Home Theater Forum and Systems banner
1 - 20 of 55 Posts

·
Registered
Joined
·
261 Posts
Discussion Starter · #1 · (Edited)
The avforums review of the KK DXD 12012 has the following statement

Ultimately, if you want a subwoofer (or any speaker for that matter) to have a real kick, snap, hit, insert adjective of choice, you need to ensure that all frequencies it emits, are emitted at the same time and stop at the same time. A complex sound like an explosion, or a drum strike, is composed of lots of frequencies which all start instantaneously. If anomolies in the phase response of the subwoofer cause certain frequencies to start/stop at different points, then the impact of the effect or sound is softened. The measure of how much a transducer slurs it's delivery is it's transient response and it is a distortion of the original signal as serious as any other, but in the time, rather than frequency domain. I'd also note that even the 'harmonizing' CEA 2010 subwoofer measurement standard, fails to take into account this sort of non-linear distortion.
The suggestion is this is just group delay. Am I missing something?

Thanks
Matt
 

·
Premium Member
Joined
·
1,562 Posts
You are correct that Group Delay is a measure of time shift. Be advised that the REW chart can be very difficult to interpret correctly however as the room modes can cloud the result.

Their assertion is that only 2 order (sealed box) SW designs can provide the optimum sound. There are many advocates of this idea, but also many that would disagree.

They are correct that the total 2nd order phase rotation (180°) is correspondingly less than 4th order or 6th order designs and that the lower the phase rotation the lower the time shift.

Note that there are numerous 4th order SW designs (ported box) and even 6th order designs that are very highly rated.

I personally do not think that 2nd order SW design is the most important parameter in achieving the sound characteristics they describe. There are tradeoffs to any design choice and so we could list all the advantages to the 4th and 6th order designs if we wanted.

More important is:
> SPL smoothness over the entire frequency range.
> The timing of the handoff from the SW to the Mains. (this impacts SPL smoothness in the XO range)
> Room acoustics.

REW provides all the necessary measurements to optimize your system given the your equipment/room/setup constraints.
 

·
Premium Member
Joined
·
1,562 Posts
The phase rotation and GD are fixed for a given SW design. We will not be able to make any adjustments or changes to it. We can measure both with REW, but they are for info only.

For a clean measurement we need anechoic space. Outside pointed up or outside with a ground plain mic position will probably be best.

Inside becomes a problem because the room modes will distort the otherwise clean SW signal. We can move the SW as far away from walls use a 1m ground plain mic position. That may provide a slightly better signal to work with, but there is not really much that can be done.

For inside measurements we need to manually filter out the impact of room from the measurement. With experimentation we can learn to do this fairly well.

Using our standard SW and mic locations the signal may be very difficult to interpret depending on the room, but it can probably still be done. Adjusting the window settings can help greatly, but also distorts the chart at the low frequencies if the window is set too short.

This is all difficult to explain and best learned with many examples. If you really want to measure your SW GD and post an .mdat file I will see if I can interpret it for you if you. You may then be able to reverse engineer what is needed. I guess the key is knowing the expected result an then adjusting the window settings to best expose that result. Basically we expect 90° phase rotation for each order.
SW design:
Open baffle - 1st order.
Sealed box - 2nd order
Ported box (bass reflex) - 4th order
High pass filters added to protect the SW from over excursion at low frequency also add the appropriate additional orders.
Knowing the SW design and any additional HP filters and also knowing the cutoff frequency we can pretty much draw the expected Phase and GD curves. Knowing this expected result then we can try to find it in the measurement.

Just to emphasize the main message: Given a SW design the result is fixed. Measuring it is for info only. The other settings I bulleted above do provide capability actually impact the sound characteristics they mentioned. I probably should have listed another one: "> House Curve"
 

·
Premium Member
Joined
·
1,562 Posts
FYI - Perhaps you have seen this.

Here is a interesting experiment that suggests that a linear phase SW is an audible improvement.
http://www.bodziosoftware.com.au/LP_MP_Subwoofer_Tests.pdf

Raczynski tested using a 4th order SW, but also then used DSP software to linearize the phase by first applying inverse EQ to the signal before it's sent to the SW. This becoming more common for those using HTPC setups. Most of the "evidence" is in showing the expected IR and step response differences, but he was also able to hear differences with test signals and "concluded" or maybe just implied that it may be significant for program material as well.

