Bone conduction is a difficult subject. Do you have any knowledge of transient hearing related to bone conduction.
-
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!
Is transient response the most important thing for the perceived audio quality in a system ?
- Thread starter Tangband
- Start date
ernestcarl
Major Contributor
Could you elaborate on your question?
What does it even mean to say “phase response is audible” or “not audible”?
Did you listen/watch the video I posted and how it can be argued either way depending on the test design?
ernestcarl
Major Contributor
The "impulse response" of many loudspeakers ain't really all that bad. But it can be improved upon with better speaker design and acoustic listening space design -- either way, this can include some additional signal processing in the chain.
Yes.
2.5 Minimum Phase and Non-Minimum Phase Behaviors
A system is defined as being minimum-phase if both the system transfer function and its inverse are causal and stable. A consequence of this definition is that if a correction or preconditioning filter can be created that corrects the magnitude response of a minimum phase system (meaning the system’s response doesn’t go to zero at any frequency), it can also correct the phase response - yielding a perfect impulse response with no latency.
Because of this property of minimum phase systems, it has often been stated that non-minimum phase effects cannot be corrected by preconditioning filters. In fact, it is only true that a non-minimum-phase system cannot be corrected perfectly. However, the imperfection may simply be latency - which in audio, is an exceptionally benign imperfection.
All loudspeaker systems are subject to significant latency, because of the relatively slow propagation speed of sound in air. Therefore, a small amount of added latency is usually inconsequential.
Frequently, a filter can be defined which corrects a non- minimum phase system’s magnitude response, and which linearizes its phase response. The net result is a system which is perfect in the sense that its impulse response is a delta function, but in which the impulse is not located at t=0.
In other cases, a filter can be defined which approximately corrects a system’s magnitude response and approximately linearizes the system’s phase response, while introducing minimal latency.
In short, many non-minimum-phase phenomena are practically correctable in the context of audio applications. Crossover phase linearity is one example.
...
3.4 Crossover Phase Linearity
Multi-way loudspeaker systems typically employ minimum phase high pass and low pass filters. When such filters are summed, a flat magnitude response may be obtained, but the phase response is non-minimum. The resulting system has an all-pass transfer function, with a phase component that is similar to that of the high pass and low pass filters from which it was constructed.
To linearize the phase response of the summed crossover, we would need a complementary filter, created by inverting the summed response. The result is itself an all-pass filter, but with a phase response that is the mirror image of the summed crossover. In the time domain, this yields an impulse response which is reversed in time. Since the crossover filters themselves have infinite impulse responses, the correction filter is unbounded in negative time.
The non-minimum phase behavior corrected for below (spanning the bass xo region) is primarily from the room itself -- i.e. "room correction" was performed to improve the transient response.
Last edited:
My point on "bone conduction" is not directly related to transient hearing, but just a reminder for you.
As you may well aware, "bone conduction" would be more important in high to super-high Fq zone than in low Fq transient sensation, as I recently pointed here regarding age-dependent hearing decline in High Fq.
Even though many people including here on this thread would usually focus on "transient hearing" in rather low-Fq sound such as kick drums in jazz, let me point and share the importance of "transient hearing" in high-Fq high-energy transient sound.
I repeatedly shared this topic on my project thread, and my post here would serve as comprehensive explanation and actual hearing example; I wrote there;
Even though a lot of people have been intensively discussing about audio reproduction of low frequency bass sound like on this thread entitled "Bass!", very few people talk about the suitable high-Fq transient sound music tracks for check and tuning of our audio system and room acoustics; this is also the case even on this thread entitled " Critical (Best) Music Tracks for Speaker and Room EQ Testing" started by amirm.
In this post, therefore, let me share with you one very unique track containing extremely high-energy high-frequency sharp transient music sound of Bimmel Bolle antique orgel.
I recently posted here a video clip of dancing/moving my 12-VU-Meter Array with the sound of that "Bimmel Bolle antique orgel", and I hope it would be allowed sharing the video clip here again for your easy understanding on my point;
I assume that "bone conduction" too would be a matter of consideration for "transient hearing", in listening to this kind of high-energy (high-gain) high-Fq sharp-transient sound; of course we need nice HiFi tweeters plus excellent super-tweeters, though. Edit: This is only my naive speculation; please refer to post #131 below on this thread.
Last edited:
The non-minimum phase behavior corrected for below (spanning the bass xo region) is primarily from the room itself -- i.e. "room correction" was performed to improve the transient response.
