Does anyone know why the software my stuff uses (Ease Focus) is so far off it's SPL predictions? Are other company's software so far off? I'm curious as to the mechanism used to create the predicted numbers, I would think the software should be more than a marketing tool. Does anyone have stuff other than QSC that uses Focus, and is it accurate? Maybe my 22 year old Rat Shack SPL meter is bad.
Prediction Software
- Thread starter Matt Lillie
- Start date
Re: Prediction Software
Matt,
The software can only be as accurate as the information handed to it by the loudspeaker profile. The fact of the matter is probably that giving a single number for SPL capability is fraught with complication. At what frequency? Over what time period? At what temperature? Are the manufacturer's specs accurate? How were they taken, and how much does the software therefore know about the actual capabilities of the box?
I always take those numbers with a grain of salt... a 6dB grain of salt at least. If I really need to get within 6dB of the box's capabilities I probably need to get a box that's 10dB louder.
What numbers are you trying to reach, and what level of problems are you having? Array aiming could be part of the problem, but in my experience the shitty auto aiming algorithm in EASE Focus is more likely to create more SPL, not less.
Matt,
The software can only be as accurate as the information handed to it by the loudspeaker profile. The fact of the matter is probably that giving a single number for SPL capability is fraught with complication. At what frequency? Over what time period? At what temperature? Are the manufacturer's specs accurate? How were they taken, and how much does the software therefore know about the actual capabilities of the box?
I always take those numbers with a grain of salt... a 6dB grain of salt at least. If I really need to get within 6dB of the box's capabilities I probably need to get a box that's 10dB louder.
What numbers are you trying to reach, and what level of problems are you having? Array aiming could be part of the problem, but in my experience the shitty auto aiming algorithm in EASE Focus is more likely to create more SPL, not less.
Re: Prediction Software
Focus has many options for SPL, can be 1/3 octave, octave, 3 octave, broadband, or user defined range. There are also options for A weighted and flat, as well as peak, program, rms and a "worst case" "flat max". If I punch in broadband, flat max, A weight, I get a result of 106.7 @ 50'. In the real world, I can touch a hundred ("A", fast) at that distance with both sides of the PA running and into the limiters. Assuming a 4-6 dB gain over one side as in the software model, that's over a 10 dB error, and it's worse than that as the software predicts about 117 dBa peak. All errors added up, that's about a 20dB difference.
Aiming isn't such a big issue, we're talking four boxes /side groundstacked in most situations, but the real issue in my mind is that a "tool" that's supposed to be used by "professionals" shouldn't be so far out of whack. Which brought me to my question. I know other manufacturers use Focus, is it a QSC issue due to the numbers they provide, or is it a Focus issue? Or my meter that's about as old as you is severely out of calibration? My eardrums tell me "no", I know what 100 dBa feels/sounds like. I'm definitely not peaking anywhere near 115!
Other than that, I'm just bored on vacation.
Focus has many options for SPL, can be 1/3 octave, octave, 3 octave, broadband, or user defined range. There are also options for A weighted and flat, as well as peak, program, rms and a "worst case" "flat max". If I punch in broadband, flat max, A weight, I get a result of 106.7 @ 50'. In the real world, I can touch a hundred ("A", fast) at that distance with both sides of the PA running and into the limiters. Assuming a 4-6 dB gain over one side as in the software model, that's over a 10 dB error, and it's worse than that as the software predicts about 117 dBa peak. All errors added up, that's about a 20dB difference.
Aiming isn't such a big issue, we're talking four boxes /side groundstacked in most situations, but the real issue in my mind is that a "tool" that's supposed to be used by "professionals" shouldn't be so far out of whack. Which brought me to my question. I know other manufacturers use Focus, is it a QSC issue due to the numbers they provide, or is it a Focus issue? Or my meter that's about as old as you is severely out of calibration? My eardrums tell me "no", I know what 100 dBa feels/sounds like. I'm definitely not peaking anywhere near 115!
Other than that, I'm just bored on vacation.
Re: Prediction Software
Matt,
Those aren't options for SPL, those are options for bandwidth. I suspect the program is trying to give you "maximum peak SPL" or something similar, for which there is no definition. Your meter, with a 125ms integration time on "fast", is surely missing those peaks. Actual measured SPL will be highly dependent on program material, frequency range, and peak to average in that case.
I agree with you that professional tools ought to deliver predictable results. This is unfortunately where we end up at the end of the number inflation game. How much QSC plays along I couldn't tell you since I've never used their products in a situation where it mattered.