The audibility of phase rotation on program material is well debated with no strong consensus that I have able to determine. My comments above are intended to point out that there are more basic setup issues that have more apparent impact to the sound that need to be given higher priority.

Any unless we have an HTPC setup or other equipment for linear phase, FIR filtering then we have no control over SW phase anyway.
 

·
Premium Member
Joined
·
1,562 Posts
In domestic listening rooms, the room and SW are essentially inseparable; at 1m the reflected sound is typically louder than the direct sound at <100Hz.

With appropriate EQ, type of SW doesn't matter.
Care to elaborate on this?

Is "appropriate EQ" only linear phase EQ? That is, where do you fall on the question of whether one can hear a difference due to high phase rotation?

You've put significant effort achieving linear phase for you system. Do you only find that beneficial for the MF and HF?
 

·
Registered
Joined
·
261 Posts
Discussion Starter · #10 ·
Given a SW design the result is fixed. Measuring it is for info only.
Thanks for the comprehensive reply. One thing confuses me immediately and that is the comparing the quote above against this earlier quote when commenting on "things that are important to SQ"

The timing of the handoff from the SW to the Mains
from a measurements perspective, if aligning manually, aren't these 2 points addressing the same concern?
 

·
Premium Member
Joined
·
1,562 Posts
The innate phase rotation of a SW is fixed depending on its design. In pointing this out I was thinking of the article you referenced. They suggested their SW design was advantageous because it had low phase rotation. If so, then all such 2nd order SW designs then have the same characteristic. This point is probably not your question or interest, but it was my main takeaway from the article. You did quote from it, but I was not sure why.

Now, in thinking of system setup given a SW (no matter the design), the issue of the handoff between the SW and mains is the main concern. We can use REW to adjust the timing by using it to set speaker distance settings in an AVR. That impacts the smoothness of the SPL through the XO area. REW provides the capability to set this timing in several different ways. Using GD is one, but I find that one to be the most problematic due to the impact of the room response on the chart.

For those interested in getting good results with a minimum of time and effort the RTA method is best choice.
 

·
Registered
Joined
·
261 Posts
Discussion Starter · #12 ·
This point is probably not your question or interest, but it was my main takeaway from the article. You did quote from it, but I was not sure why.
just for context basically, the quote I placed in bold was the one I was seeking clarification on exactly what that means from an objective measurement perspective.

The proximate cause of my question is "what is the most effective way to setup my system such that I can enjoy an optimal response (given room constraints)?". The ultimate cause is that I am curious & that going into the detail of loudspeaker response feels somewhat like peeling an onion blindfolded :) but once an itch is scratched then it can stay scratched for a while.

By this I mean there is a v large gap between the basic summary information (which at best relies on practical methods useful to the layman & at worst is a mass of old wives tales and half truthes) & the underlying detail (which is clouded by psychoacoustics and/or manufacturer claims and/or the difficulty of the subject itself and/or the relatively unfamiliar and unexplained terminology). It feels like this gap is started to close as I become somewhat more familiar with the terminology and read more though.

The relevance of sealed/ported is really just because I am imminently moving from a large ported SVS sub to a larger (in driver/power terms) sealed sub.

Thanks for your replies (to this and other threads) btw, I was considering trying the method you described in the thread from a while ago about using a narrow band sweep (XO +/- 3Hz IIRC) on sub & mains for checking alignment using impulse response as a next step. I just need to get time to actual do some measurements as opposed to reading about what measurements to take.
 

·
Premium Member
Joined
·
1,562 Posts
If you go beyond the basic RTA setup to learn more what the actual phase impact is, I can advise or assist if you like.

If you use the narrow IR method suggested by Bill (laser188139), I will be very interested to see how well it works. I have only tried that method once and had good results.
 

·
Registered
Joined
·
261 Posts
Discussion Starter · #14 ·
If you go beyond the basic RTA setup to learn more what the actual phase impact is, I can advise or assist if you like.

