I have been following this thread from the outset, but.... May I pose a stupid question? If I had to place a reputable speaker in a well designed rigid box and play some music through it containing well defined bass, whatever your preferred genre in a well treated room, would you be able to tell whether the speaker has Q=0.5 or Q=0.707 by listening to it.
No AB comparison, just one speaker. If you could then please advice what I must listen for and how to interpret the sound.
I have three very different speakers, very small IB, particularly large BR and a 1.2 m tall TL. Each of these has quite well defined and punchy bass but a different presentation completely.
The small speaker has beautiful textured bass, but it sounds, well small. The BR is really quite authoritative, but shakes everything including the roof, still acceptable but can make you anxious in a way. The transmission line is, well flat, very neutral with good controlled bass not thick not thin but not interesting nor exciting.
All the drivers comes from the same manufacturing company. Both BR and TL have identical woofers mid and tweeters, the small speaker actually utilizes the same mid as a woofer and identical tweeters to the other. I have to place the speakers in a particular area when listening, none in the same spot but differently. If of any importance, the small IB and the BR speaker dates from 1970s, the TL from the 80s All three designed by very influential and reputable designers/teams of designers.
My question is how do I know these people knew what they were doing and understood all the issues discussed in this thread - how can I tell without measuring the step response which speaker should perform the best and most accurate following the input. My genre is mainly rock from the mid sixties to early eighties, some orchestral and lesser folk.
No AB comparison, just one speaker. If you could then please advice what I must listen for and how to interpret the sound.
I have three very different speakers, very small IB, particularly large BR and a 1.2 m tall TL. Each of these has quite well defined and punchy bass but a different presentation completely.
The small speaker has beautiful textured bass, but it sounds, well small. The BR is really quite authoritative, but shakes everything including the roof, still acceptable but can make you anxious in a way. The transmission line is, well flat, very neutral with good controlled bass not thick not thin but not interesting nor exciting.
All the drivers comes from the same manufacturing company. Both BR and TL have identical woofers mid and tweeters, the small speaker actually utilizes the same mid as a woofer and identical tweeters to the other. I have to place the speakers in a particular area when listening, none in the same spot but differently. If of any importance, the small IB and the BR speaker dates from 1970s, the TL from the 80s All three designed by very influential and reputable designers/teams of designers.
My question is how do I know these people knew what they were doing and understood all the issues discussed in this thread - how can I tell without measuring the step response which speaker should perform the best and most accurate following the input. My genre is mainly rock from the mid sixties to early eighties, some orchestral and lesser folk.
Last edited:
Hmmmm... I will continue to study your work, it is interesting.
I would humbly suggest this. It seems the control is supposed to represent a bass pluck at 30hz? But is not an actual bass pluck? What is it then? It seems you want to see steady state performance so I'd guess its just a Sine wave at 30hz with the amplitude over time, adjusted to mimic something that might come out of a stand up bass. In your report you say that Q=1.0 is identical to the input, but that is untrue. Resonance brings group delay and I see the second oscillation in the test having higher amplitude than the first. Note worthy, as to what happens as Q increases.
It does appear to be a simple sine wave compare n contrast, input vs source. In my opinion, you are overcomplicating the issue of spl. All you need to do is match the start up transients of each test situation. You don't have to eq, to level match, you use gain. If you use a peak filter to change levels at Fc you just changed Q and group delay. The group delay present in the Q=1 sample throws a crook in the game spl matching. Time to max spl is not 0ms so matching the start up transient further skews things. It might be more productive to match spl of the sine waves much later in the signal.
The more I study your results the more obvious it is to me that Q=0.5 is more accurate and that the other samples are simply closer to input due to the alleviation of the attenuation caused by the high pass filtering allowing the signal to play closer to the amplitude shown in the input. 0.7 and 1.0 both have distortion in amplitude with the first 2 oscillations, group delay. Q0.7 second oscillation matches the amplitude of the first oscillation, and Q1.0 second oscillation is higher in amplitude than the first... everything else seems to be related to amplitude of input vs output, rather than accuracy.
Whats funny about this topic is, it only affects the signal at and around the filter knee. Q, as in damping, changes with frequency and Spl but in regards to Q and F, Q is much lower in the parts of passband that has no drastic changes in level. Less Q, Less group delay, sooner to max spl, sooner to rest.
Even more reason to create situations where shallow roll offs can be used with plenty of headroom leftover or roll offs that occur below content.
Seems Qes is so much more important.
Last edited:
@Maxine01 Can you please outline why you might consider that the step response and group delay of a Q=0.5 system are better than that of a Q=0.7071 system? It's not entirely clear why that would be the case. Maybe you can provide some objective data to clarify the situation.
Last edited:
I did do something along those lines. However, it wasn't quite enough. The trick is to adjust the amplitude to ensure that the power contained in the two signals was more or less identical. As this condition was approached, any discernable audible differences between the two sets of response simply disappeared.