I have very little experience with EASE Focus. The program seems OK to me, aside from the auto splay calculator... it is certainly more usable than some other options. I have honestly always done my SPL calculations when they mattered based on continuous output capability, and considered peak to be gravy for sound quality... even then I've never tried to see what an array would do in the real world based on its prediction software, instead I've done the math and then built an array to a goal based on those numbers.
Matt,
Those aren't options for SPL, those are options for bandwidth. I suspect the program is trying to give you "maximum peak SPL" or something similar, for which there is no definition. Your meter, with a 125ms integration time on "fast", is surely missing those peaks. Actual measured SPL will be highly dependent on program material, frequency range, and peak to average in that case.
I agree with you that professional tools ought to deliver predictable results. This is unfortunately where we end up at the end of the number inflation game. How much QSC plays along I couldn't tell you since I've never used their products in a situation where it mattered.
I have very little experience with EASE Focus. The program seems OK to me, aside from the auto splay calculator... it is certainly more usable than some other options. I have honestly always done my SPL calculations when they mattered based on continuous output capability, and considered peak to be gravy for sound quality... even then I've never tried to see what an array would do in the real world based on its prediction software, instead I've done the math and then built an array to a goal based on those numbers.
Re: Prediction Software
Yeah, they're SPL options within the defined b/w, there are other options for peak, rms, program and "flat". In theory, you can pick 2k5, 1/3 octave, rms, and get a pretty little curve that shows you graphically what your spl is at whatever distance at that frequency and b/w. OR you could pick broadband, peak, it will spit out a different curve. None of the results, however, as far as I can tell, have anything to do with the actual measured results. I can put on pink w/6dB crest and get a "slow" reading 20 dB off from predicted? Spec sheets have been marketing for years, so now the tools are as well?
Try it with some other prediction software, see if the numbers are even close. I could live with, like you say, maybe 3-6 dB margin of error. I'm curious. What does your ADR stuff use? Think about some guy sitting in a cubicle somewhere, specing a system for XYZ install, and the RFQ calls for, let's say, 105 dBa "program" 100-10k at 75', even coverage, blah blah blah. Guy puts in the numbers, specs a rig, sells it, installs it, and fires it up. He's 6 or 10 dB from hitting the spec, based on the numbers provided in a "prediction" program. Hopefully this doesn't really happen but if it does, who eats the cost of doubling/tripling the rig? There's probably a disclaimer in the fine print.
Yeah, they're SPL options within the defined b/w, there are other options for peak, rms, program and "flat". In theory, you can pick 2k5, 1/3 octave, rms, and get a pretty little curve that shows you graphically what your spl is at whatever distance at that frequency and b/w. OR you could pick broadband, peak, it will spit out a different curve. None of the results, however, as far as I can tell, have anything to do with the actual measured results. I can put on pink w/6dB crest and get a "slow" reading 20 dB off from predicted? Spec sheets have been marketing for years, so now the tools are as well?
Try it with some other prediction software, see if the numbers are even close. I could live with, like you say, maybe 3-6 dB margin of error. I'm curious. What does your ADR stuff use? Think about some guy sitting in a cubicle somewhere, specing a system for XYZ install, and the RFQ calls for, let's say, 105 dBa "program" 100-10k at 75', even coverage, blah blah blah. Guy puts in the numbers, specs a rig, sells it, installs it, and fires it up. He's 6 or 10 dB from hitting the spec, based on the numbers provided in a "prediction" program. Hopefully this doesn't really happen but if it does, who eats the cost of doubling/tripling the rig? There's probably a disclaimer in the fine print.
Re: Prediction Software
Hey Matt,
I just tried to check ease focus 2 with the CODA audio design preset i use at the last festival i did. If I select broad band spl A weighted it shows a value in the 120db A, witch i think is proberly too high. If I select 1 oct at 4 or 2 khz it show a value in around 112-115 dbA that seems to correspond with I measured on site. We where about 104-106 dbA in average with peaks at 112-114 dbA and never in limit with the system. Measured with Smaart set to slow.
I will look more in to it at next gig.
Hey Matt,
I just tried to check ease focus 2 with the CODA audio design preset i use at the last festival i did. If I select broad band spl A weighted it shows a value in the 120db A, witch i think is proberly too high. If I select 1 oct at 4 or 2 khz it show a value in around 112-115 dbA that seems to correspond with I measured on site. We where about 104-106 dbA in average with peaks at 112-114 dbA and never in limit with the system. Measured with Smaart set to slow.
I will look more in to it at next gig.