound. I would suspect the sound is similar, but would be interested to know if that is what you find."
pos has mentioned that "rePhase is not meant to be a room correction software". This is where at least one difference is. See attachment. Note that the excess phase correction FDW parameters are user adjustable for the measurement at the LP. As I adjust the values and generate new filters, and ABX them, I can not only hear the difference, but I can tell which one has the more excess phase correction.
It is going to take a couple of days, but let me convolve some music with different excess phase corrections as the only changing variable and post up and we will see if folks can hear the differences or not. Cheers, Mitch
That's part of it as there is more to the conclusion, but folks can read the rest for themselves.
jtalden, "Acourate and other software use very different EQ approaches than typical PEQ filters. These EQ approaches are more likely the primary contributor to the improvements you hear. I suggest that since you have optimized your setup using Acourate that you can now use RePhase to create a linear filter with a typical speaker's phase rotation and use it to convolve some of your music selections so that you can then ABX test. In that way you should get a good sense of just how much the direct phase rotation correction portion of Acourate is contributing to the sound. I would suspect the sound is similar, but would be interested to know if that is what you find."
pos has mentioned that "rePhase is not meant to be a room correction software". This is where at least one difference is. See attachment. Note that the excess phase correction FDW parameters are user adjustable for the measurement at the LP. As I adjust the values and generate new filters, and ABX them, I can not only hear the difference, but I can tell which one has the more excess phase correction.
It is going to take a couple of days, but let me convolve some music with different excess phase corrections as the only changing variable and post up and we will see if folks can hear the differences or not. Cheers, Mitch
Attachments
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.