• Welcome to ASR. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Understanding the state of the DSP market

Seems like the speakers being sealed is really the only necessity for ease of integration. A ported sub only has phase/GD issues around the port's tuning frequency, which should be well below where it works be interacting with the other speakers. Or am I overlooking something?

Yes, if you look at it from purely subs-mains integration perspective you are right and there are lot of fine ported subs out there and you will be getting benefits of ported enclosure in terms of bass extension, sensitivity and size.

My reasons to go for sealed were multifold:
- lot of subs in my system - with 8 subs in relatively small room [rear subs approx 2m behind seating position] port noise would be audible
- room walls are 30 cm concrete - I get room gain below 1st modal frequency of almost 6dB, so “room is my bass reflex” ;-)
- it’s easier to integrate 8 subs between them, if they are sealed [on the low end]
- its easier to integrate infra subs later or combine several types of subs, if you intend to build your system gradually [which I do]

As some have mentioned, maybe even more important id the quality of internal DSP - quite often you have subs being pushed to the limit for couple of Hz if low end extension to pimp up specs and you end with 12 in woofers and 500W amps “18Hz in room extension” This is why I have decided also to go for passive subs, dumb amps and shallow enclosure [to have as much as possible boundary gain]. I handle all the DSP via processor [including HPF, limitations etc] and it allows for better thermals and cable management.

But overall, my philosophy - do as much as you can in “analogue domain” - have good speakers with proper off axis behavior, place them properly, have as much excess capabilities, ideally absolute overkill in FR extension, SPL levels, watts etc - e.g. if you want to cross over at 80Hz make sure your speakers are good down to at least 40Hz and 110dB ;-) and most importantly have your room properly treated. Then just add some finishing touches by DSP.

They are not playing it fair though since they are exclusively using MIMO capabilities of the receivers. To you and me, only a graphic equalizer (not even parametric) is available and even that has been obliterated by removing any equalizer for subwoofers from the equation.

This is my system and my measurements. And I have nothing else than my individual listening impressions to support the statement, that there is audible difference between them. Both FR charts were almost identical.
 
You're right, it's hard to find this kind of information since those who know often don't share it, and I don't remember the exact source, but I recall seeing that number in a forum discussion years ago. I later tested and confirmed it myself with many inverted phase filters (which are essentially "compensate" mode all-pass filters in rePhase, or can be generated in REW with gated inversion) in my own system. More recently, I also discovered that Dirac uses 4.5 ms, and even Audyssey left a 4.5 ms buffer seemingly intended for inverted all-pass filters, though it appears they decided against implementing it in production.

You can easily test this yourself (just not with headphones, of course). Pre-ringing, or pre-echo in most cases, is quite audible with the right track.

Can you show me how you would visualize or determine such in an REW plot? You can use the subwoofer impulse response recorded from my desk LP I uploaded earlier as the test "subject".


1747498615892.jpeg

changing Q and center freq values in "real time"

*BTW, FIR filter/config for the sub only contains the phase EQ as the IIR EQ is kept separate via JRiver's DSP parametric plugin


I mentioned it before, but I do test for audibility primarily using a cycling kick drum beat in continuous replay which extends in the region where center freq. of the sub GD peak is ~83.5 - 84 Hz -- naturally shifts with position change. When listening to that subwoofer driver by itself at the desk MLP, it really is quite easy to hear a marked difference when convolving an inverse phase all pass filter in the chain. It is also possible to hear a very subtle difference when incrementing or decrementing by about 0.2 Q steps. To me, the initial hit/strike is much more apparent without the all pass inversion, but the tone of the kick seems quite a bit off. With the phase inversion, the very initial hit/strike definitely can be described as less apparent but the overall tonality actually sounds much closer to "correct" and even just a tad bit louder overall -- now, I know from numerous previous experimental testing with other filter variants (for better or worse) and system setups (couch vs desk LP) it can go the other way around -- so, there's really no point in me making definitive statements.

