AVS Forum banner

61 - 80 of 154 Posts

·
Registered
Joined
·
3,801 Posts

Quote:
Originally Posted by J_Palmer_Cass /forum/post/16789517


Anechoic measurements are not made by anyone on this forum. Most people here tend to post measurements of the in room FR on a very limited frequency basis (AKA subwoofer area).


This is not rocket science!

There are plenty of people on this forum who do gated quasi-anechoic measurements spliced to woofer nearfield measurements. That's about the closest you can come to an anechoic measurement without access to a chamber. See augerpro's (the guy you replied to) measurement pages for examples.



And there is a bit of rocket science involved to do things right. The Harman philosophy is to design speakers that are flat anechoically and roll off in the room with an ungated measurement. To get to the right in-room measurement involves speaker design, speaker placement, room treatment and EQ. Doing too much of it with EQ mucks up the flat anechoic response (first arrival) which is a big contributor to sound quality.
 

·
Registered
Joined
·
6,604 Posts
The analog bar graph of the digital unit represents the needle on the analog unit. The AVE reading is an average reading that is updated in time and response (fast or slow). The MAX button stores the peak SPL reading of that analog bar graph for the past few seconds.


I don't think that the analog meter tells you as much as a digital meter. The digital meter calculates and average and maxium value, and the analog meter calculates nothing.


I always wondered how to equate the readings of an analog meter with the digital meter. An instantaneous analog reading is not a calculated average reading nor is it a maximum reading.


Anyhow if the digital unit measures 5 dB higher than the analog meter, then if you calibrate the subwoofer to be "flat", then flat may or not be flat depending on the SPL meter that you use. I have always used a digital SPL meter, and a "flat" subwoofer calibration has never sounded good.


I do not measure flat with True RTA either!
 

·
Registered
Joined
·
6,604 Posts

Quote:
Originally Posted by counsil /forum/post/16790580


Here is a little bit more information for you guys...


I calibrate my Denon 3808 with Audyssey Pro. After calibration my digital RS meter reads 74/75 for the sub. My analog meter reads 69/70.


Since the pro mic is calibrated to be +/- .5dB, I tend to think that a 70dB reading on an analog RS meter is more like 75 instead of the popular 72/73dB reading (in my room anyway). What do you guys think?


I suspect that you are right. That means that people who calibrate the subwoofer level with an analog meter and claim that they are "flat" may be in reality boosting the subwoofer by 5 dB.


I tend to use REW and True RTA to measure relative levels. The SPL meter does not show the peaks and dips in FR that can lead to misleading results.
 

·
Registered
Joined
·
3,801 Posts

Quote:
An instantaneous analog reading is not a calculated average reading nor is it a maximum reading.

The analog meter is averaged (with an analog circuit) with a choice of fast or slow averaging. No max or instantaneous readings are available. I don't know about the accuracy of the two but I prefer the analog meter's readout because you can read fractions of a dB. The digital meter might display 74 with a 74.4dB signal and jump a whole dB if the signal were only 1/10 dB louder.
 

·
Registered
Joined
·
2,019 Posts
JPC> the point I was trying to make is not whether that "target" was good or bad, accurate or inaccurate, but that you post it with no context, no underlying assumptions. Just post a pic and take a swipe at people here.


Now we know that it is NOT a target at all. At least not one based on pyschoacoustics (like Toole's work or even F-M curves). It's just a trend of "typical" studios. In fact the logic is pretty poor: "this is what everyone else is doing, so it must be good and you must do it too." In fact the only reason it would even be valuable is IF ALL the studios do not EQ bass flat, AND assume that the consumer does not either. Now that may have been the case 20 years ago, but is it now? NO studio EQ's the bass or uses room treatment to flatten and smooth?


Point is: now that we know a little more background on the magic plot, it takes on different meaning no? I assumed right away like the other poster that this was a real target curve developed from decades of Harman work. Now we know it is only a target IF certain assumptions are met, and I'm not sure that they are or aren't. I do feel however that they are less and less true as time goes by. And maybe even for different formats, i.e., music mastering versus HT mastering and what the "typical" consumer has for bass performance for each of those.


FWIW I do find the find the plot interesting, and I personally like my bass a little "hot". I'm just not ready to make a bunch of assumptions about plots that get thrown around without any background information, context, or givens.
 

·
Registered
Joined
·
6,604 Posts

Quote:
Originally Posted by catapult /forum/post/16791008


The analog meter is averaged (with an analog circuit) with a choice of fast or slow averaging. No max or instantaneous readings are available. I don't know about the accuracy of the two but I prefer the analog meter's readout because you can read fractions of a dB. The digital meter might display 74 with a 74.4dB signal and jump a whole dB if the signal were only 1/10 dB louder.


