REW Beta Release Confusion over combined speaker response

[jtalden]

jtalden
hello. What does negative group delay mean?

Check Wiki. I think it is just the 1st derivative of the phase slope.
It is negative here as a result of the chosen FDW. The spikes here are just highlighting the strong early room effects that cause strong phase flips - typically nulls.
A larger choice of FDW may make those spikes positive, but also brings in more of the later room influence and the direct sound GD is thus hard to pick out. The point here is to see the group delay of the direct sound rather than the chaos of the later room effect. So the positive or negative spikes are just ignored.

 

[jtalden]

Double post

 

[sm52]

My question was not about current measurements. Looking at them, I remembered that I had long wanted to understand what negative group delay means at the physical level. Positive - this frequency range is delayed relative to others. And negative - is this frequency range ahead of the rest?

 

[jtalden]

I was only looking to see the direct sound trace and to show that there is no sudden offset in the curve at the XO frequency that would suggest poor XO delay timing.

I can't answer your question definitively. I just assume the negative spikes at some FDW settings to be artifacts of the interaction of the FDW filter phase distortion coupled with the the room phase distortion. I would not assume it to be anything more than that.

 

[Exponential]

HI all!

It's scorching weather here in England and I'm inside doing more measurements!
I love it except for the fact I haven't much clue if I'm doing it right!

Anyway, I have had a further play around with positioning and this is now the absolute final place my speakers will be sitting in.
I just need to nail down this 100hz crossover and I think I'm done?!?!?

MDAT is attached with new data which I hope is useable for you @jtalden

Lets see what you can conjure up!

 

[Exponential]

@jtalden Do you have time to quickly look at my MDAT please?

Sorry to bother you!

I'm going around in circles trying to get it right and I'm struggling to sync everything up.

 

[jtalden]

Post-30 data Analysis
The current SWs delay setting is still correct when using the inverted polarity on the SWs as per the previous data. This time however the SPL support in the XO range is not quite as favorable. There is less SPL support in the XO range.

This new analysis finds that the SPL support is now better when the SWs are not inverted. With this change the SWs delay also needs to be increased by 5.5 ms.

These are still the 2 most favorable settings and both are usable. There are tradeoffs of course, but I would suggest this new non inverted setting is a little better from the analysis perspective. If you are into experimentation, you may want to listen to both. Be sure to apply appropriate EQ to each of them.

Left Channel - current (blue) vs new (red)

right Channel - current (brown) vs new (blue)

 

[Exponential]

Post-30 data Analysis
The current SWs delay setting is still correct when using the inverted polarity on the SWs as per the previous data. This time however the SPL support in the XO range is not quite as favorable. There is less SPL support in the XO range.

This new analysis finds that the SPL support is now better when the SWs are not inverted. With this change the SWs delay also needs to be increased by 5.5 ms.

These are still the 2 most favorable settings and both are usable. There are tradeoffs of course, but I would suggest this new non inverted setting is a little better from the analysis perspective. If you are into experimentation, you may want to listen to both. Be sure to apply appropriate EQ to each of them.

Left Channel - current (blue) vs new (red)
View attachment 43572

right Channel - current (brown) vs new (blue)
View attachment 43573

Hi @jtalden.

Thanks for doing this! I really appreciate it.

I will apply your new settings to "profile 2" and keep the original in "profile 1" in the minidsp. Will be really easy to A-B them.

Looking at the mess that is the range between 60hz to 300hz, do you feel that changing the crossover may help in this area?
(Don't worry, I'm not going to be posting any more MDATs)
I'm asking because I don't want to go through all that effort to end up either back where I was originally or more confused than I already am!

It's hard enough remembering the steps required to get to where you got as it is!

Mike.

 

[jtalden]

You already tried relocating your main speakers with limited success. You can also try different LP distances. That often makes a significant difference in this frequency range. Often this is just shifting the problems to slightly different frequencies, but sometimes it cleans up the phase tracking and smooths the response.