Thank you for sharing!
Very early stage in my multichannel project, I too utilized REW's Wavelet analysis for check and validation of group delay tuning in DSP software (in my case EKIO) as shared in my old posts here through here. Even though that was done before the arrival of multichannel DAC (OKTO DAC8PRO) at my home, I too could confirm what you kindly pointed just before on this thread.
Yep. Ideal phase is perfectly flat at zero degrees. (with all constant delay removed)
Pragmatically, it is achievable everywhere but at very bottom of the spectrum, where inevitable response rolloff brings lagging phase.
Yes. However linear phase is not a requirement for flat phase response.
An acoustic first-order crossover has flat phase,
as do full-rangers, and single driver headphones, etc
Most speakers don't give a good impulse response, unless the designer went to some effort to time and phase align....
and use either very low order IIR crossovers, or better, linear phase crossovers.
Probably not so important for a simple two-way, but for each 'way' added, the benefits of flat phase / good impulse, grow quickly ime.
ernestcarl
Major Contributor
Indeed, I have seen those. While I don't think there is any single plot view that can fully represent/illustrate a loudspeaker's transient response since the transfer function varies (e.g. angle of microphone and positioning and test signal and filtering applied), the wavelet spectrogram is definitely one of the most useful tools for this.
- Joined
- May 1, 2021
- Messages
- 492
- Likes
- 1,557
The absolute shortest I could make it
Does anyone have an example of a file that had audible phase issues? If so how many "degrees" of phase shift do you need over what frequency range? I just made an all pass filter in Rephase with 2000 degrees of phase shift between 25 Hz and 1000 Hz and I can't hear any effect at all.
fpitas
Master Contributor
Some of the finest speakers made use LR4 throughout, the Salon2 being only one example. After all these years of being told we need to have perfect square wave response, I'm a little skeptical about the importance of it.
ernestcarl
Major Contributor
I uploaded an "extreme" example a while back of a kickdrum bass subwoofer test track with/out all pass filters applied. I told people to use Foobar and ABX test the files. If I recall correctly, people said they could hear it -- although one person seemed to back out? I have to look for it... but it's uploaded here on ASR. Personally, I can hear the effect of a 2nd order all pass filter myself.
I've also done ABX listening tests (~2 kHz LR4 xo phase linearization -- where magnitude response remains identical) of my KH120s in the nearfield at ~1m distance (very quiet, acoustically dry room setup) with some pop tunes at moderate to low SPL volume... And, yes, even with those rather random selections (on the top my head I remember songs by Shawn Mendes and Sheryl Crow), with much effort and practice, I could hear subtle audible differences. *see attached files
Audible differences doesn't necessarily mean better, of course.
In the case of the former wavelet spectrogram example I posted, roughly half a dB of shelving boost in the bass is required to somewhat psychoacoustically match the uncorrected response with the corrected response -- i.e. FIR filtered bass actually sounds louder and more intense if magnitude response is kept evenly the same.
While as to the latter mid/HF xo phase post-linearization, a certain softening/smoothing in some of the transients was noted. Subjectively, to me this was good as I felt this reduced some harshness of the KH120s by the tiniest bit despite the rather small time processing delay cost.
Both -- assuming were performed correctly -- resulted in an overall increase in clarity.
To be clear, I cannot vouch for others. But this is what I heard from my own listening AB and ABX test setups.
Last edited:
dasdoing
Major Contributor
how would the manifestation of a sluggish response to abrupt input changes appear in the frequency response?
fpitas
Master Contributor
Bad high frequency response.
dasdoing
Major Contributor
It might be able to accurately reproduce a high-frequency sine wave, yet it may still require a few cycles to reach its maximum sound pressure level.
fpitas
Master Contributor
That sounds like it has a resonance. Why else would it do that?
dasdoing
Major Contributor
Inertia?
Can speakers reach their maximum amplitude within the first cycle starting from zero?
fpitas
Master Contributor
The FR defines all of that, once again. I think we're going in circles...
You mean maximum movement outwards in space?
- Joined
- May 1, 2021
- Messages
- 492
- Likes
- 1,557
This. The frequency response (magnitude and phase) at positive frequencies are a manifestation of the underlying transfer function which alongside knowledge of the system being causal and real (physically realiazable) gives us not only the response to sinusoidal signals but also what we need to calculate impulse response, step response, and so on.
dasdoing
Major Contributor
Just wondering if the requested SPL at a given frequency is immediately present from the beginning.
