Sub calibration contradiction (1 Viewer)

[MingL]

Would integrating a subwoofer be best done with an RTA or with an SPL meter? Wouldn't using a RTA be a better alternative than using a SPL meter?

On the RTA, we can see how smooth the mains cross over to the sub. Any differences can be seen clearly in terms of stair-stepping or obvious difference in SPL at before and after 80hz cross-over point. SPL meters don't do this and results very dependent on the accuracy of the meter at lower frequencies. SPL meter method would be more like an assumption method since room modes and resonances would throw the accuracy off, which otherwise RTAs can detect.


So, isn't calibrating speaker and sub levels be better on the RTA? Or am I missing something?

 

[Robert AG]

You are quite right that an RTA is better for calibrating a sub, or any audio system for that matter. The reasons you cited are all the right ones and the ability to see how the entire spectrum is behaving at once is invaluable.

You need to use full range pink noise however, and not the band-limited pink noise test signals on most calibration DVDs and in receivers and pre/pros. This signal source can be had on the Avia DVD test disc.

A good RTA program is "Ture RTA", (www.trueaudio.com) which also will generate pink noise. You need a calibrated microphone too for best accuracy, although the Panasonic 1/4" microphone capsules available from Digikey Electronics (http://rocky.digikey.com/WebLib/Pana...ata/WM-60A.pdf) are really flat enough without calibration for use in a home theater. You will need to build a simple battery power supply, but the schematic for this is in the DigiKey catalog. (www.digikey.com)

 

[Edward J M]

Ming:

TrueRTA generates a 10-25K Hz "chirp" that can evaluate the plot the system's FR in about 1 second.

The chirp is gated, but the gate is wide enough to allow room acoustics to be fully included, which is what I assume you are looking for.

Running the chirp with the speaker outside in a quasi-ground plane environment will show you the difference the room is adding to the response.

If you are only after the sub response, plug the RTA sound card output right into the sub and only look at say 10-100 Hz.

If you are after the sub/mains with the BM xo's included in the loop, then plug into an analog AVR input and do it that way.

Of course, you can still use a SPL meter and 1/12 ocatve test tones and manually plot the response with meter correction factors included. More tedious and time consuming, but still fairly accurate.

 

[Wayne A. Pflughaupt]

Ming, everything you noted in your post is correct,(except that RTAs will indeed react to room modes just like a SPL meter). The problem, however, is that it’s all academic unless you first equalize the sub.

If not - where you set the sub’s level in relation to the mains will ultimately be determined by the sub’s hottest peaking frequency. So it doesn’t really matter if you are able to “see” that frequency on the RTA or not - the reading the SPL meter shows will be the hottest frequency.

Regards,
Wayne A. Pflughaupt

 

[Robert AG]

I calibrate my mains and surrounds by sound level meter, then look at the entire bandwidth including the subwoofers with RTA software to match the subs to the rest of the spectrum. I use full bandwidth pink noise for this test as it gives a continual update on the RTA so you can see changes you make. If you use 1/24th octave resolution, you can see the various peaks and dips in response with excellent resolution, and how they relate to the rest of the spectrum.

I do mixes of music for motion pictures with this system, and have found no incompatibilities in levels when taking mixes done here to a film dubbing stage. The subwoofer levels relative to the rest of the spectrum have been dead-on.

 

[MingL]

The useful part about using an RTA is that it enables to let me see whats going on and wrong in the frequency domain. This is especially useful for subwoofer integration with the mains.

But isn't 1/24 octave a bit too detailed?

What I normally do is that I'd play wideband pink noise (or white noise depending on the rta I'm using) and view the frequency response of all 7 main speakers on the RTA. With this info, I can use my peqs and geqs to make similar the frequency response of all the 7 speakers.

The sub calibration follows by playing a sweep or noise signal on the L/R/C channels and viewing how the frquency response is, paying more attention to the x-over region and other room resonance points. My BFD takes care of modes that affect the subwoofer regions.

However, I find 1/24 octave resolution a bit too detailed, seeing too many narrow band peaks and dips makes EQ-ing tough.

Is this method the right or wrong way to calibrate the system on an RTA? Should I use 5.1DD sweeps/noise or can I use PCM sweep/noise?

 

[Kevin C Brown]

I use this:

http://www.etfacoustic.com/

 

[Robert AG]

You should not have all 7 speakers playing if you are trying to balance your system. Do each speaker indivitually, and individually with the subwoofer (s). I don't use 1/24th octave exculsively - different resolutions give different information. For instance, 1 octave resolution is good for getting a feel for the overall bass to treble balance, so you can see the relative weighting across the entire band of frequencies. 1/24th octave is good for seeing very small peaks and valleys in the response.

Kevin: Thanks for that link. The software looks very useful!