Home Theater Forum and Systems banner

1 - 14 of 14 Posts

·
Registered
Joined
·
2,261 Posts
Discussion Starter #1 (Edited)
Hi! Sophomore REW user here regarding "Calibrating the SPL Reading" section of REW Help.

I came across a thread that made me a little nervous about entering a value accurate to one significant decimal place. So I have a couple of questions:
  1. How important is it to enter a precise value into the REW SPL Meter?
  2. Maybe more importantly, when would you need to enter an accurate SPL value?
In my experience, a few tenths difference doesn't significantly affect the reported maximum SPL that can be measured. Even with my external SPL meter is set to C-weighted and Slow Response, the reading fluctuates naturally with the pink noise test signal. The fluctuations are only on the order of 1dB, so I just average them. As an example: for a swing between 74dB and 75dB, I'll enter 74.5dB into REW's SPL Calibration window.

Am I over-thinking this?
Thanks in advance for any advice/insights.

Incidentally, I came across an example in which an SPL value accurate to 0.1 dB actually is needed. It's contained in jtalden's alternate method for calibrating REW's SPL meter. To quote jtalden: "...it provides a quick easy way to get a highly repeatable REW SPL cal level from calibration to calibration using only a simple loopback cable. The input level and thus the resulting headroom can also be easily changed without any impact to the REW SPL meter level." , in which an SPL value accurate to 0.1 dB is needed
 

·
Premium Member
Joined
·
1,562 Posts
I mislead you. There is little need for high accuracy for most all the common uses of REW. It usually doesn't matter to us if our trace is located at 75dB when the true level is 3 dB higher or lower. We are typically more concerned with the relative levels of the trace across the frequency range. Those relative levels are more a result of the mic calibration file and the soundcard calibration.

The method I outlined for REW SPL calibration is probably no more accurate than the other methods commonly used. The resulting accuracy of the total measuring system is likely to be within 2-3 db of true SPL. Rounding off the calculated calibration number to 1 dB is fine. I didn't mean to imply that 1 or 2 decimals added any value to the process. It was just the result of the my example calculation and an oversight in not rounding it to the nearest dB.

The primarily advantages to that process is; only a suitable loopback cable is needed and that the process is fast and easy to accomplish. [No SLM or multimeter is needed.]

That method does require us to lookup the specs on the mic and soundcard and convert them into the needed units to get the needed calibration number. We only have to do that once however. Then, we just use that same number each time we want to recalibrate the REW SPL meter.
 

·
Banned
Joined
·
4,838 Posts
My own approach involves a little splitting of hairs. While there is NO real benefit, many of us have some OCD behavior going on in places in our AV hobby behavior. Mine involves symmetry, keeping cables untangled & unknotted, and nice round numbers for readings. Again, there is NO real benefit to this, if just feels better.

There are cases where repeatability is important, where matching of driver or speaker levels is desired, and a fraction of a dB for matching makes a difference. But for general SPL readings it does not.


Plus, as is often pointed out, common SPL meters have only +/- 2 dB absolute accuracy anyway. You must pay a LOT for better than that.
 

·
Super Moderator
Joined
·
9,252 Posts
So I have a couple of questions:
  1. How important is it to enter a precise value into the REW SPL Meter?
  2. Maybe more importantly, when would you need to enter an accurate SPL value?
Typically in professional situations, such as noise mitigation in a factory setting for worker safety reasons, or a peace officer enforcing noise ordinances. These situations can’t reasonably utilize the cheap Class 3 meters we use, that technically should spec @ +/- 1.5 dBA but are often actually +/- 2 dBA or even worse. Due to the potential of legal challenges, accuracy is necessary, and as such meters can cost hundreds for Class 2 devices (+/- 1 dBA or better) or even thousands for Class 1 devices (+/- .7 dBA).

As far as measuring with .1 dB accuracy is concerned, good luck with that. Laboratory-grade Class 0 meters (cost - who knows? I can't even find any for sale) offer accuracy to “only” +/- .4 dBA.

Regards,
Wayne
 

·
Registered
Joined
·
1,784 Posts
You are very well informed - genuinely interesting! I had no idea of the resolution needed (or not needed).


