Is transient response the most important thing for the perceived audio quality in a system ?

[Killingbeans]

Hmm… I also wondering if a stiff loudspeaker cone ( like metal ) in a good driver are better with transients than softer ones.

If the moving mass and the motor strength is the same, then the only difference is the breakup modes. So, no.

It's been said a million times already: bandwidth and "speed" is the same thing.

 

[ernestcarl]

I have a DIY 4 way active system (All crossovers are analog active LR4). I use REW and Rephase for speaker / room FR correction and spent a considerable amount of time messing around with "phase". One big issue is it is not easy to get in room phase measurements that you can trust even with gating and all the other tricks. I created a separate filter in Rephase just for "phase" so I could quickly switch it in and out and the difference was subtle to non-existent. I finally settled on using the Rephase built in normalization filters for the crossovers and the speaker box, it gets me a long ways toward "flat" phase and is quick and easy and based on reliable inputs rather than potentially dodgy in room phase measurements. I still can't hear much difference but it sounds good and makes me feel better I am "close enough" to phase coherence.

With commercially available products, I'd prefer the loudspeaker manufacturer to perform the anechoic measurements and provide the "linear" transient response "design" themselves, say, via DSP (e.g. Adam, Neumann, Genelec, Meyer Sound, Fulcrum Acoustic) -- or passively via analog all-pass filters e.g. PSI. When it comes to the final "in-room" time-domain linearization correction... well, that's really up to the overall system designer/installer/calibrator.

 

[ernestcarl]

Might as well re-post this this in-depth discussion by @René - Acculution.com here:

What is audible? Results very much depends on the listening test design. You can argue either way whether this or that phase/transient response distortion in loudspeakers is audible. But, how important is it, really? Eh, I personally think it's likelier to be more substantial for "hi-fi" than any incremental SINAD improvements in the latest up-and-coming DACs.

 

[Axo1989]

I didn't see how you could take a swept sine wave and figure out where the "edges" would be.

So an experiement:

Create a transient of the worst (and for digital, illegal) case - a single full scale sample. Play that trough the speakers and record it. IIt sounds like a "click".

Then do a swept sine frequency response measurement, and display the imoulse response. It wounds like a "woooooooooooop" from low to high frequency.

The analysis of the impulse resonse derived from the frequency response looks all but exactly like the recorded "click" (transient).



Single Full Scale Sample sent through speakers, audio playback recorded in Audacity via UMIK-1:

Impulse response calculated from a ten second 10-24kHz sweep tone in REW:

A recorded and calculated Step response are equally similar.

There obviously is some math beyond my pay scale going on there.

Nice. And kind of amazing really. That's why it's correct and also pretty useless when someone says "look at the FR". I can look at the FR until the cows come home and still not see the impulse response, step response, etc (even distortion is in there too). I have to ask my silicon-based servant to transform it for me.

 

[Axo1989]

Might as well re-post this this in-depth discussion by @René - Acculution.com here:

What is audible? Results very much depends on the listening test design. You can argue either way whether this or that phase/transient response distortion in loudspeakers is audible. But, how important is it, really? Eh, I personally think it's likelier to be more substantial for "hi-fi" than any incremental SINAD improvements in the latest up-and-coming DACs.

Four hours? Thank you (non-sarcastically). I often complain when there's not enough depth. Also, next time I can't sleep ...

 

[dualazmak]

2. Yes, very important - explain why​

(not the most important, but one of the very important factors)​

Subjective sensation on transient sound would be greatly dependent on;

1. transient behavior/characteristics/specification of each of the SP drivers,
2. transient behavior of two SP drivers around the XO Fq, especially woofer to midrange because of the very large difference in "inertial mass" (mass of moving parts) between the two drivers. If you have sub-woofer(s), the XO from sub-woofer(s) to woofers greatly affect on low-Fq transient sound,
3. time alignment for all the SP drivers in 0.1 msec accuracy/precision,
4. room acoustic environment, of course

For measurement and tuning of transient behavior and time alignment of your audio system (SP system) and room acoustics, you need to have/establish your "well validated", fully understandable, reproducible, and high-precision methods which should not have any "black-box" type hardware/software signal processing.

