Step Response: Does It Really Matter?

[ctrl]

I find another test with manger and waterfall. https://www.i-fidelity.net/testberichte/high-end/manger-z1/labor-seite-6.html here the waterfall look better as at troeslgraven, but can not see how much better it is as mtm or eris or kali

Yep, I can't see the superiority of the Manger in terms of "speed" either.

The CSD from i-fidelity.net only considers 20dB attenuation and you should only compare them within one source.
There, the Manger's decay is not bad, but there are better ones:
https://www.i-fidelity.net/testberichte/high-end/canton-reference-8-k/seite-6.html
https://www.i-fidelity.net/testberichte/high-end/t-a-talis-r-300/seite-5.html

 

[bennybbbx]

Yep, I can't see the superiority of the Manger in terms of "speed" either.

The CSD from i-fidelity.net only considers 20dB attenuation and you should only compare them within one source.
There, the Manger's decay is not bad, but there are better ones:
https://www.i-fidelity.net/testberichte/high-end/canton-reference-8-k/seite-6.html
https://www.i-fidelity.net/testberichte/high-end/t-a-talis-r-300/seite-5.html

In the compare with your linked speakers can see the canton is much faster as the 2 way talis-r-300. I look at the 16. slice line from behind of both speakers and the canton have much lower level on 16. slice line. but the waterfall in compare to my waterfall look extrem slow. this is a 60 cm away measure oif Eris 3.5. you can see how it look with diffrent rise times. i think ifidelity use too much rise time. faster speakers are more smoothed than slower speakers. at rise time of 2 ms there can not see much diffrence to kali. so important for to see if a speaker is fast or not is a short rise time of maybe 0.1 ms. what do you think, use short rise time show better fast speakers is good science or esoterik science i think many waterfall in speaker tests are wrong because a frequency of 10 khz have a period time of 0.1 ms. but in many waterfalls can see that the fall time is more than .5 ms.

 

[bennybbbx]

here are posts of 16. slice(can choose in REW with perspective setting which slice is show on front and the value cursor show values from front slice)with 3 ms rise time and 2 ms rise time and 0.1 ms rise time

 

[Andysu]

I don't bother with step. One the room has to be 0 quiet and the step depending on the step frequency length can take hours and if any noise appears in the room or outside the mic will pick it up and the step nothing more than waste of hours. I just do the sweep quick and fast will suffice. Otherwise the step is like watching wet paint drying on the walls.

 

[bennybbbx]

I don't bother with step. One the room has to be 0 quiet and the step depending on the step frequency length can take hours and if any noise appears in the room or outside the mic will pick it up and the step nothing more than waste of hours. I just do the sweep quick and fast will suffice. Otherwise the step is like watching wet paint drying on the walls.

the step response need no additional measure time. a step response is calculate from a normal sweep measure. same all other views.When you have REW(its donationware) the best view is the spectrogram with slice enabled . but other programs have this not. in REW there can see the step response good at a specific frequency. see screenshot in the spectrogram the cursor is at 967 hz. and the below curve show the step response at 967 hz.

 

[ernestcarl]

I don't bother with step. One the room has to be 0 quiet and the step depending on the step frequency length can take hours and if any noise appears in the room or outside the mic will pick it up and the step nothing more than waste of hours. I just do the sweep quick and fast will suffice. Otherwise the step is like watching wet paint drying on the walls.

Not sure what you mean about measuring an IR taking hours. Capturing a few sweeps only takes a couple of minutes. Vector average several measurements if you want to get a cleaner IR with less noise. Though, if there's stil a lot of noise, do it at a time when the house is quiet.

 

[Andysu]

Not sure what you mean about measuring an IR taking hours. Capturing a few sweeps only takes a couple of minutes. Vector average several measurements if you want to get a cleaner IR with less noise. Though, if there's stil a lot of noise, do it at a time when the house is quiet.

That is what I saw when I used it thinking it would be great. then saw how long it would take to do the step response based on the frequency length and tied it and went back 90 so mins later the pc was in sleep mode and the REW step response was all messed up. Maybe I did it wrong?

 

[ernestcarl]

That is what I saw when I used it thinking it would be great. then saw how long it would take to do the step response based on the frequency length and tied it and went back 90 so mins later the pc was in sleep mode and the REW step response was all messed up. Maybe I did it wrong?

It just takes a few seconds per sweep. You don't need more than 1M length -- I usually just go for 512k

 

[Andysu]

It just takes a few seconds per sweep. You don't need more than 1M length -- I usually just go for 512k

That I seem to recall yeah I used 1M length of course this was early last year when I used step response.

 

[ernestcarl]

Taking measurements can be quite easy. It's the process of how one uses or interprets that information in the correct context which takes far, far longer to learn.

 

[bennybbbx]

That I seem to recall yeah I used 1M length of course this was early last year when I used step response.