The accuracy of either unit is +/- 2 dB. Who cares about .4 dB one way or another. All you are doing is fooling yourself if you think that means something. It's just an SPL meter. If you want detail, you need to use a PC based RTA program.


You also do not understand the difference between the digital and analog units. The digital unit's analog bar graph averages every .2 seconds (Slow) or .5 seconds (Fast), and the digital display updates the results once a second. Max just holds the peak reading on the analog display for a few seconds. Peak readings can be stored for 2 seconds or as long as you want.


Analog units can not measure peaks because the meter movement is too slow. The users eyes and judgement are also not dependable. Still, the analog manual says that the fast setting is for peaks.
 

·
Registered
Joined
·
6,604 Posts

Quote:
Originally Posted by augerpro /forum/post/16791151


JPC> the point I was trying to make is not whether that "target" was good or bad, accurate or inaccurate, but that you post it with no context, no underlying assumptions. Just post a pic and take a swipe at people here.


Now we know that it is NOT a target at all. At least not one based on pyschoacoustics (like Toole's work or even F-M curves). It's just a trend of "typical" studios. In fact the logic is pretty poor: "this is what everyone else is doing, so it must be good and you must do it too." In fact the only reason it would even be valuable is IF ALL the studios do not EQ bass flat, AND assume that the consumer does not either. Now that may have been the case 20 years ago, but is it now? NO studio EQ's the bass or uses room treatment to flatten and smooth?


Point is: now that we know a little more background on the magic plot, it takes on different meaning no? I assumed right away like the other poster that this was a real target curve developed from decades of Harman work. Now we know it is only a target IF certain assumptions are met, and I'm not sure that they are or aren't. I do feel however that they are less and less true as time goes by. And maybe even for different formats, i.e., music mastering versus HT mastering and what the "typical" consumer has for bass performance for each of those.


FWIW I do find the find the plot interesting, and I personally like my bass a little "hot". I'm just not ready to make a bunch of assumptions about plots that get thrown around without any background information, context, or givens.



They mentioned the curve is based on varpous listening rooms and years of research, not "studios" FR curves. "By surveying many listening rooms over the years, JBL engineers have established a low-frequency response characteristic that delivers the most consistently satisfactory results. This has been confirmed through extensive subjective evaluation tests."


Your opinion of what the JBL target curve represents leads me to believe that you did not read the instruction manual for in room EQ adjustments that apply to those target curves.


My in room FR measures similar to the JBL "target" curve. An 8 dB boost in the bass is not anywhere near a flat response!
 

·
Registered
Joined
·
3,801 Posts

Quote:
Originally Posted by J_Palmer_Cass /forum/post/16791249


The accuracy of either unit is +/- 2 dB. Who cares about .4 dB one way or another. All you are doing is fooling yourself if you think that means something. It's just an SPL meter. If you want detail, you need to use a PC based RTA program.

The absolute accuracy isn't really important. That can be checked with a calibrated mic and you can come up with a fudge factor -- add so many dB to what the meter says at this frequency. What most people use a meter for is getting all their channels adjusted to the same volume with pink noise and the 1dB steps of the digital meter are too crude for that.
Quote:
You also do not understand the difference between the digital and analog units. The digital unit's analog bar graph averages every .2 seconds (Slow) or .5 seconds (Fast), and the digital display updates the results once a second. Max just holds the peak reading on the analog display for a few seconds. Peak readings can be stored for 2 seconds or as long as you want.


Analog units can not measure peaks because the meter movement is too slow. The users eyes and judgement are also not dependable. Still, the analog manual says that the fast setting is for peaks.

Please don't tell me what I don't understand. Nothing personal but I'm pretty sure I understand a lot more than you do about measuring speakers.
Peak measurements aren't particularly useful. Pink noise is random and has a crest factor so slow averaging is the only useful measurement with a meter. Of course a calibrated mic and software are better in every way but many people aren't willing to deal with the complexity.
 

·
Wireless member
Joined
·
27,116 Posts

Quote:
Originally Posted by catapult /forum/post/16791434


The absolute accuracy isn't really important. That can be checked with a calibrated mic and you can come up with a fudge factor -- add so many dB to what the meter says at this frequency. What most people use a meter for is getting all their channels adjusted to the same volume with pink noise and the 1dB steps of the digital meter are too crude for that.Please don't tell me what I don't understand. Nothing personal but I'm pretty sure I understand a lot more than you do about measuring speakers.
Peak measurements aren't particularly useful. Pink noise is random and has a crest factor so slow averaging is the only useful measurement with a meter. Of course a calibrated mic and software are better in every way but many people aren't willing to deal with the complexity.

Or the cost. The RS meter is good, good (and inexpensive) enough so that the "good" is the enemy of the "better." IMO.
 