Sent from my iPhone using Tapatalk
I do not believe that in setting up speakers in a room, those close tolerances are needed. I agree that in designing a crossover of speaker construction those tolerances may be needed. Further these 0.1db or 0.2 db will ever be heard. Some say we need 3.0db of difference before the listener notices a difference, I tend to think it may be lower than that actually. Never the less use a good SPL meter, I have the Galaxy CM140 and although it does have 10th of a decibel readouts, I don't even pay attention to them.
 

·
Registered
Joined
·
4,015 Posts
Jack, I'm curious about the galaxy meter. I've always used my RS digital, which if nothing else was consistently, itself. I noticed the other day a discrepancy between it and the REW meter via hdmi with calibration file. So for the first time used the REW meter to set my channel levels. The rolling decimals are irritating btw. Haven't had time to evaluate anything but in the past, on certain material, I've been able to notice 1db deficits in 2ch. I've always used my RS because it's simple and (I thought) effective, and faster than REW. Now me thinks it's time to go shopping.
 

·
Registered
Joined
·
1,784 Posts
Jack, I'm curious about the galaxy meter. I've always used my RS digital, which if nothing else was consistently, itself. I noticed the other day a discrepancy between it and the REW meter via hdmi with calibration file. So for the first time used the REW meter to set my channel levels. The rolling decimals are irritating btw. Haven't had time to evaluate anything but in the past, on certain material, I've been able to notice 1db deficits in 2ch. I've always used my RS because it's simple and (I thought) effective, and faster than REW. Now me thinks it's time to go shopping.
I got it Willis because it was a component suggested in the REW instructions...and because my RS meter was having fits. The Galaxy seems to do quite well and I love the max and min hold features. Not the cheapest horse in the barn but it works quite well...every time.
 

·
Banned
Joined
·
4,838 Posts
I do not believe that in setting up speakers in a room, those close tolerances are needed. I agree that in designing a crossover of speaker construction those tolerances may be needed. Further these 0.1db or 0.2 db will ever be heard. Some say we need 3.0db of difference before the listener notices a difference, I tend to think it may be lower than that actually. Never the less use a good SPL meter, I have the Galaxy CM140 and although it does have 10th of a decibel readouts, I don't even pay attention to them.
Jack, apples and oranges.

3 db diff to hear a change is for sine waves with direct level comparison (pretty sure) .

In frequency response, under controlled conditions, one band being increased by, say, 1/2 db, that gain being switched in/out, can easily be heard.

SS&I can be made/broken by 1/4 to 1/2 db differences.

But these are all RELATIVE measurements, and the OP asked about ABSOLUTE accuracy. BIG difference.
 

·
Registered
Joined
·
4,015 Posts
Wayne, I can't truly say to what level db I can detect, but I believe your statement about SS&I. I've experienced what you're saying. Even with my jbls which do ok in that regard.
 

·
Banned
Joined
·
4,838 Posts
Wayne, I can't truly say to what level db I can detect, but I believe your statement about SS&I. I've experienced what you're saying. Even with my jbls which do ok in that regard.
One of the reasons that I slightly prefer Dirac Live PC version (full) or DDRC-88A/BM over lesser versions: the slight changes in timing and level per channel can be taken back out again manually if desired. I always do it because they were always perfect, or close, before calibrating, and always find an improvement in imaging.
 

·
Registered
Joined
·
1,784 Posts
Jack, apples and oranges.

3 db diff to hear a change is for sine waves with direct level comparison (pretty sure) .

In frequency response, under controlled conditions, one band being increased by, say, 1/2 db, that gain being switched in/out, can easily be heard.

SS&I can be made/broken by 1/4 to 1/2 db differences.

But these are all RELATIVE measurements, and the OP asked about ABSOLUTE accuracy. BIG difference.
Thank you for this information as I had never heard it before. I do not believe that at my age I could hear those differences on my best day and based upon most rooms I have visited, I would most heartily suspect that such small changes would be irrelevant. Barometric pressure, headaches, head movement and a persons mood would contribute or take away far more information involving SS&I that would the 1/4 to 1/2 db differences. I know we have to have some number basis for things we are involved in, but I would think that we would need to have young ears in a silent environment or an environment with perfectly stable levels to detect such changes....at least in a home environment not using headphones. Just my 1/4 cents
 
1 - 14 of 14 Posts
Top