DiyMobileAudio.com Car Stereo Forum banner

21 - 25 of 25 Posts

·
Registered
Joined
·
8,484 Posts
You are all aware of the averaging feature in REW's RTA, aren't you? I think the fan idea is cool, but I'm not sure it's necessary most of the time. By setting the RTA to 32 averages you'll see that graph gets pretty tight after only about 16 averages, so you don't need to move the mic back and forth 100 times, a few slow and smooth passes around the listening position will get you the data you need.
 

·
Registered
Joined
·
756 Posts
You are all aware of the averaging feature in REW's RTA, aren't you? I think the fan idea is cool, but I'm not sure it's necessary most of the time. By setting the RTA to 32 averages you'll see that graph gets pretty tight after only about 16 averages, so you don't need to move the mic back and forth 100 times, a few slow and smooth passes around the listening position will get you the data you need.
I think the thought process is to have repeatable movements.

I've definitely struggled to replicate measurements depending on my posture or position. Even back to back. Even with 150 averages.
 

·
Registered
Joined
·
8,484 Posts
I think the thought process is to have repeatable movements.

I've definitely struggled to replicate measurements depending on my posture or position. Even back to back. Even with 150 averages.
I sit in the back seat and get pretty repeatable measurements. Repeatable enough to clearly see the problem areas.

There are obviously several techniques that work, hence the purpose of this thread, but I do think that it's possible to over-analyze, and all you need to do is get a good representation of the overall output to level match, and a clear idea of where the problem areas are. I think that even a few fixed mic measurements can do this without too much variation.

When there are significant variations between measurement methods, I think that most of those variations can be ignored. You should see very similar problem areas regardless of the technique you use.
 

·
Moderator
Joined
·
18,544 Posts
Discussion Starter #24
The fan method is for:
1) easy to stand somewhere else because sitting in the car may be uncomfortable,
2) continually run the RTA while you make adjustments "live" without having to hold the mic, measure, pause, make changes, rinse/wash/repeat
3) repeatability.

I use correlated pink noise.


"Significant variations" can be relative. To some 2dB is significant. To others 5dB is significant. The purpose of this video was to show the differences are there and it is important to consider them and how you use a target curve. Not to say one method is better than the other.

I hope that clears things up.
 

·
Registered
Joined
·
4,982 Posts
I've abandoned the usual methods of measuring since I got the APL dsp. I find it to be faster and more reliable, it uses the sound power response from multiple points in the entire listening space. If you compare it to spatial averaging, it's similar in the lower frequencies but tend to differ quite much in midrange area and up. I still use a "normal" measurement, basically a one-point sweep in front of the headrest to measure the response in the modal passband (lows) since it's more accurate than the above method in the lows.

The audibility of variations between L/R are also dependent on the Q of the frequency passband. A wide/broad difference is more significant than a high Q/narrow band is with small variations in amplitude... i.e if the the difference is wide (perhaps between 2000-8000Hz) a very little amplitude difference is audible. I notice 0,5dB in such situation. With a high Q variation/narrow (like 2000-2100Hz) , it might not that audible. Such variations could cause a wandering stage if you should have the L/R-difference in the vocal region for example. Tonality wise, it's basically the same deal but perhaps less important. Should aim for a spatial average that looks smooth. High Q/narrow band EQing only works in the modal region generally. It's possible to use high Q filters to attain a smooth response in conjunction with other adjacent filters like the autoEQ function in RoomEQ but that's another topic.

Here's an interesting pdf on spatial averaging. Spatial averaging (pdf)
 
21 - 25 of 25 Posts
Top