Master B,
I've already seen you treating enclosures differents path impulses separatly. Is there somewhere a guideline doing so ? Analyse-wise ? Where do this <10 ms guideline come from ? It's the first time i see mention of the third, pre-pulse short. Where are you on net refining your analyzing skills ?
Side note : from Voicecoil tests, BCspeakers seems (for split voicecoil even, but maybe it's more a cure of offset) not to have at all flat BL curves. In hornresp ROAR sims, reducing BL gives good results. That should compensate at high volume. On the contrary, for high excursion application (PA/Car contest..), maybe it should be a good practice to underestimate BL in sims (a bit like Just A Guy added lossy Le option do, while first beeing added to conform sim to reality with big guys motors).
dear all,
Since i am a beginner and learner, at this moment I can not debate most of the comments made by experienced people like you ... but I am reading and learning from all your inputs. Thanks a lot ...
Since i am a beginner and learner, at this moment I can not debate most of the comments made by experienced people like you ... but I am reading and learning from all your inputs. Thanks a lot ...
Mark100 -
How much latency does that FIR add? How was the transient bass subjective sound? It certainly measure better.
I've been considering some FLH builds with either FIR or poor man's FIR to cancel reflections. Just wondering if it's worth the effort... ?
-Robert
I was testing these two different tracks with different LP-filtering.
One is a 50 hertz decaying sinewave, the other is a Risset drum with a 50 Hz centered fundamental and a sharp steep transient start.
With a 60 Hz 48 dB/octave LP filtering I had a difficult time differentiating the two different tracks from each other.
But with a 80 Hz 48 dB/octave LP filtering the difference was quite easy to hear.
I did not use any top system. This was only my HROAR with a 10 inch car audio driver.
I must still maintain that a non-sinusoidal capable but low pass filtered bass box does reproduce the very important part of a transient - the physical and most tactile steep pressure wave. A BR box is not very good at reproducing a non-sinusoidal waveform within the passband (-10 dB) of the port, since the port contribution is very smoothened into something much more similar to a sine wave instead of a steep transient pressure wave.
You can hear the difference between these two tracks through a BR because the direct radiation from the driver, but the BR will not recreate the low frequency pressure of the transient. It will sound dull and lack the punch and tactile attack of a bass horn or other high order QW tuning.
When adding my Dayton Audio PS95-8 as a nice top-system, my HROAR still sounds much more physical punchy and tactile within its passband then a comparable BR.
Hello everyone
The central question here is: does the port and the cone sum up together in a minphase response behavior?
If their total output in reality is minphase, then its response, phase, rise and decay can be inverted by any suitable EQ.
The inversion or reversion of any input can be done by the best FIR, affecting both cone and port output the same way. Problem is, they run both powered by same input signal, obviously. But they dont operate the same way, and not same time.
I doubt the results simulated by HornResp in any graph contain a complex signal, rather than a simple hilbert-transformation.
cheers
Josh
The central question here is: does the port and the cone sum up together in a minphase response behavior?
If their total output in reality is minphase, then its response, phase, rise and decay can be inverted by any suitable EQ.
The inversion or reversion of any input can be done by the best FIR, affecting both cone and port output the same way. Problem is, they run both powered by same input signal, obviously. But they dont operate the same way, and not same time.
I doubt the results simulated by HornResp in any graph contain a complex signal, rather than a simple hilbert-transformation.
cheers
Josh
Hi Robert, that trace had 64ms latency.
It's kind of funny, but I've found that reducing sub group delay has less to do with the sound of bass transients, than it does with the entire full-range sense of rhythm and timing.
With regards to bass alone, flattening the phase seems to be able to impart a texture to the bass.
It sounds/feels like the deeper bass notes are carriers for higher frequency content riding on top of the notes, riding more where they are supposed to be.
I dunno...but one sign timing is right is when nose, arm, and leg hairs tickle with high SPL bass lol
All that said, I work very hard to maintain flat phase throughout the spectrum (aka perfect timing), so it's a little hard to say exactly where the enhanced sense of rhythm and timing comes from.
