Difficulty with precise Subwoofer calibration

[lexx2004]

I have read many forums where individuals seem confident that they've calibrated their sub to say, for example, exactly 82db (to compensate for the RS meter's inherent inaccuracy when calibrating all channels to 85db) . I'd like some feedback to determine how to get an exact sub calibration. I have both the analog RS meter and the Galaxy CM140 meter. I've been using the AVIA test disc warble tone which works in conjunction with another speaker, say center channel and sub, or left speaker and sub. I usually calibrate using the center channel and the sub. From what I've read, the warble tone is a wide band test signal so my room likely has some peaks and dips which is causing the meter to fluctuate. My room is acoustically treated with more than enough bass traps and HF panels but I still cannot get a precise calibration as so many individuals claim to have gotten. Are these individuals using the "average" as indicated on the meter and then claiming to have dialed in their sub to that exact db? That's the only explanation I come up with. Maybe someone could suggest a different test signal for subwoofer calibration. I've also tried the THX test signals which is present on a few DVD movies but I get different readings using that method versus the AVIA for not only the sub but all the other channels as well. Any advice would be appreciated.

[thewire]

Those test disc are for a quick and easy setup. For best results you would use 1/12 octave RTA in REW and playback some Pink PN in the singal generator. I'm thinking this will be like the Full Bandwidth but it will take out some of the averaging. You would calibrate so that the speakers are around 70dB on the Spectrum window and if your using an RS meter, the sub is going to measure around 4dB higher on the RMS input. You set this 10dB higher than the center channel using a 1/3 octave scale. After all this you might find that your THX optimizer would read about 75dB for the speakers, then the sub could measure around 74dB - 75dB. Keep in mind that the sub operates on a different voltage singal than the other speakers, but it should still measure 85dB at -20dB FS or 75dB with the optimizer. There is no need to do SPL calc for measuring pink noise.

The reason you may be getting different results could be because of the decoder in your system doing the signals. In this case you would want to verify that this is the problem by burning a CD using REW of the Pink PN at -33dB FS for 75dB at -10 on your receiver or other decoding/pre/amp device. Remember that reference level is 0dBv and you should never listen higher than this.

If it sounds like there are dips measure, or try to keep this in mind when setting the level.

[Wayne A. Pflughaupt]

lexx2004 wrote:

My room is acoustically treated with more than enough bass traps and HF panels but I still cannot get a precise calibration...

How is it that you are certain about that?

Regards,
Wayne

[lexx2004]

Wayne A. Pflughaupt wrote:

How is it that you are certain about that?

Regards,
Wayne

I guess I should have been more specific. I have HF panels at all 1st and secondary reflection points (11 in total). I have a single bass trap in each corner (8 in total as some are also used throughout the room). Visually, the room looks "full" of panels and traps (the room is 15x20 feet). I'm lucky that my girlfriend is very understanding in this regard. Certainly I could add more traps (eg. at all tri-corners, ceiling bi-corners etc. but that would have a far more detrimental impact on room aesthetics since I do not have a dedicated room. So, Wayne, thanks, I should have quantified what I meant by "more than enough treatment". I find that most enthusiasts prefer to EQ than to treat the room but I prefer the latter. Some would argue that it's best to do both but I prefer not to go the EQ route at all.

[brucek]

I should have quantified what I meant by "more than enough treatment".

hehe, and Wayne should have been more clear in his post also. What he was saying (I assume, since it was the same question in my head), was how do you know how effective (or not) your efforts have been when you haven't taken a response measurement or waterfall plot of your room with a program such as REW?

brucek

[lexx2004]

brucek wrote:

hehe, and Wayne should have been more clear in his post also. What he was saying (I assume, since it was the same question in my head), was how do you know how effective (or not) your efforts have been when you haven't taken a response measurement or waterfall plot of your room with a program such as REW?

brucek

I should have mentioned that I did take measurements before and after treatment. The room was very "bright" prior to treatment. I did a 1/6 octave test and noticed a significant difference before and after the tests (in terms of leveling out peaks and dips). I haven't used REW as yet or tested every hz from say 20 hz to 80 hz using test signals so there are bound to be peaks and dips in there regardless of how much treatment I do. You can only do so much when you don't have a dedicated room.

[brucek]

I did a 1/6 octave test and noticed a significant difference before and after the tests (in terms of leveling out peaks and dips).

The effectiveness in treatement is best measured observing decay using waterfall data. You would need to use REW for that. It's not that hard to setup and use. We know you already have a computer, so you're half way there.....

brucek

[JimP]

Something you mentioned is that you use the center channel and sub for calibration.

Giving this some thought, I'm thinking that it make more sense doing this than necessarily the right speaker which is next to the sub (my approach). Seems that I have been calibrating using the right speaker then doing the listening test in part using voices which mainly come out of the center channel.

What do you guys think?

[brucek]

The majority of the information that you want to integrate with the sub comes from the mains. Voice doesn't have much information in the subwoofer bandwidth. Use the mains..

brucek

[JimP]

You've got enough interaction between the sub and the center channel to tell a lot. You sure can tell a lot if the voices sound too thin or bassy