If you use the narrow IR method suggested by Bill (laser188139), I will be very interested to see how well it works. I have only tried that method once and had good results.
that is a most generous offer and I will be happy to take you up on it. Would you recommend starting with the narrow IR method or an alternative method?
 

·
Premium Member
Joined
·
1,562 Posts
> What is your AV equipment?
> Do you have the capability to do loopback measurements?
> Are you intending to confirm or fine tune the SW timing of an automated setup program like Audyssey or is this a manual setup?
> Will you be doing manual SW PEQ?
> Anything else I should know?

[I need to have a better idea of scope of the job before answering.]
 

·
Registered
Joined
·
261 Posts
Discussion Starter · #16 ·
> What is your AV equipment?
> Do you have the capability to do loopback measurements?
> Are you intending to confirm or fine tune the SW timing of an automated setup program like Audyssey or is this a manual setup?
> Will you be doing manual SW PEQ?
> Anything else I should know?

[I need to have a better idea of scope of the job before answering.]
The relevant AV equipment includes

  • Marantz AV 7005 processor (has audyssey multeq xt)
  • BFD 1124 for PEQ
  • SVS PB 12 Ultra sub
  • M&K S85 LCR
  • M&K Surround 55 Mk2 tripole surrounds

In the next few weeks, the sub/speakers will be replaced by a 15" sealed sub powered by an inuke, MK MP150Mk2 speakers across the front & MK S150T tripoles for surrounds.

For measurement, I have

  • art usb dual pre mic preamp
  • Dayton EMM6 from CSL
  • PC connected to the processor via HDMI (with ASIO working) or laptop connected via front analogue inputs

I am familiar with operating REW having used it for years for EQing frequency response, venturing into time domain measurements is relatively new to me.

I am mildly suspicious of audyssey. I find it fails to eq the sub anywhere near adequately hence the BFD so current method is to EQ the 2 main modes and then run audyssey. Previous measurements suggest audyssey is aligning LCR accurately but that the surrounds are way off. I find it fails to do much for the crossover region though room simulation suggests this is a function of the MLP. The MLP will be moving forward slightly away from the rear wall in the next few months and simulation suggests this should help. I intend to take some measurements to confirm or refute this (though the MLP is going to move forward anyway due to new sofa).

Therefore the intention is first to confirm/tweak audyssey and then to enable me to accurately setup this up manually so I can do a proper comparison of audyssey vs my manual setup.

As to anything else you should know... room size/construction perhaps? The room size is not unusual for this sort of Victorian house in the UK (4.1m x 3.6m x 3.05m) with suspended wooden floors and brick walls. Flooring is currently wood but soon to be replaced by a carpet. Room treatment is largely impractical due to size of room and mixed use. LCR are horizontally aligned underneath the projector screen, sub is positioned to the right of the R speaker in the corner. There is a pic in this post

Anything else?
 

·
Premium Member
Joined
·
1,562 Posts
Okay thanks!
The only thing there that concerns me a little is providing a manual EQ for the SW and then running Audyssey on top of it. This may work okay, but any personal house curve will be removed and replaced with Audyssey's house curve. Would we then manually EQ again on top the initial BFD settings?

I would tend let Audyssey do its EQ first and then confirm/adjust the SW to Main timing and then follow that with any additional BFD EQ that is needed. We should also keep in mind that the experts suggest that it is not a good idea use narrow filters to try to remove narrow dips from room modes.

Okay back to the subject.
It sounds like you intend to make several setups and compare the results. We can pick just one that we work together and then you can do the others.

Options:
1. Setup using RTA. We can then still run sweeps to just see/confirm the Phase and GD results.

2. Setup using the narrow IR method. We can then still run sweeps to just see/confirm the Phase and GD results.

3. Setup using Phase to first select the which SW to Main timing we prefer to target.

The first is the most straight forward and normally used by those that want a well setup system but not wanting to delving into all the gory details. The second is likely to find the same result, but I am not sure of that as I have little experience with it. For either of these the optional confirmation full range sweep measurements will still show us the actual phase alignment we found. The third option is more complicated as we spend time looking at optional phase/GD alignment tradeoffs before we chose one to target. The first two try to target a conventional alignment. The third provides more insight into the Phase GD tradeoffs and provide the opportunity to setup and evaluate other options.