Big Yep to the FLH's. They are ideal for high power bass without needing a whole lot of FIR phase work, ime.
They simply have a flatter raw phase trace to work with, like sealed subs, don't they.
But I still end up using a bit more FIR, because I like to cross as steep as reasonable between mains and subs to help reduce lobing.
I simulated that beast.
1. Lets take the usual driver. A hipass, Q 0.5, 50Hz Resonance notched out where the tunnel is. Minphase response of course.
2. Then, same frequency we have a bandpass second order, 50Hz. Also Minphase, no delay, but inversed polarity.
Thats how these components look:
(brown is driver, blue is the port)
Now when you merge them into 1 you have the almost flat sum, whatever
-> green step response
BUT the same curve created as a native minimum phase response has different decay -> black step response
1. Lets take the usual driver. A hipass, Q 0.5, 50Hz Resonance notched out where the tunnel is. Minphase response of course.
2. Then, same frequency we have a bandpass second order, 50Hz. Also Minphase, no delay, but inversed polarity.
Thats how these components look:
(brown is driver, blue is the port)
Now when you merge them into 1 you have the almost flat sum, whatever
-> green step response
BUT the same curve created as a native minimum phase response has different decay -> black step response
Now for another way to look at it, how fast the response is, a wavelet.
These are my settings used in REW:
1. a closed woofer would act purely like minimum phase
2. a ported sub with that drift in step response would clearly have group delay issues, as you see on the dotted center line
3. now really interesting, if you advance the delayed lows via FIR to linearize the group delay, this is the outcome:
The group delay is literally perfect.
But it does not represent the duration, but the peak only.
Watch the black on blue outline, at 50Hz port frequency, indicating the -12dB line.
On closed box its on -78 msec.
On the ported woofer its 2msec spread earlier preringing and longer decay.
These are my settings used in REW:
1. a closed woofer would act purely like minimum phase
2. a ported sub with that drift in step response would clearly have group delay issues, as you see on the dotted center line
3. now really interesting, if you advance the delayed lows via FIR to linearize the group delay, this is the outcome:
The group delay is literally perfect.
But it does not represent the duration, but the peak only.
Watch the black on blue outline, at 50Hz port frequency, indicating the -12dB line.
On closed box its on -78 msec.
On the ported woofer its 2msec spread earlier preringing and longer decay.
Actually. That doesnt give the whole picture yet. The same woofer without port would have a different response. Likely the Q would rise from 0.5 to 0.7 and the fs from 50 to 80Hz. And it would remain 2nd order.
So a real CB vs ported would look like that:
(brown closed, green ported)
note the brown CB fast decay, and ported overshoot
then the corresponding wavelet of that closed box looks dramatically better:
(same settings!)
and last but not least, lets compare the complex sum of ported sub vs. closed in a waterfall:
closed
ported
So a real CB vs ported would look like that:
(brown closed, green ported)
note the brown CB fast decay, and ported overshoot
then the corresponding wavelet of that closed box looks dramatically better:
(same settings!)
and last but not least, lets compare the complex sum of ported sub vs. closed in a waterfall:
closed
ported
There are no transients in a signal that's low pass filtered. Take that kick drum sound and filter it as if it was going in to a sub then look at the waveform. The first transient peak will be gone! This is math.
Hi Joshcpct,
The Step Response for the Port cannot start at the same Time as the Driver does.
In your Picture they both start at relative Time Zero making the comparison of the Port + Driver Step response not conclusive. I never use minimum Phase for events that is separated in Time.
b🙂
Hi bjorno. Why should the port not start same time?
As the cone moves out, the port sucks in.
It happens in instant correlation.
The relative phase-shift is only related to bandpass response.
Natural elements always act causal.
Only FIR can be non-causal.
What is your theory? How do you do it without minphase then?
Please if pointed at something wrong, come with solution straight, don't wait for invitation 😉 Otherwise ping pong response gets chewy...
