Beginner trying to figure stuff out.

Try Mitch book too. It is on Accurate but a lot of the same principles apply here. In the filter designer try playing with the "true time domain subwindow" increase the 10Hz and 24khz keep the 10hz higher than the 24khz. So if 24khz is 6 than 10hz will be 7 same in the measurement and correction window. This worked for me. Also the target curve matters. Try using mixed target of minimum phase and linear phase start with 50%, have your target close to the actual measurement with the adjustment you want, it helps. Each filter you create look at the simulated step response, the step you have posted (I think it was step response) actually looks really good. You have 2 speakers placed pretty close together you should be able to reach a good step response that is on target. When you add the sub it might be a bit more challenging. A good step response really improves imaging. One last thing to remember, this is not applicable now. True time domain /driver will only be achieved in a full active 3-way system if you choose to go that route. Good luck.
 
Last edited:
  • Like
Reactions: moy
Hi @Ofer Updates on targer as suggested now:
Screenshot 2025-11-05 214853.png

Screenshot 2025-11-05 213518.png


And the IR Simulation looks nice I guess
Screenshot 2025-11-05 213558.png


The CPD window as before. Here I don't understand what you suggested to try:
Screenshot 2025-11-05 215111.png

Sound is great, getting to understand a bit more the effect of target curve on filters outcome.
 
Ok. Looks good. Glad to hear it sounds good too.
Did you try a mixed target curve?
You can try to follow the decline you have in the top end with the target. Or introduce a downward tilt like from -15db you have in 90hz to -15@20khz. These are just suggestions for you to try.
What I meant in the cpd window is something like increasing the TTD subwindow from 10hz-5ms to 10hz-7ms and 24000hz-3ms to 24000hz-5ms. Same idea in the frequency bump 5ms to 7ms and 8 to 10ms. See if that improves your simulated step response which looks pretty good as is.
 
  • Like
Reactions: moy
Ok. Looks good. Glad to hear it sounds good too.
Did you try a mixed target curve?
You can try to follow the decline you have in the top end with the target. Or introduce a downward tilt like from -15db you have in 90hz to -15@20khz. These are just suggestions for you to try.
What I meant in the cpd window is something like increasing the TTD subwindow from 10hz-5ms to 10hz-7ms and 24000hz-3ms to 24000hz-5ms. Same idea in the frequency bump 5ms to 7ms and 8 to 10ms. See if that improves your simulated step response which looks pretty good as is.
Decline target updated
Screenshot 2025-11-07 080913.png


Simulation

Screenshot 2025-11-07 081901.png


Testing 1.dB Max correction. Will do a next test with ms time update to compare

Screenshot 2025-11-07 082002.png


Here is the TTDS updated version
Screenshot 2025-11-07 084145.png


Screenshot 2025-11-07 084414.png
 
Last edited:
Yeah it looks good. I recommend giving each new procedure another name. That way it would be easier for you to play with them. Also a quick sanity check. You placed the mic exactly between the speakers? After the measurement you should aspire to 0.05ms between speakers.
 
  • Like
Reactions: moy
Also a quick sanity check. You placed the mic exactly between the speakers? After the measurement you should aspire to 0.05ms between speakers.
Yes! Got 0.02 this time, I kept adjusting the mic until center. Many thanks for the procedure profiles! Great tip
Screenshot 2025-11-08 182438.png
 
Last edited:
Nice, I see you are reluctant to add the sub. Can understand, or am I missing something?
 
Nice, I see you are reluctant to add the sub. Can understand, or am I missing something?
Sorry, I don’t know how to add the sub on the measurements, I just measured it as 2.0, not 2.1 :(
 
Sorry, I don’t know how to add the sub on the measurements, I just measured it as 2.0, not 2.1 :(
Oh! What you meant is no need to have it? Will listen to it off
 
The measurements you post, the sub is playing in those measurements?
 
OK. That explains the extension to 30hz. From the picture it seems that the sub is almost on the same plain as the speakers. That is why you probably get a relatively nice step response. What frequency is are the sub and the speakers crossed?
 
Looks like the crossover is at ~60hz. Kind of makes sense as this is where you base hump start to drop. Meaning that the sub outputs more SPL than your bookshelves. If you like this ~15db base boost from 30-90hz (aprox) than keep it and tame it with AL, as you did. You could also decrease the volume in the sub to have it closer to the speakers. There is no right answer here. In my system there is also a sub boost just because.
 
Last edited:
  • Like
Reactions: moy
There is no right answer here. In my system there is also a sub boost just because
Totally awesome for your kind feedback! You really helped me a lot 👏🏻
 
Back
Top