You can chose which you want to do. Just be advised that the third option can get pretty confusing as can be seen in this thread. My involvement started at Post 129.
 

·
Registered
Joined
·
261 Posts
Discussion Starter · #18 ·
The only thing there that concerns me a little is providing a manual EQ for the SW and then running Audyssey on top of it. This may work okay, but any personal house curve will be removed and replaced with Audyssey's house curve. Would we then manually EQ again on top the initial BFD settings?
I used to do that but don't at the moment, I just eq to flat so that audyssey has little to do.

Just be advised that the third option can get pretty confusing as can be seen in this thread. My involvement started at Post 129.
I will read that thread then decide if that's ok.

Cheers
Matt
 

·
Banned
Joined
·
4,835 Posts
Backtracking just a wee bit on the thread, which I have followed with fascination...

I try to be objective about what we hear and what we think we hear. Phase response and the use of linear phase filtering and/or design approaches is something that SEEMS like should be more important because it SEEMS like we should be able to hear it. That is probably mostly wishful and idealistic thinking on my part. As I stated in post #6, I know that relative phase/timing between speakers at mid & high frequencies affects imaging. And I have heard speakers that had a punchiness where percussive instruments really pop out at you - and have imagined that tight phase coherence made that possible - but alas have no proof for it.

In the review referred to in post #1, the writer says "if you want a subwoofer (or any speaker for that matter) to have a real kick, snap, hit, insert adjective of choice, you need to ensure that all frequencies it emits, are emitted at the same time and stop at the same time." Speaking with no authority at all here, but from an intuitive/aesthetic standpoint, it seems like any frequencies contributing to "snap" or "kick" in a sound are going to be well above subwoofer frequencies. A snappy kick drum or plucked bass string can have frequencies well into the Khz range, most likely where the "snap" is being experienced.

In the article referred to in post #5, the author says he heard an "audible buzz" (MP design) in one case, and what was more like a "pop" or "click" (MP design) than a thump (LP design) in another. Almost sounds more like something going nonlinear and producing higher frequency distortion components than a sound that can be attributed solely to phase shift.

Bottom line, as far as I can gather from all this: I am having a hard time being convinced that LP design in a subwoofer IN ITSELF is a super important factor. No harm in it whatsoever, and as a somewhat obsessive perfectionist and "I think I can hear it" phase-response wishful thinker, I would probably go for it all other factors being equal - those were some mighty impressive measurement waveforms in that article, that HAS to sound better, right? But in reality, paying due attention to actual psychoacoustic research, other factors like integration with main speakers at crossover and overall frequency response - as already pointed out - seem like far more important factors in the real world.

My 2 cents worth.

jtalden: Thanks, as always, for your in-depth and valuable contributions.
 

·
Registered
Joined
·
261 Posts
Discussion Starter · #20 ·
I speak here with no authority as I understand v little of how the auditory system actually works. Furthermore one can read reams about hifi/av and essentially come away none the wiser because the authors are, most likely, similarly uneducated even if they are experts in the mechanics of how a speaker functions. This produces the subjectivity common to most reviews & discussion forum and explains why so many threads are derailed into what are essentially "he said, she said" arguments.

I think this is anathema to the more objectively minded individuals hence the desire to investigate through objective measurements. Whether this directly contributes to a more enjoyable audio experience is hard to say, if not impossible, given that there is no fixed target to aim for but it surely *indirectly* contributes because it leads one to more critically appraise/compare different setups and hence one becomes more aware of what is better/worse.

To give one example, I am currently pursuing the understanding to be able to successfully time align my setup by hand yet one source I read says that humans enjoy the "fullness of sound" presented by a multi sub setup that is naturally (unless you place them miles apart) slightly out of time alignment. Is this true? I have no idea. Is it true but is a side effect of some other thing that appears to derive from multi sub setup? Again, no idea.

If anyone has any recommendations of a good laymans "this is how we think the auditory system works" (popular science basically a la stephen pinker) then that would be interesting to read.

A somewhat OT ramble there, back to the 21 page thread on phase alignment!
 
1 - 20 of 55 Posts
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top