Is it really true that the port acts instantaneously? or is this a theoretical approximation?
I have my doubts, seeing as that in a Helmholtz function you have a volume of a very compressible fluid, air, that would introduce a time difference compared to the front facing cone, partly due to the distance the sound needs to travel, but also due to the resonance build up time, there should also be an amplitude ramp-up time since the the pressure acting on the port will be reduced due to the compressible fluid volume.
This to me seems to indicate both a time and an amplitude (resonance raise time) issues if looking at the port response in comparison with the driver direct sound as in a combined response plot, and should then not be represented as having the same starting point, just thinking out loud.
I have my doubts, seeing as that in a Helmholtz function you have a volume of a very compressible fluid, air, that would introduce a time difference compared to the front facing cone, partly due to the distance the sound needs to travel, but also due to the resonance build up time, there should also be an amplitude ramp-up time since the the pressure acting on the port will be reduced due to the compressible fluid volume.
This to me seems to indicate both a time and an amplitude (resonance raise time) issues if looking at the port response in comparison with the driver direct sound as in a combined response plot, and should then not be represented as having the same starting point, just thinking out loud.
Only if the speed of sound is infinite... 🙂.
There is air between the driver diaphragm and the port, not a rigid connecting rod.
Why is there no ramp up in case of Risset drum sample? It should have.
Isnt it true that all musicals instruments have 'attack', 'sustain' and 'delay'?
Last edited:
Yes, there would be build up time. Question is how much is this build up time as compared to the music it is reproducing, coz even the music has build up time?
The steep initial transient created from the drumstick hitting the drum skin and displacing lots of air is only a few ms long as seen in the screenshot of Audacity.
Since the air inside a BR box is compressible and the slug of air inside the port has mass (hence mass inertia), there will always be a lag in the port response.
Port velocity is the enemy of transient response since the mass inertia is a square law function. Not a nice linear function.
Since the air inside a BR box is compressible and the slug of air inside the port has mass (hence mass inertia), there will always be a lag in the port response.
Port velocity is the enemy of transient response since the mass inertia is a square law function. Not a nice linear function.
The time of flow (>1000 ft/s) to the port may be relevant if one designs with huge offset.
Like port on lower rear side. Thats not BR problem but users. Same if I place two drivers apart. Shouldn't be included here.
If it was, the result would only get worsened.
Unless the port is placed up and closer to your ear, what then? 🙂
What you guys struggle with is an imaginary picture of slow reacting mass. Like flow of honey, bounce of a balloon, or stretch of rubber.
You make up stories how transients would get soft like it drowns in the honey. The port delays transients like a rubber tube.
Unless your port has a gas with significant lower speed of sound and kilometers of length.
The shape of transients of a minimum phase response system is according to its amplitude response. That is what was called earlier "math".
Take a tweeter, put an electric low-pass on it, see how the transients become chewy like rubber 😉
The port will start instantaneously because there is no digital delay buffer.
All real world elements behave causal and excess-phase can only occur via (relevant) distance. Inertia is related to minphase.
Please stop making up "bouncing theories"
cheers
Josh
Like port on lower rear side. Thats not BR problem but users. Same if I place two drivers apart. Shouldn't be included here.
If it was, the result would only get worsened.
Unless the port is placed up and closer to your ear, what then? 🙂
What you guys struggle with is an imaginary picture of slow reacting mass. Like flow of honey, bounce of a balloon, or stretch of rubber.
You make up stories how transients would get soft like it drowns in the honey. The port delays transients like a rubber tube.
Unless your port has a gas with significant lower speed of sound and kilometers of length.
The shape of transients of a minimum phase response system is according to its amplitude response. That is what was called earlier "math".
Take a tweeter, put an electric low-pass on it, see how the transients become chewy like rubber 😉
The port will start instantaneously because there is no digital delay buffer.
All real world elements behave causal and excess-phase can only occur via (relevant) distance. Inertia is related to minphase.
Please stop making up "bouncing theories"
cheers
Josh
- Status
- Not open for further replies.