A gain change produces essentially the same result. The EQ approach was just another option that came to mind. It wasn't meant to be definitive in any way, shape or form.
That's all very true, and I was aware of it. But, if we get the power output levels similar, then the results sound very, very similar.
And that's the entire crux of the problem. The attenuated levels are the dominant psychoacoustic phenomenon at play. If I adjust the levels to be the same, the sounds of the two systems become indistinguishable from each other. The "Q=0.5 is better" doesn't seem to stand up to scrutiny.
It's tricky doing these types of comparisons, isn't it? I'm glad that I'm not the only one that had some trouble attempting to do so.
In this thread, a guy shows the frequency response, step response, and theoretical group delay for Q factors of 0.5, 0.707, and 1.0 of a second-order system:
Theoretically, the change is small, but a Q factor of 0.5 to 0.7 is really a fine adjustment that can make a difference for some people, for others not so much.
It's too late where I live now, so tomorrow I'll try to better explain my opinion on this small Q-factor change.
The images are from his post.
Attachments
These step responses and stuff is just plain nonsense and I will just continue to listen to things I like. No one could give me an indication of how speakers of different Q will sound. So I wasted my time on this thread, it is superficial and plain nonsense. Keep up the Q work and step responses guys. I will stick to electronics.
The Rogers LS3/5a I have is specified as a Q of 1.2 which produced a bump in the mid bass. This was done to balance the bass to the rest of the sound spectrum since a flat speaker sounds lean if it's bass resonance is high like the LS3/5a. And is probably the best sounding speaker in my collection.
Hi Perry,
In what it is different from Keele's 6th order assisted alignement?
Well i could give you an indication of how i feel they sound but words being words and as you are not me i'm not sure it will convey anything constructive.
And we don't know what you listen to, not music genre, but in general. We all favour different kind of renderings so the best loudspeaker on earth for someone could be the worst for his neighbour...
If you are into electronic more than electroacoustic then i suggest you to build a LT circuit ( if you don't want to play with eq on a computer which is way faster) and discover by yourself, nothing beat dfirst hand experience.
If you don't get why step response is important or the acoustic part of a loudspeaker then sure it can seems like non sense. Too bad you seems to be waiting for absolute answers. There is none, it's all relative.
If the ls3/5a is the better loudspeaker for you then it gives some clues about what you like, at least some guess.
It's a mini monitor so was developed by BBC to check the medium range in remote facility and difficult 'room' ( control room implemented into trucks). So it's 'credible' in the critical vocal range ( 300hz/3khz), outside this range it's either faked ( in bass using a trick to make you think there is some there) or of poor interest as iirc FM ( bandwidth 50hz/15khz) didn't even exist when they were developed and their initial program was broadcast monitoring of the days.
So either your preference is for low bass music contents ( which it seems to be given the music genre you listed) either for mid focused reproduction. No judgement on this, it's your preference, it's perfectly fine.
Mini monitor are tools i use as they focus on the most important range when i work but they don't give a full view of what you track or mix, that's why we have mains in studio, which are widebandwidth high dynamic system which can (theorically) reproduce any acoustic instrument at it's natural level and without compromising there bandwidth either.
Try reproduce a timpani or a double bass at same acoustic level than original instrument on ls3 and see if it's close to the original, despite a double bass isn't a 'loud' instrument.
If you want to approach this i would suggest same thing as for Maxine: some sawdust to build a sub, a car audio sub in a sealed cabinet and from there modify qtc to see which one is the closest to the real instrument. This suppose you have a double bass player availlable, the whole recording chain and a not too obstrusive room to chek results.
There is chance you'll find qtc 0,5 with low fc more realistic this way IF your goal is to be true to source. If it's not your goal then...
I started in music as bass player and i had a combo amplifier which used a Qtc 1,2 ( or more can't remember) cabinet/driver system. A nice way to 'fake' louder fundemental of my then 4 string and keep cabinet volume reasonable (15" driver). It was nice to have a bit more output to fight the 100w amp of guitarist and my hammer fisted drummer. But sound of it was blurry, not accurate enough to my taste and genre played ( metal), it was totally acceptable for rock, blues, reggae,....
When i switched to 5 strings it was useless. Low B wasn't audible on it.
It's all trade off and compromise...
Last edited:
The Keele 6th order alignment uses a 2nd order electronic high pass which while helpful is not steep enough to minimize excursion immediately below Fb.
In the AX article I show a reflex tuned lower than normal so it has a soft knee, combined with a shelf filter with a Q of around 2. The combination is flat down to Fb and the shelf filter cuts the signal by about 20dB in 1/2 octave, then levels off. This reduces phase shift and the combination is quite elegant.
This is a crude sketch of the difference.
In the AX article I show a reflex tuned lower than normal so it has a soft knee, combined with a shelf filter with a Q of around 2. The combination is flat down to Fb and the shelf filter cuts the signal by about 20dB in 1/2 octave, then levels off. This reduces phase shift and the combination is quite elegant.
