Home Theater Forum and Systems banner

1 - 8 of 8 Posts

·
Registered
Joined
·
17 Posts
Discussion Starter #1
Hello all, My name is Garth and I probably know just enough about REW to be dangerous. I’ve spent a good portion of the day generating ~25 frequency sweeps trying to optimize my new subwoofer integration with my main speakers for 2 channel audio. My sub is a Rythmik E15HP so many of the sweeps were adjusting the sub amp parameters.

I know the basics that you want a reasonably flat frequency response. I would like your opinion if I am close: Good, bad or average. First image is a full scan with 1/12 octave smoothing. Second image is up to 200 hz with no smoothing. I am basically hoping to get your help and advise as to what further measurements I need to do, If I need acoustical treatments and where.

I’ve tried to get the phase right near the crossover frequency of the sub (~64 hz). To be honest, I don’t know if I’ve achieved this. I’ve used the PEQ to reduce a dip around 38 hz.
I don’t really know how to interpret the impulse response and all of the things that it implies.
All of the measurements are without acoustic treatments.
I'm not sure how this post will appear with the images so I will post now and add additional images shortly.

Thanks, Garth
 

Attachments

·
Registered
Joined
·
17 Posts
Discussion Starter #2
These two images are the impulse response. As I mentioned, I don't really know what to do with this graph. Meaning how do I take action in a meaningful way based on the data that is shown in the graph. I guess I'm supposed to see that there is a peak a couple of microseconds after the initial impulse which indicates that I have a reflection x feet away. Therefore I should put a panel at locations x feet away in different places until that peak disappears?
 

Attachments

·
Registered
Joined
·
17 Posts
Discussion Starter #3
The last image I will post is of the group delay and minimum phase. I understand that you only want to eq areas that are flat minimum phase. My front speakers are Salk Songtowers that I have plugged the ports to start the roll off earlier than ported. I can post the frequency response and phase of the songtowers alone if helpful. Sorry if this is excessive posting but I will appreciate anyones help and advise.

Thanks, Garth
 

Attachments

·
Premium Member
Joined
·
1,562 Posts
> From your IR it looks like you may have measured with 2 mains both operating? This is okay for LF, but most often causes problems at HF. If the mic is exactly centered then the impact is less. To be safe it is best to measure one main at a time for MF and HF EQ purposes.

> The SPL chart looks very good.

> Yes, you can add treatment at near by reflective points as indicated by any significant ETC peaks arriving 1-8 ms after the initial peak.

> The IR is only used for timing and it is difficult to interpret for other purposes; thus all the other chart types.

> Your SW timing can be evaluated by measuring; SW, Main, SW+main and then overlay the 3 SPL measurements. There should be good SPL support through the XO range. To adjust it initially just set the distances via tape measurement and then use PinkPN with the RTA and adjust the SW distance setting a little as needed to get the greatest SPL throughout the XO range.
 

·
Registered
Joined
·
17 Posts
Discussion Starter #6
Wayne, Thanks for the link to the formatting of graphs. I may or may not reformat the graphs above but I will format further graphs to spec.

Garth
 

·
Registered
Joined
·
17 Posts
Discussion Starter #7
Thank you jtalden for your reply, I’ve seen people measuring speakers one at a time and thought it to be too much trouble. I didn’t know about the ill effects at higher frequencies. Is there a process to then sum the two individual plots to create one comprehensive plot that you can draw conclusions from?

When you say the SPL chart looks good, are you referring to the dB vs frequency chart?

What does ETC mean? You state that one should only concern themselves with peaks between 1 and 8 ms, why is that? Therefore, all of the large peaks just after the initial impulse can be disregarded. I was surprised because I placed a row of pillows on the wall directly behind my head which appeared to have little to no effect on the early msec peaks in the impulse response plot.

Regarding your last point, I am not using bass management via my receiver. My system is being used for both home theater and 2 channel audio. For 2 channel audio I don’t use the receiver so I am trying to optimize the frequency response without the receiver and use this scenario for both applications. So instead of using distance in the receiver, I am using the phase on the subwoofer. To my surprise, as I adjusted the phase, the SPL at the xover frequency did not change but the response on either side of the xover frequency changed slightly. Would it make sense to take note of the phase of the main speakers at the xover frequency when measured without the sub playing, then adjust the phase of the subwoofer in order to get the same phase value?

Thanks, Garth
 

·
Premium Member
Joined
·
1,562 Posts
Is there a process to then sum the two individual plots to create one comprehensive plot that you can draw conclusions from?
If you have REW loopback timing capability then it is also possible to use REW 'trace arithmetic' to calculate the FL+FR response from the individual measurements.

However, I would not suggest drawing any conclusion about the SPL response or make any EQ decisions for the MF and HF based on a measurement of FL+FR. There is interaction between main speakers playing the same signal (comb filtering and reflective interference). The severity of this effect varies significantly depending on centering of the mic and asymmetry of the room setup.

[I suppose a FL+FR measurement is useful in helping to highlight the severity of the problem. As such, it may help with identifying efficacy of any corrective actions taken. To do that we could compare the simple SPL average of FL and FR Vs the FL+FR measurement.]


When you say the SPL chart looks good, are you referring to the dB vs frequency chart?
Yes. It is a good measurement and looks comparable to many others I have seen.

What does ETC mean?
ETC is the common chart for identifying reflections. Time to read the REW help on the ETC chart. There is lots of great info there. Other good sources of info are wiki and probably Acoustic Frontiers. Just browse for "Energy Time Curve" if "ETC" doesn't find good sources.

You state that one should only concern themselves with peaks between 1 and 8 ms, why is that?
The <1ms can be addressed, but it may not be as straight forward. A good portion of that is due to the spread of the phase rotation and some more of it to speaker edge diffractions. If you have very close reflective surfaces then they may fall in this range also. Normally we try to place speakers and LP far enough away from very close reflective surfaces so that significant reflections in this range don't occur.

Would it make sense to take note of the phase of the main speakers at the xover frequency when measured without the sub playing, then adjust the phase of the subwoofer in order to get the same phase value?
That may be helpful. It's difficult for me to provide any clear comments in this case. My experience is with the more conventional setup including the ability to adjust standard bass management and EQ controls. Different approaches may provide good results however. It really just depends on your setup configuration and the resulting measurements.
 
1 - 8 of 8 Posts
Top