It gets complicated when you add additional drivers for the purpose of bass management mixing/upmixing -- in this case, the LCR channels (muted the surrounds) -- as any hints of such "pre-echo" artifact that's easily heard only with the sub in isolation becomes masked furthermore. In fact, the "off-ness" of the tonality in the min-phase version sounds a bit worse in my desk setup. The FIR linearized version actually sounds better and more focused overall. When critically listening through that kick drum test track, I can still hear vague hints of "pre-echo", of course, but in actual practice listening with real music at leisure even with a clear, strong bass line it's also completely inaudible.

In terms of perceived "loudness" when equalized with the same summed magnitude gain, it kind of really depends with the specific test track or music passages being listened to... The bass line in some music test tracks sound a bit louder in the min-phase version vs the inverse phase linearized version and vice versa. Again, it kind of ends up being a random toss thing so doesn't really matter.

I have different priorities for different listening sessions so I keep multiple DSP presets saved. Unfortunately, there is no ultimate champion DSP filter settings that comes out the winner for me here.


-----------


A couple more listening notes:

The completely min-phase sub on its own ends percussion notes a bit softer and also sounds more localizable -- I can tell it's coming from the floor in the front wall where the sub is physically sitting. The phase inverse EQ'd sub sounds harder at the tail end (seems pre-causal slippage is slightly softening the leading edge) although it's also a bit less localizable -- listening intently, I can tell the origin is in the front, but the bulk of the impact itself feels more forward: almost approaching in or between my head kind of feeling.

Finally, together with the LCR main speakers, the sense of focus/coherence and solidity of percussive hits on the whole is readily more apparent in the phase linearized version. I mean, the minimum phase version is still very good sounding -- no doubt there's not really much to complain, BUT... with the main speakers playing, it just is missing a little bit of extra "hardness" to the percussive kicks. I suppose that's "icing on the cake" as they say... Nevermind the extra latency and slight hint of pre-causal slippage detectable when one is really, really paying close attention -- which I'm not 99.9% of the time anyway. DSP presets with built-in inverse all pass phase compensation/correction is worth it in my own setup despite the hassle involved in testing and setting up.
 
Last edited:
Is not about distortion and harmonics, but about fundamental frequency of a note and harmonics, that make the timbre of the instrument or voice. E.g. this is frequency spectrum of open E bass note. You want them all to arrive at the same time.

View attachment 451398

Actually Group Delay is one of the more important things to get right if you want really high performant system, especially below 100Hz. Usually sealed speakers behave much better in that respect and ported speakers trade-off their bass extension for worse Group Delay.

To get FR right is trivial.
Yes, I was imagining a graph like that in my head when I was trying to work out what some of the problems might be when you have a different delay for the lower frequencies. I think for sure it's not ideal. I imagine there's a point where you wouldn't detect a difference, but like people have been saying it seems that there's not much study on audibility of Group Delay.
Group delay audibility, especially at lower frequencies, is not settled science unfortunately. The most common concern with Group Delay is with something like a "drum hit" where ideally you would want all the frequencies that are part of the drum hit to play at the same timing as original so you get "impact" of the original. You also bring up some interesting points about how group delay would interact with distortion which I have never considered. The safest and most reasonable conclusion is that minimizing distortion and group delay is a reasonable way to improve sound quality as well as making integration easier. The levels of distortion and group delay we are dealing with for many subs is quite high and is not at all like chasing SINAD from 100 dB to 120 dB.
Yes, I can see how Group Delay is not ideal to have in your sub, but yeah we don't know the cut off points for what's acceptable. But yes, I do now agree that if you're chasing the best sub you can get then you would also take Group Delay into account along with the more common attributes of distortion percentage & frequency response that we use here to assess speakers & headphones. It might well be found though that high distortion subs have quite a lot of group delay too if the low range has been extended with DSP so praps in a lot of cases people can do pretty well by just following distortion & frequency response, but I can see now that if you wanted to chase the best then you'd also look at Group Delay to be sure.
 