·
Registered
Joined
·
3,801 Posts
Forget it Brandon, JPC isn't interested in the finer points of measurements and probably doesn't care that you've corresponded with the chief "JBL engineer" on this very subject and know a little more about it than JPC does.
 

·
Registered
Joined
·
2,019 Posts
He's right, I did unwittingly use the word "studio" instead of "listening rooms". Whether that substantially changes my point is up to the reader



I think in many situations the JBL curve may be a reasonable target, but if you don't understand its underpinning assumptions we might have a bunch of people out there cutting the frequencies above 12khz by 20dB just so they can use the "JBL curve". Again I'll let the reader decide on the merits of doing that in their listening rooms.


But ya know, I had so much peace this winter when I didn't post on the forums. I think I'll go back to that. I'll just post data and let others argue about what they do or don't see in a CSD waterfall...
 

·
Registered
Joined
·
6,604 Posts

Quote:
Originally Posted by catapult /forum/post/16791434



The absolute accuracy isn't really important. That can be checked with a calibrated mic and you can come up with a fudge factor -- add so many dB to what the meter says at this frequency. What most people use a meter for is getting all their channels adjusted to the same volume with pink noise and the 1dB steps of the digital meter are too crude for that.


What a silly man. If they made AVRs with level adjustments in 0.1 dB steps, then maybe you would have a valid point. My best AVR allows for 0.5 dB level adjustments, and with 0.5 dB steps are hard enough to notice a difference in steps.


In addition, since SPL measurements vary as the FR of each speaker varies by speaker location and listener location, the entire matter is just a bunch of theoretical cornswaggle. I use my ears to fine tune levels, not the SPL meter.




Quote:
Originally Posted by catapult /forum/post/16791434



Please don't tell me what I don't understand. Nothing personal but I'm pretty sure I understand a lot more than you do about measuring speakers.
Peak measurements aren't particularly useful. Pink noise is random and has a crest factor so slow averaging is the only useful measurement with a meter. Of course a calibrated mic and software are better in every way but many people aren't willing to deal with the complexity.



You are way too narrow minded. There is more to life than speaker design. I don't design speakers. Most of my speakers have been tested, and all I have to do is read the test results on line. They all measure pretty flat.


This thread is not about speaker design and nearfield measurements. It is more or less about in room measurements at the listening position. A "flat" speaker is no longer a flat speaker when the room effects come into play. There is pink noise of all sorts, and there is AVR calibration noise (AKA not pink noise), and there is calibration disk noise. They all give you different end calibration results, subwoofers in particular.



You still have to come up with some target curve in order to setup your speakers in room. An SPL meter alone in not adequate to do the job.



Just as a note, I use the MAX readings on the digital SPL meter to measure peaks of actual program material. Things like cannon shots in Master and Commander. You can calculate whether your master volume / subwoofer / speakers are calibrated properly if you know how to back into the theoretical numbers.
 

·
Registered
Joined
·
1,314 Posts

Quote:
Originally Posted by penngray /forum/post/16790238


Well this thread will probably get back on topic with toddRiffic posts again. Its really about his problem and I still want to understand what Audyssey is doing and Im sure he wants to fix his problem, which is the fact that Audyssey creates a null when calibrating his new sub.



I did like the NXT info, I always wondered why the did their EQing that way.

Hey guys,


I have to run to work,, so I have to brief. I got up this morning and broke out the audyssey mic, tripod and RS spl meter. I reran audyssey calibration, placing the tripod in my main listening position only. My low end is back again! Don't have time for the details now, but I got measurement of 31.5Hz @ 80dB, 25Hz @ 80dB, 20Hz 73dB. These are the readings directly from the RS meter with no corrections applied.
Thanks for your support penngray.
I'll post more later.
 

·
Registered
Joined
·
6,604 Posts

Quote:
Originally Posted by augerpro /forum/post/16791819


He's right, I did unwittingly use the word "studio" instead of "listening rooms". Whether that substantially changes my point is up to the reader



I think in many situations the JBL curve may be a reasonable target, but if you don't understand its underpinning assumptions we might have a bunch of people out there cutting the frequencies above 12khz by 20dB just so they can use the "JBL curve". Again I'll let the reader decide on the merits of doing that in their listening rooms.



There are a lot of people who don't know what they are doing, so what difference does your silly scenario make? You have to be able to measure first, and you can not measure 12 kHz properly with a RS SPL meter even when you use the so called correction "factors".


That JBL equipment does not EQ above 800 Hz the way that I read it, so what's the big deal on the high end of the frequency scale? For that matter, who EQ's up high anyhow? I don't know anyone who EQ's up the high end.