The 2 delay timings we found will not be affected by any resonable LP distance change so there is no need to repeat that work. You can just see if the SPL is improved at different LP distances.

The data doesn't suggest to me that changes in XO frequency or XO filter settings would be helpful with this room geometry.

The SPL dip at 90 Hz is in both mains and SWs. A relocation of the SWs may improve this for the SWs, but it is not clear that the combined response would be improved unless the XO frequency was also increased to at least 150 Hz and sharper XO filters may also need to be used.

 

[Exponential]

You already tried relocating your main speakers with limited success. You can also try different LP distances. That often makes a significant difference in this frequency range. Often this is just shifting the problems to slightly different frequencies, but sometimes it cleans up the phase tracking and smooths the response.

The 2 delay timings we found will not be affected by any resonable LP distance change so there is no need to repeat that work. You can just see if the SPL is improved at different LP distances.

The data doesn't suggest to me that changes in XO frequency or XO filter settings would be helpful with this room geometry.

The SPL dip at 90 Hz is in both mains and SWs. A relocation of the SWs may improve this for the SWs, but it is not clear that the combined response would be improved unless the XO frequency was also increased to at least 150 Hz and sharper XO filters may also need to be used.

Okie doke.

I'll have a very limited play around with crossover slopes and PEQ when I get time.

I enjoy doing it so it's no big deal to me.

I do know that, psychoacousti ally, it's not THAT bad.

Again, thank you for taking the time to analyse my measurements.

Enjoy the rest of your weekend!

Mike.

 

[Exponential]

@jtalden

Hi.

I have made a couple of changes and I need a bit of confirmation really if you don't mind.

I'm looking at my latest measurements and I can't seem to find an adjustment that improves what I have?
I have probably done it wrong though.

Can you take a look to see if there is anything you would do to improve things? How does it look from your perspective?

Looking at the summed response from my subs to my mains together, the only thing I can see is a little hump in my subs at around 65hz that I could tackle with some PEQ.

They are all on a 100hz xover L/R24.

Thank you.

 

[jtalden]

Looks Good
The timing is a very favorable option. It is your Option1 in the chart below. The SWs leading the FL+FR by 1/2 WL. This timing is used a lot.
If you want another option to experiment with, there is Option2 shown in the chart below in blue. The SWs polarity is reversed from the current setting and 5 ms of additional delay is added to them. The chart below shows the SPL tradeoff. Option 2 timing is the more conventional target timing, but that does not suggest the sound quality is improved.

I recommend the next step is to EQ to your preferred house curve. Yes, that should include the the peak at 61 Hz.

 

[Exponential]

Looks Good
The timing is a very favorable option. It is your Option1 in the chart below. The SWs leading the FL+FR by 1/2 WL. This timing is used a lot.
If you want another option to experiment with, there is Option2 shown in the chart below in blue. The SWs polarity is reversed from the current setting and 5 ms of additional delay is added to them. The chart below shows the SPL tradeoff. Option 2 timing is the more conventional target timing, but that does not suggest the sound quality is improved.

I recommend the next step is to EQ to your preferred house curve. Yes, that should include the the peak at 61 Hz.

View attachment 43836

@jtalden

You are a complete legend! Thank you my friend.

I will have a play around with it tomorrow through the day and post the final MDAT.

I feel that option 2 is my preferred option as that peak at around 160hz would be very audible in option 1.

I post back tomorrow.

Mike.

 

[Exponential]

Hi @jtalden

I have applied your suggested option 2 as I felt that, although I suffered a loss of SPL, overall, the frequency response was more favourable to me.

There are a couple of nasty peaks and dips still but, to be honest, I can live with those as they will be hardly noticeable when the room takes over.

I hope you agree that we have both (you mostly) managed to get a really good overall response and I am now happy to put away the measurement mic etc for a while and just enjoy the music.
It sounds awesome by the way! and its all thanks to you!

Best wishes,

Mike.

 

[jtalden]

Looks Good

 

[Exponential]

Looks Good

Thanks my friend.

I appreciate your help.

Mike.