Can you show me how you would visualize or determine such in an REW plot? You can use the subwoofer impulse response recorded from my desk LP I uploaded earlier as the test "subject".


View attachment 451639
changing Q and center freq values in "real time"

*BTW, FIR filter/config for the sub only contains the phase EQ as the IIR EQ is kept separate via JRiver's DSP parametric plugin


I mentioned it before, but I do test for audibility primarily using a cycling kick drum beat in continuous replay which extends in the region where center freq. of the sub GD peak is ~83.5 - 84 Hz -- naturally shifts with position change. When listening to that subwoofer driver by itself at the desk MLP, it really is quite easy to hear a marked difference when convolving an inverse phase all pass filter in the chain. It is also possible to hear a very subtle difference when incrementing or decrementing by about 0.2 Q steps. To me, the initial hit/strike is much more apparent without the all pass inversion, but the tone of the kick seems quite a bit off. With the phase inversion, the very initial hit/strike definitely can be described as less apparent but the overall tonality actually sounds much closer to "correct" and even just a tad bit louder overall -- now, I know from numerous previous experimental testing with other filter variants (for better or worse) and system setups (couch vs desk LP) it can go the other way around -- so, there's really no point in me making definitive statements.

It gets complicated when you add additional drivers for the purpose of bass management mixing/upmixing -- in this case, the LCR channels (muted the surrounds) -- as any hints of such "pre-echo" artifact that's easily heard only with the sub in isolation becomes masked furthermore. In fact, the "off-ness" of the tonality in the min-phase version sounds a bit worse in my desk setup. The FIR linearized version actually sounds better and more focused overall. When critically listening through that kick drum test track, I can still hear vague hints of "pre-echo", of course, but in actual practice listening with real music at leisure even with a clear, strong bass line it's also completely inaudible.

In terms of perceived "loudness" when equalized with the same summed magnitude gain, it kind of really depends with the specific test track or music passages being listened to... The bass line in some music test tracks sound a bit louder in the min-phase version vs the inverse phase linearized version and vice versa. Again, it kind of ends up being a random toss thing so doesn't really matter.

I have different priorities for different listening sessions so I keep multiple DSP presets saved. Unfortunately, there is no ultimate champion DSP filter settings that comes out the winner for me here.


-----------


A couple more listening notes:

The completely min-phase sub on its own ends percussion notes a bit softer and also sounds more localizable -- I can tell it's coming from the floor in the front wall where the sub is physically sitting. The phase inverse EQ'd sub sounds harder at the tail end (seems pre-causal slippage is slightly softening the leading edge) although it's also a bit less localizable -- listening intently, I can tell the origin is in the front, but the bulk of the impact itself feels more forward: almost approaching in or between my head kind of feeling.

Finally, together with the LCR main speakers, the sense of focus/coherence and solidity of percussive hits on the whole is readily more apparent in the phase linearized version. I mean, the minimum phase version is still very good sounding -- no doubt there's not really much to complain, BUT... with the main speakers playing, it just is missing a little bit of extra "hardness" to the percussive kicks. I suppose that's "icing on the cake" as they say... Nevermind the extra latency and slight hint of pre-causal slippage detectable when one is really, really paying close attention -- which I'm not 99.9% of the time anyway. DSP presets with built-in inverse all pass phase compensation/correction is worth it in my own setup despite the hassle involved in testing and setting up.
My experience is similar, pre-echo can get away undetected with many tracks until it doesn't with one particular track and unfortunately these are often the better sounding filters otherwise. Btw, I couldn't look too deep but didn't see anything that would cause pre-echo in the filters in your mdat. Everything is silenced by more than 43dB already before 5ms:
1747547724677.jpeg
 
I'm slowly reading through the thread. This seems different from REW's AutoEQ. Is it something that can only be done in Acourate?
 