This is a crude sketch of the difference.
Last edited:
Human intellect does not function like that. They can tell you the Qtc of a driver in a cabinet only after they had it measured. It's an alternate way of describing a certain frequency response under specific room conditions.
The LS3/5 has Q =1.2 the Kef Concerto a Q =0.7, the Transmission Line that of 0.45 as per specs. Besides, I can't recall if I ever heard an instrument being played by Dire Straits, Metalica, Pink Floyed, Uriah Heap, Inx,etc. or any of the 60s to 80s rock bands. I am led to believe that the thump thump is referred to as bass drum, the bong bong maybe a bass guitar, twaat twaat may be a trumpet but I have no idea, the sound coming from them are great but different in every way, some move me internally some chases the cats away, others make my neighbours 50 meters away complain. I will ask them which Q they most likely approve of.
I have answered my own question as to what are you supposed to hear to identify low Q from High Q. If I hit a cardboard box it goes baf, if I hit an empty gas drum it goes boooom so I guess the cardboard box is Q<0.1 and the fuel drum is Q>5, because it rings for some seconds longer. All I need is to find which type of cardboard box (multi layer, corrugated or whatever) is most likely to float my Q and give my current speakers to the salvation army. 🤣 🤣 🤣 🤣
Numbers are not what counts subjectively at all.
I have answered my own question as to what are you supposed to hear to identify low Q from High Q. If I hit a cardboard box it goes baf, if I hit an empty gas drum it goes boooom so I guess the cardboard box is Q<0.1 and the fuel drum is Q>5, because it rings for some seconds longer. All I need is to find which type of cardboard box (multi layer, corrugated or whatever) is most likely to float my Q and give my current speakers to the salvation army. 🤣 🤣 🤣 🤣
Numbers are not what counts subjectively at all.
So what is the point discussing it. You can modify Q to anything you want by either changing compliance or weight, so which sound better, pointless to say the human intellect does not function like that. Are we looking at graphs or listening to music. I guess that you print your speaker Q and response, frame it and hang it on the walls. When friends come around you can show them what your speakers sound like and they will be very impressed without even hearing anything. That is an audiophile. 🤔
I guess you have to be specific when one says "better"... 0.7 allows more of the response to remain neutral and that isn't a fail, per say, either, but in essence a designer might favor 0.5 Q with a 12db/oct slope for his Crossovers
Which is the problem, Filter Q and voicing are not the same thing.... A person who is saying a Q0.5 sounds thin has absolutely no grasp on what he is describing, If you won't accept that you will remain lost. Stop treating filter Q as a specific place in the spectrum. The response on the left has a Q 0.5 and a f3 of 10hz the response on the right has a Q of 18 and f3 of about 1200hz.... which one do you think sounds thin in application?
Q doesn't even specify a slope order...Q doesn't even specify a slope order...Q doesn't even specify a slope order... lol
If its crossover, the whole "thin" aspects makes even less sense
What's more important? Understanding the relationship between Filters and group delay, resonance, and decay.
TLDR = A smooth Frequency response is the highest priority.
Lets look at an undistorted FR
The FR changes, cause distortion. Regardless the Q, Regardless the Slope Order. The more gradual the changes are in FR, the less distortion
Last edited:
Just think a little. Qtc = 0.5 in a typical real-world scenario means a relatively high F3 (in an anechoic environment), which means that the frequency response is audibly distorted, so despite the clean impulse response, a critically damped 2nd-order system still does not provide the final fidelity as long as F3 doesn't go below 20Hz and the low frequency roll-off Q (including room gain etc.) is still 0.5..
Last edited:
So what I understand is that critical damping sucks out second harmonic? Will definitely affect the sound signature. With higher third and less second the system will sound more "clinical" is that what we want? I am starting to figure out what to listen for.
YSDR yeah thats the logical conclusion. And I'd almost bet that if response can't be extended that low then any boost down there would make sound better even though group delay picks up, so use what ever Q if it extends lows it'll make sound better up to some point.
Basically, simplified, this makes rule of thumb that what ever one does bigger system will work better, as it's response can be extended lower with lower Q for given SPL. The smaller the system the higher the Q needs to be to get some extension, as it's eventually volume displacement that limits SPL at some low frequency.
Basically, simplified, this makes rule of thumb that what ever one does bigger system will work better, as it's response can be extended lower with lower Q for given SPL. The smaller the system the higher the Q needs to be to get some extension, as it's eventually volume displacement that limits SPL at some low frequency.
I am really starting to figure out why the Rogers LS3*5a sounds different. The bass is very fast and punchy but not thick, with Q-1.2. There is no listeners fatigue at all. Maybe that was one of its design features. So that recording professionals can listen for hours. When they are positioned correctly on my desk there is no lack of bass just seems very controlled and precise.
EDIT: Exactly the conclusion of TMUIKKU. We wrote our responses at almost the same time.
EDIT: Exactly the conclusion of TMUIKKU. We wrote our responses at almost the same time.