speaker location and graph help

[muse77]

I have been playing with REW,speaker location and trying to understand different graphs (i'm still new at this). I have done so many sweeps that my wife and son are walking around imitating them. Luckily it is winter and the neighbors can't hear them . Seriously though here are the results. As of right now the room has no treatment yet and the speakers are angled across a corner.

Am I going in right direction by doing this?
bass graph.jpg

The next graph I put the mic 4in. in front of each speaker and covered mic. I the covered the speaker and mic with a blanket to avoid room reflections. Would these results be usable?
Speaker at 4in with blanket.jpg

Also the next graph is time energy of the mic on the couch.
yellow sub green full range from couch 20 ms smoothing.jpg
The help files are great for this graph but I want to make sure I understand the graph.
Are the spikes on the left from the room itself?
What I am trying to understand is what distortion is caused by the room and/or the speakers. This is so when I do room treatment I'll know if it was effective. Also by knowing the difference I hope to make simple modifications to the speakers to make improvements.


To walk a mile one must take one step at a time.

Let me know if it would have been better to split this up.
Thanks Bryan

[Wayne A. Pflughaupt]

muse77 wrote:

I have been playing with REW,speaker location and trying to understand different graphs (i'm still new at this). I have done so many sweeps that my wife and son are walking around imitating them.

That's great!

Just tell them they should be grateful that you’re not using pink noise!

Your first graph is a bit unusual, based on the info you provided on their placement (assuming you didn’t shift the blue line upwards). Typically the nearer to a corner you are, the more bass increases. The top trace looks smoother, if you can live with your speakers seven ft. out in the room.

Your second graph, where you plugged the port on one, both the green and blue are virtually identical. You shouldn’t be able to hear a difference between them one way or the other

The next graph I put the mic 4in. in front of each speaker and covered mic. I the covered the speaker and mic with a blanket to avoid room reflections. Would these results be usable?

Typically you want to measure from the listening position.

Re the graph with the yellow traces, I know nothing about time energy graphs.

Regards,
Wayne

[JohnM]

The scaled down versions of the main energy-time plot you see on the LHS are distortion components, each is due to a single distortion component (2nd harmonic, 3rd, 4th etc). The further to the left you go, the higher the distortion order. You may find these easier to see on the main impulse response graph, the level of each peak on that graph relative to the main shows the level of that distortion component, broadly speaking. It is easiest to see the distortion responses with short sweeps, as the longer the sweep the further into negative time the distortion elements are pushed so fewer harmonics are visible.

Distortion comes from the system, in particular the speakers, not the room.

[muse77]

Thanks guys. Actually the speakers sit across a corner and there is a fireplace angled across the corner. Maybe a picture would be better. I wasn't sure if the port being plugged might make an audible difference.

I will play with shorter sweeps to see the difference. I hope to understand this part. Could I plug the speaker cables in to the sound card and get a measurement distortion from my system? I am thinking I could use a digital meter to check the input voltage at 75db from the Radio Shack meter. Then match the voltage from my stereo so I don't do something nasty to the soundcard.

To walk a mile takes a step at a time.

Thanks again, Bryan

[brucek]

Could I plug the speaker cables in to the sound card and get a measurement distortion from my system? I am thinking I could use a digital meter to check the input voltage at 75db from the Radio Shack meter. Then match the voltage from my stereo so I don't do something nasty to the soundcard.

You certainly could not plug speaker cables from a power amp into a soundcard, but you can plug the output of the preamp (low output impedance, line level) into the soundcard.

But what reason would you do this? The signal is perfect from the electronics.
Oh, it may exhibit some distortion < 0.01% no doubt, but it's nothing you would ever hear. All the trouble comes when we feed the signal to a speaker and the room gets a hold of it.

brucek

[Wayne A. Pflughaupt]

All the trouble comes when we feed the signal to a speaker and the room gets a hold of it.

Well-put.

[muse77]

brucek wrote:

You certainly could not All the trouble comes when we feed the signal to a speaker and the room gets a hold of it.

brucek

THANK YOU! The light bulb has blinked on in my brain. That was the answer I needed. Now I can focus the speakers and the room. I was begining to think that I was chasing my tail on a wild goose chase.

What my goal is: When treating a room when to understand when the problem is the room or when the problem starts with the speakers and to understand what the different graphs are telling me.

Thanks, Bryan

[brucek]

When treating a room when to understand when the problem is the room or when the problem starts with the speakers

It's hard to get an accurate reading of a full range speakers response in a room. You need an anechoic chamber, or some people take readings outside with good success (it's a little cold right now though).

With a sub, you can do a near field and it gives a fairly good reading because there's only one driver. With full range (as you've seen), it's difficult.

I think with full range you have to rely on the speakers published spec and assume everything else is the room...

brucek

[muse77]

I would probably have a better chance at buying the winning lottery ticket than getting published specs for the Bose 701 series 2 speakers. I did read an interview with Siegfied Linkwitz where you have a mic on axis with the speaker and give it a 1/3 octave pulse. The pulse is short enough that reflections from the room don't have time to interfere with the signal. Anything you pick up would be from the speakers and any leftover energy in it. I might be comparing apples to oranges but I think that is what John kind of meant " in post#3". When he said to use shorter sweeps to see distortion in the impulse response graph. I think I am going to give it a try.



I am learning from this experience and really appreciate the help.

Bryan