REW and Acourate are at the opposite end of the spectrum. Both are very manual tools, and the tools need to be deployed in the correct order and in the correct situation. When I say "manual", I mean that you have to inspect the measurements yourself and decide what you want to do. Acourate has a few more "luxury" features compared to REW, which is why I prefer it. REW can not be used on its own, you need RePhase. Both are extremely flexible, but also close to impossible to use if you do not know what you are doing.
It seems like REW doesn't have advanced phase correction features like those you mentioned in the thread. So, I was wondering if these features are only available in that specific Acourate software.
 
It seems like REW doesn't have advanced phase correction features like those you mentioned in the thread. So, I was wondering if these features are only available in that specific Acourate software.

REW does not, but rePhase does. If you really know what you are doing, you can do 90% of what Acourate does without paying any money. But it will take you a looong time because Acourate automates some processes and there is absolutely no automation in REW. One example, suppose you have a 4 way stereo system (8 drivers). You take a measurement, invert over a target curve, and come up with a final correction. Now you have to make a .WAV file for each driver. In Acourate, it's one button push and presto, you get 8 .WAV files along with a config file for Roon. With REW, you manually convolve the correction with each crossover, adjust the level, then save the .WAV file with the correct tap length and impulse position, and repeat 8 times. Then write the config file for Roon yourself. Even then there are still features missing, e.g. no pre-ringing compensation. It's not impossible, but it's tedious and repetitive.

This is because REW was not written to design linear-phase FIR filters. It has to be adapted to do so with third party software. OTOH REW is excellent for designing biquads for IIR, much better than Acourate if that's what you want to do.

In Acourate there are a few ways to correct phase. Some of them are automatic, e.g. excess phase correction and driver phase linearisation. Some use a special tool, e.g. matching the phase between left and right. Or if you want, you could forego all that and do it completely manually. No need to export to another program. I have found that you can do the same corrections with rePhase, but you really need to understand the steps and execute them in the correct order because it's just a tool, there is no guidance.
 
REW does not, but rePhase does. If you really know what you are doing, you can do 90% of what Acourate does without paying any money. But it will take you a looong time because Acourate automates some processes and there is absolutely no automation in REW. One example, suppose you have a 4 way stereo system (8 drivers). You take a measurement, invert over a target curve, and come up with a final correction. Now you have to make a .WAV file for each driver. In Acourate, it's one button push and presto, you get 8 .WAV files along with a config file for Roon. With REW, you manually convolve the correction with each crossover, adjust the level, then save the .WAV file with the correct tap length and impulse position, and repeat 8 times. Then write the config file for Roon yourself. Even then there are still features missing, e.g. no pre-ringing compensation. It's not impossible, but it's tedious and repetitive.

This is because REW was not written to design linear-phase FIR filters. It has to be adapted to do so with third party software. OTOH REW is excellent for designing biquads for IIR, much better than Acourate if that's what you want to do.

In Acourate there are a few ways to correct phase. Some of them are automatic, e.g. excess phase correction and driver phase linearisation. Some use a special tool, e.g. matching the phase between left and right. Or if you want, you could forego all that and do it completely manually. No need to export to another program. I have found that you can do the same corrections with rePhase, but you really need to understand the steps and execute them in the correct order because it's just a tool, there is no guidance.
Keith my friend, I will have to kindly disagree with you on that one. There's nothing REW cannot do that rePhase does (except for saving filters in .bin format for MiniDSP). I agree rePhase makes it a lot easier to fiddle with phase but all of them are doable in REW. Feel free to challenge me :)
 
Last edited:
Keith my friend, I will have to kindly disagree with you on that one. There's nothing REW cannot do that rePhase does (except for saving filters in .bin format for MiniDSP). I agree rePhase makes it a lot easier to fiddle with phase but all of them are doable in REW. Feel free to challenge me :)
Not a challenge but a question. How can you get REW to create linear phase filters without Rephase?
 