Quote:
Originally Posted by augerpro /forum/post/16791819



But ya know, I had so much peace this winter when I didn't post on the forums. I think I'll go back to that. I'll just post data and let others argue about what they do or don't see in a CSD waterfall...



Then you missed the Geddes thread. I learned a lot from the "experts" on that thread!
 

·
Wireless member
Joined
·
27,116 Posts

Quote:
Originally Posted by toddRiffic /forum/post/16791891


Hey guys,


I have to run to work,, so I have to brief. I got up this morning and broke out the audyssey mic, tripod and RS spl meter. I reran audyssey calibration, placing the tripod in my main listening position only. My low end is back again! Don't have time for the details now, but I got measurement of 31.5Hz @ 80dB, 25Hz @ 80dB, 20Hz 73dB. These are the readings directly from the RS meter with no corrections applied.
Thanks for your support penngray.
I'll post more later.

In spite of my poking you in the ribs on the other thread, I am very interested in hearing what you did differently to achieve these results!
 

·
Registered
Joined
·
3,801 Posts
I've got yer digital meter right here, pal.....
Pink noise showing about a 12dB crest factor.


 

·
Registered
Joined
·
3,801 Posts
Oops, sorry about that. It's a recording of pink noise for 35 seconds. Each line is a different way of measuring SPL. In the legend on the graph, all the measurements are C-weighted so they start with LC.


LCpk = peak value over the last second

LCI = RMS value with impulsive averaging (35 ms rise time, 1500 ms fall time)

LCS = RMS value with slow averaging (1000 ms)

LCF = RMS value with fast averaging (125 ms)

LCeq = RMS value averaged over the whole 35 seconds
 

·
Registered
Joined
·
2,575 Posts
Cat,


Thank you for the follow up response. It does lead me to one more question though. Is the graph meant to show the large db level between LCpk and the rest of them, and thus showing a weakness in the digital SPL meter? Thanks again.


James
 

·
Registered
Joined
·
4,575 Posts

Quote:
Originally Posted by augerpro /forum/post/16790121


I think the "Hard Knee" is just an acknowledgment that the room gain does not come on gradually (as in a curve) but somewhat abruptly at a given frequency. For example when doing the driver measurements in garage you can quite clearly see the room take over at 30hz, and it's not real gradual.

Hard knee would properly be the opposite of room gain curve to equalize it, not the same, to boost it even further.


Actually, 'hard knee' house curve comes from exactly the same place the original house curve mania came from.


Wayne admits to discovering the idea in the pro sound web forums.


Live sound requires a house curve because it's live sound. It's being mixed at the board in real time for the live audience.


The problem here is simple. At home, you're re-playing the already mixed version. If you add a house curve, you're adding that curve to the one that's already been added in the recording process.


Wayne's (or any one else's) personal subjective leanings notwithstanding, it's absurd to think that recordists leave the mixing duties to the end user.

Quote:
Originally Posted by LTD02 /forum/post/16789789


while most folks around here would disagree, the idea of rolling off at 30hz has its advantages.


systems designed in that way tend to maximize spl in the 30-40hz range.


big spl in that region is probably subjectively more impactful in blind listening tests than extension down to 5hz or whatever.


the lower your tuning frequency, the lower your efficiency and maximum spl in the 30-40hz region.


it's all tradeoffs.


i hate to beat a dead horse, but the frequency response of the mastering studio must be considered. those guys are not running systems that are flat to 5hz. they rolloff; most of them as the jbl frequency response target. as a result, if you have a system that is flat to 5hz, you are just way overblowing the subharmonics. and while this may be fun, it is a distortion. a closing car door should not sound like a dinosaur stomping the ground.


ok, flame suit on.

This is more of the same nonsense.


You could start by posting some in-room FR graphs of any mix desk in any soundtrack studio, which you can't, so roll off of the various systems is pure speculation.


If you don't have a system with an in room flat response to 5Hz, how would it ever be possible to accurately replay a recording of the closing of a car door?


Conversely, if you do have such a system, how is playback of the recorded event of the closing of a car door 'overblowing the subharmonics'? This belief assumes that the recording was mixed with the sub harmonics grossly boosted, in which case, you'd still be playing it back accurately.


Dinosaur foot stomps may or may not be exaggerated because there's no way to know what the spectra of the actual event is, but the spectra of the sound of black hawk rotor blades is well known.


There is no advantage to a home playback system that's EQ'd to boost any part of the BW by 6-10dB, other than personal taste. "Big SPL" in any BW should be mentioned in the context of headroom, not a house curve.


What's required is a spectrogram of the system's playback vs one taken from the disc. Keith Yates performed this test and included those graphs in his WDD article. None of the subs accurately reproduced the subsonic content, but he recorded the playback outdoors.


Bosso
 
61 - 80 of 154 Posts
Top