At least in my case, I did these by rather primitive but "fully validated" reliable methods, as shared on my project thread;

- Precision measurement and adjustment of time alignment for speaker (SP) units: Part-1_ Precision pulse wave matching method: #493
- Precision measurement and adjustment of time alignment for speaker (SP) units: Part-2_ Energy peak matching method: #494
- Precision measurement and adjustment of time alignment for speaker (SP) units: Part-3_ Precision single sine wave matching method in 0.1 msec accuracy: #504, #507
- Measurement of transient characteristics of Yamaha 30 cm woofer JA-3058 in sealed cabinet and Yamaha active sub-woofer YST-SW1000: #495, #497, #503, #507

- Identification of sound reflecting plane/wall by strong excitation of SP unit and room acoustics: #498

- Perfect (0.1 msec precision) time alignment of all the SP drivers greatly contributes to amazing disappearance of SPs, tightness and cleanliness of the sound, and superior 3D sound stage: #520

- Not only the precision (0.1 msec level) time alignment over all the SP drivers but also SP facing directions and sound-deadening space behind the SPs plus behind our listening position would be critically important for effective (perfect?) disappearance of speakers: #687

In addition to the precise recorded wave shape matching, as you may find in above reference posts, 3D (time-Fq-gain) color spectrum representation by Adobe Audition (in my case ver.3.0.1) of the recorded room air sound at your listening position would greatly contribute to visual-objective observation and tuning of transient behaviors and time alignments of your audio setup.

 

[gnarly]

Nice. And kind of amazing really. That's why it's correct and also pretty useless when someone says "look at the FR". I can look at the FR until the cows come home and still not see the impulse response, step response, etc (even distortion is in there too). I have to ask my silicon-based servant to transform it for me.

Yep, when they say look at frequency response and they exclude phase, impulse can be anything from excellent to dog food.

It's kinda funny to me how it's hard to look at the impulse response itself, and know what needs to be corrected to improve it.
Same for step response..hard to interpret without a lot of experience.
Whereas it's pretty easy to look at a phase trace and see what needs to be corrected.

I've learned not to even bother trying to interpret impulse or step responses... i simply smooth/flatten phase and both impulse and step head towards excellence.
This of course assumes frequency magnitude response is kept smooth..(the easy half of getting freq response right )

 

[thewas]

I've learned not to even bother trying to interpret impulse or step responses...

I also prefer looking at group delay plots which are better to interpret in terms of possible audibility.

 

[tmtomh]

The math is above my pay grade too, but I understand its principal implication very well.
Frequency response = transient response.

I think this is a really important, if perhaps basic, point, and it gets to the "psychological domain" point @Neuro made earlier in this thread. The way I see "transient" being used in a lot of online audiophile discussion is in a way basically equivalent to "speed," as in, "the transients really snap with these speakers" or "the rhythm and pacing of this amplifier was just better than the other one," or "44.1kHz sample rates are too slow to capture the fastest transients that people can hear."

In all cases, there seems to be a lack of understanding that no sound can be "faster" than its frequency; that frequency = speed.

So I think Neuro is right that multiple conceptions of "transient" tend to get mixed up together in questions like the one the OP here posed.

 

[Axo1989]

Yep, when they say look at frequency response and they exclude phase, impulse can be anything from excellent to dog food.

It's kinda funny to me how it's hard to look at the impulse response itself, and know what needs to be corrected to improve it.
Same for step response..hard to interpret without a lot of experience.
Whereas it's pretty easy to look at a phase trace and see what needs to be corrected.

I've learned not to even bother trying to interpret impulse or step responses... i simply smooth/flatten phase and both impulse and step head towards excellence.
This of course assumes frequency magnitude response is kept smooth..(the easy half of getting freq response right )

I also prefer looking at group delay plots which are better to interpret in terms of possible audibility.

Good points both. Group delay and wavelet spectrogram are great to see when thinking about loudspeakers.

Somewhat tangential but it blows my mind a bit when zooming in to the impulse that I can see the pre-ring just prior. Not because that worries me (amplitude and duration are tiny) rather that it survives from the DAC (assuming that's what I'm seeing) through the amp and speakers then to the microphone at LP. While our systems aren't perfect, they do a creditable job in many respects.

 

[RayDunzl]

This was my most successful attempt at creating what I would consider a "transient" as could occur in music.

Everything else I tried was slower in the "attack" phase.

A single tap of two stainless coffee spoons, via UMIK-1

 

[Deleted member 48726]

Not sure. We noted it and moved on. He was happy taking care of half the phase down there, because he's super-picky about bass. He left the phase-flattening on at the crossover too, just because. Thomas, the creator of Re-Phase, wasn't real surprised at our findings.

Hm. Okay. It's interesting.
How did the two configs sound / differ to you?

Thanks for sharing.

 

[Neuro]

We have at least two sensors of sound in the human body that is relevant in this context. Roughly the ears and the thorax.
The thoracic resonance is clearly noticeable at loud sound levels, clearly above 100 dB. Depending on body size, the resonance is between 40 - 60 Hz. A short signal pulse or continuous sinus can trigger this perception. No harmonics are required. This resonance is often perceived as a transient sound.