Not a challenge but a question. How can you get REW to create linear phase filters without Rephase?
Any REW measurement (or filter) you save as .wav in REW is a FIR filter ready to use in a convolution engine - linear, min or max phase. Specialty of FIR filters are not their linearity, it's their capacity to be non-linear if needed.
 
Keith my friend, I will have to kindly disagree with you on that one. There's nothing REW cannot do that rePhase does (except for saving filters in .bin format for MiniDSP). I agree rePhase makes it a lot easier to fiddle with phase but all of them are doable in REW. Feel free to challenge me :)

After you mentioned in the other thread that I can create AP filters using the EQ tool, I tried it. But it's almost impossible to design an AP filter without the phase response on the screen in front of you as a reference. Unless there is a better way to do it in REW (you can chat to me offline about that) I think rePhase is the superior tool for correcting phase.
 
After you mentioned in the other thread that I can create AP filters using the EQ tool, I tried it. But it's almost impossible to design an AP filter without the phase response on the screen in front of you as a reference. Unless there is a better way to do it in REW (you can chat to me offline about that) I think rePhase is the superior tool for correcting phase.
Yellow: response after allpass is applied, green: the allpass filter itself:
1748026280949.png
 
In the EQ window the graph below the filter adjust graph can be selected as Phase, showing measured and predicted phase.
Hm, I spent some hours with REW, but I have no idea to look for this "show as phase" option.
Is this in a submenu to the preferences of the EQ window or in the preference to a sub window to the EQ dialogue of the EQ window? Actually I could find it in neither place.
Can someone give directions?
 
The lower graph of the EQ window, here showing the waterfall, can also show impulse response, phase, group delay or a pole-zero plot.
Thank you!
The lower part was hidden for me (probably since ever) and the tiny bar did never catch my attention.
 
There's nothing REW cannot do that rePhase does
I agree with OCA's opinion.

(except for saving filters in .bin format for MiniDSP)
1748044628059.png

1748044643028.png

1748044656316.png

I've seen some users export WAV files, then import them as raw files into Audacity and change the file extension to .bin to use them.
I haven't been able to test it myself since I no longer have a MiniDSP.
So, I'm not really sure if that actually works.


------------------------------------------------

By the way, A few simple examples.
(I still don’t fully understand the API (always grateful to @JohnPM for REW), but users like @OCA automate it and use it effectively for their purposes. - So, if you have a good knowledge of API, you can either automate it or do it manually.)


1748045521243.png


1748045791854.png

1748046781750.png
 
Last edited:
My experience is similar, pre-echo can get away undetected with many tracks until it doesn't with one particular track and unfortunately these are often the better sounding filters otherwise. Btw, I couldn't look too deep but didn't see anything that would cause pre-echo in the filters in your mdat. Everything is silenced by more than 43dB already before 5ms:
View attachment 451735

I still find it a bit difficult to tell with absolute certainty just looking at the (ETC above) graphs to ascertain pre-ringing/echo audibility -- esp. coming from single-point measurements only.

Actually, minimum phase traces (included above) all have zero pre-ringing potential -- unusually, perhaps, the subwoofer starts a tad bit early, but there is no pre-causal artifacts whatsoever. This can be tested when measuring at multiple spatially spread out microphone positions.
 
Keith my friend, I will have to kindly disagree with you on that one. There's nothing REW cannot do that rePhase does (except for saving filters in .bin format for MiniDSP). I agree rePhase makes it a lot easier to fiddle with phase but all of them are doable in REW. Feel free to challenge me :)

I mostly agree...

1748057198572.png


Doing everything in REW isn't always the most convenient and there are some unique GUI previewing (drag and drop files and automatically see filtering effect) options that are only found in rePhase e.g. 'energy' centering and filter 'optimization' -- very useful when trying to get as good a fit as possible and minimizing rippling/slippage while using fewer taps.
 
  • Like
Reactions: OCA
Back
Top Bottom