I use 128 kb and 2 repetitions in REW. take 5.9 seconds. for the 2 measure cycles. REW play then automatic 2 times the 128 kb measure tone. many professional measure systems use shorter measure tones as 128 k and measure more often. more but shorter measure cycle i think better to reduce random noise. for a example a loud car drive by. the next measure cycle start after 3 sec, in this time loud cars get silent

for step response or waterfall or spectrum ito see if speaker is fast or slow it is also good usable to measure 1.5 cm from system away.

 

[Andysu]

I use 128 kb and 2 repetitions in REW. take 5.9 seconds. for the 2 measure cycles. REW play then automatic 2 times the 128 kb measure tone. many professional measure systems use shorter measure tones as 128 k and measure more often. more but shorter measure cycle i think better to reduce random noise. for a example a loud car drive by. the next measure cycle start after 3 sec, in this time loud cars get silent

for step response or waterfall or spectrum ito see if speaker is fast or slow it is also good usable to measure 1.5 cm from system away.

Can you kinder, please do a few videos of this with the speakers subs surrounds.

 

[bennybbbx]

Can you kinder, please do a few videos of this with the speakers subs surrounds.

you have a surround system ?. have you a surround receiver that do phase correct ?. this is usefull also for the subwoofer
the step response measure you can use also to look if your speakers are all in correct position. then you need enable acoustic timing reference to 1 speaker and measure the other speakers

I have no surround system anymore(i sell this some weeks ago, because need it not. fast speaker produce better reverb as surrround i think). I only use from time to time the kali lp6 as subwoofer correct with t.racks mini dsp. it is 90 cm more way as speakers. and without phase correct can hear clear bass loss and gaps due to phase cancelation. to measure step repsonse or other of satellite better measure without subwooofer play

here is my rew setting

 

[ernestcarl]

Use the EQ window button and apply filter(s).

Then under the file menu, export the data with the EQ filters -- I believe I used the export as wave option.

Also, I have noticed that depending on how one exports the IR, the response can change.

After checking a couple more times, REW's IR export feature with EQ filters activated still causes the HF to droop (not all the time -- if I remember correctly) both in the menu and All SPL view tab. Anyhows, I've found a way to bypass it when applying a HPF by multiplying the original IR with another in which a HPF was applied (with rePhase) using trace arithmetic (essentially by convolution). You can use minimum phase filters and/or linear phase filters.

Might be useful.

 

[bennybbbx]

After checking a couple more times, REW's IR export feature with EQ filters activated still causes the HF to droop (not all the time -- if I remember correctly) both in the menu and All SPL view tab. Anyhows, I've found a way to bypass it when applying a HPF by multiplying the original IR with another in which a HPF was applied (with rePhase) using trace arithmetic (essentially by convolution). You can use minimum phase filters and/or linear phase filters.

Might be useful.

I see in the LP version that phase is linear in the HP version not but when look at rise time, there can see clear whats faster. see screenshots. also remember the Mp and HP files i have zoom more as the rise time of the speakers. if you want see a step impulse of a fast speaker, then look here. https://www.audiosciencereview.com/forum/index.php?threads/step-response-question.15375/ I notice with view of step response that this speaker must be very fast.

 

[ernestcarl]

this speaker must be very fast.

I'm past the driver speed and step response fixation, actually.

 

[bennybbbx]

I'm past the driver speed and step response fixation, actually.

and for what reason you have post the HP and LP 200 hz and for what reason need such 200 hz EQ ?. Your arguments is that the step response of a speaker change when use a EQ (which do phase shift) and so step response is not usefull. but i do not understand what sense it have to use 200 hz HP filter. i think this is simular as when somebody tell frequency diagram are not usefull and he show when he use a EQ the frequency diagram show diffrence. the sense of step response and frequency diagram is 1. to show the quality of a speaker or 2: to correct mistakes and verify if it then is better. We talked when you show your 200 hz HP results about 1, quality of speakers and here i really not understand wy you not show your results without EQ.

If you think your room is a problem then use windowing. When you are 38 cm away from any reflective thing with microphone and speaker, remember to reach the microphone it need then another 38 cm to go back to microphone. so overall its 76 cm. and audio speed in air is around 38 cm per ms. so can be sure 2 ms of your measure are without room influence. and for step response it is enough to see first 2 ms to see speed

 

[ernestcarl]

Might be useful [information].

Nothing more to it.

 

[bennybbbx]

and here how look 3 peak eq 1 at 50 hz 2. at 1 khz and 3. at 5 khz 4 db + and Q2 with frequency and phase and step response. there can see the rise time stay same. so when measure 1.5 cm near speaker there can good see the rise time of the system and compare it with the rise time of the impulse. the step response need zoom alot to see rise time good. see pictures in post https://www.audiosciencereview.com/...oes-it-really-matter.1999/page-14#post-705341

 

[Thomas_A]

I have not read the thread in detail, but I have a general question. Can step response be used to evaluate detail retrieval from a speaker, comparing the difference of the anechoic response compared to a standardised "normal-room" setup? I was wondering if the difference signal could be used to extract early reflections within 2 ms and be used as a "detail index"?

Below step response 50 cm from a speaker placed near a wall.

Below step response measured at listening position.