Tonal sound is sensed between approx. 16 - 5000 Hz with the ears in anechoic rooms. Sound outside this range is sensed more atonal. It is difficult to localize sounds below 80 Hz in a common listening room and the perceived loudness varies in the room. Audible sounds below 80 Hz require high sound levels. High sound levels mask higher frequencies and make the whole perception more diffuse. Even a short signal pulse below 80 Hz is perceived as muddy. The perception of sounds between 16 - 80 Hz in common listening rooms has atonal elements.
A perceived clear initial sound attack in relation to surrounding sounds has the potential to be experienced as transient. The sound should be prominent in perceived sound level and frequency relative to surrounding sounds. If the sound deviates from the expected sound, the transient experience increases. The transient sound perception is often enhanced by triggering the precedence effect.
It is frequencies above 80 Hz that the ear senses as potentially transient.
The wall of sound is not perceived as transient due to masking and lack of prominent sound attacks.

 

[audiofooled]

We have at least two sensors of sound in the human body that is relevant in this context. Roughly the ears and the thorax.
The thoracic resonance is clearly noticeable at loud sound levels, clearly above 100 dB. Depending on body size, the resonance is between 40 - 60 Hz. A short signal pulse or continuous sinus can trigger this perception. No harmonics are required. This resonance is often perceived as a transient sound.

Tonal sound is sensed between approx. 16 - 5000 Hz with the ears in anechoic rooms. Sound outside this range is sensed more atonal. It is difficult to localize sounds below 80 Hz in a common listening room and the perceived loudness varies in the room. Audible sounds below 80 Hz require high sound levels. High sound levels mask higher frequencies and make the whole perception more diffuse. Even a short signal pulse below 80 Hz is perceived as muddy. The perception of sounds between 16 - 80 Hz in common listening rooms has atonal elements.
A perceived clear initial sound attack in relation to surrounding sounds has the potential to be experienced as transient. The sound should be prominent in perceived sound level and frequency relative to surrounding sounds. If the sound deviates from the expected sound, the transient experience increases. The transient sound perception is often enhanced by triggering the precedence effect.
It is frequencies above 80 Hz that the ear senses as potentially transient.
The wall of sound is not perceived as transient due to masking and lack of prominent sound attacks.

This is very interesting. It has been my observation that transients in music which are relatively long in duration and contain a wide range of frequencies (I've seen as much as 40-110 Hz) can be perceived by body sensors and ears as a bulk of information - just a single transient with distinct timbre.

Big subs vs Small subs

I now show how the port output contributes to the total output here: https://www.kvalsvoll.com/blog/forum/topic/t6-a-t6-compact-horn-subwoofer/#postid-495 This is how it works: References to another manufacturers product were mentioned, I found the product page, looking for specifications...

audiosciencereview.com

 

[antcollinet]

Yep, when they say look at frequency response and they exclude phase, impulse can be anything from excellent to dog food.

It's kinda funny to me how it's hard to look at the impulse response itself, and know what needs to be corrected to improve it.
Same for step response..hard to interpret without a lot of experience.
Whereas it's pretty easy to look at a phase trace and see what needs to be corrected.

I've learned not to even bother trying to interpret impulse or step responses... i simply smooth/flatten phase and both impulse and step head towards excellence.
This of course assumes frequency magnitude response is kept smooth..(the easy half of getting freq response right )

What is the ideal phase response from the point of view of impulse response.

I'd assume linear phase (fixed time delay) - would that be correct?


If so, how do most speakers with wildly varying phase response give anything like a decent impulse response?

 

[Neuro]

Give me a validated and reproduced double-blind study that significantly verifies that the phase response is audible.

 

[dualazmak]

We have at least two sensors of sound in the human body that is relevant in this context. Roughly the ears and the thorax.

We should not forget-about, should not ignore, our "bone conduction".

 

[Deleted member 48726]

We should not forget-about, should not ignore, our "bone conduction".

I resist with every fiber in my body, when I see a " your mom joke" like this present itself.

 

[Keith_W]

Give me a validated and reproduced double-blind study that significantly verifies that the phase response is audible.

BTW what is that in your avatar? It looks like a meningioma growing somewhere near the 7th nerve? I can't really see.

 

[Neuro]

BTW what is that in your avatar? It looks like a meningioma growing somewhere near the 7th nerve? I can't really see.

Very close. It is a verified acoustic neuroma on the eighth cranial nerve on the left side. Unfortunately, the patient has no hearing on the left side.