Rerurn to Romy the Cat's Site

Audio Discussions
Topic: Redding the RTA data requires some skills.

Page 1 of 1 (10 items)


Posted by stuck.wilson on 08-18-2007

hey romy and all-

i recently came across a strange phenomenon (to me.. maybe not to everyone else!).

I just started using RTA as a method of evaluating my results with my own 2 way speaker system, and I'm finding that I'm ending up liking the sound of my system a GREAT DEAL MORE without the absurd notion that 20-20k is what we 'need to hear'.. but it seems even more extreme than this.

primarily-- the frequency response of my system, as measured when I really appreciate it's sound, is relatively flat within the parameters of the altec 414 i'm using as the fundamental driver for below 1k.  so until approximately 2k, the system is flat-- but with a small bump in the midbass: not perfect-- but appreciable listening considering my budget and absolutely minimal space.

where it gets strange is-- after 2k-- if i'm still pleased with the sound-- rolls off DRAMATICALLY-- where i'm at LEAST 10db down by 10k, and further from there-- to where it looks as if it's rolling off a total of almost 20db from 2k- 20k!  when i lessen the reins on the horn-- the sound is SO agressive and unlistenable (on a BMS 4552nd -- the opinion of which you've shared) that i can't stand it.  upper midrange HELL. period.   so it's all fundamentals, essentially, that i seem to appreciate most!  audibly-- i don't feel as if i'm missing ANYTHING with this configuration, and i grew up playing in orchestras and listening to the same!  not that that makes me in ANY way an expert.. that's for sure!

the BMS driver, you've described as blanched and horrible.. and if measured 'flat'.. i've got your back on that.  but with a crossover made of vitamin q 100v'ers running first order with well over an octave from the lower end of the exponential horn i've got it on.. it can be, for it's modest upbringing, pretty pleasant on the whole. if attenuated 'properly' (to my ear)-- it's devoid of the  metallic shrieking and nastiness.. but of course.. it's missing TONS of information, if the measurements are worth a damn.

the only thing i DO appreciate about the absurd overblown 'flat' sound is the addition of time cues.. but they're not worth it at the expense of.. well.. my ears.  granted.. i'm pushing the driver over 6 octaves, which isn't optimal by ANY stretch, and it's lending to the 'TV speaker' effect..

so i wonder if it's the 'chicken' of 20-20k being a myth, or the 'egg' of my misuse of a 2 way horn/direct radiator... a BUNK driver,  or in fact, just all three. OR.. it's a matter that the RTA and/or my microphones are just THAT inaccurate.  or i'm a caveman.. which is also possible (!) Wink.

any light to shed on this phenomenon?  i realize that there could be an incredibly complex bevy of factors at work here..  but i wonder if anyone else has had a civilized experience with listening this way.

thanks for your input-

dan

Posted by drdna on 08-18-2007

Now first of all our ears were designed to hear up to 20,000 cycles there-abouts if not perhaps more, a scientific fact I can't really argue against.  So I find it difficult to accept a proposition that we don't enjoy listening to sounds from 2,000 Hz on. 

What seems much more likely is that you don't like the way that your tweeter sounds, and the more tweeter you hear, the more intolerable you find the music.  So you are rolling it off to avoid the horrible screeching. 

Could the rta be inaccurate?  Maybe.

You could try substituting other speakers or other tweeters and seeing if you still have the same reaction.

In any case, just tune your system the way you like it best. 

Adrian


Posted by Romy the Cat on 08-18-2007

 stuck.wilson wrote:
I just started using RTA as a method of evaluating my results with my own 2 way speaker system, and I'm finding that I'm ending up liking the sound of my system a GREAT DEAL MORE without the absurd notion that 20-20k is what we 'need to hear'.. but it seems even more extreme than this.

2 way speaker system can not be 20-20k. It can hardly be 63-12.5K…

 stuck.wilson wrote:
where it gets strange is-- after 2k-- if i'm still pleased with the sound-- rolls off DRAMATICALLY-- where i'm at LEAST 10db down by 10k, and further from there-- to where it looks as if it's rolling off a total of almost 20db from 2k- 20k!  when i lessen the reins on the horn-- the sound is SO agressive and unlistenable (on a BMS 4552nd -- the opinion of which you've shared) that i can't stand it.  upper midrange HELL. period.

It might be because many reasons but never take electricity out of accounts. The 4552nd has ligh cone and they uselessly becomes very flat but aggressive with bad electricity.

 stuck.wilson wrote:
so i wonder if it's the 'chicken' of 20-20k being a myth, or the 'egg' of my misuse of a 2 way horn/direct radiator... a BUNK driver,  or in fact, just all three. OR.. it's a matter that the RTA and/or my microphones are just THAT inaccurate.  or i'm a caveman.. which is also possible (!)  .

A horn is 3-4 octaves devise. The 20-20k is not a myth but to get  it properly is difficult and not only at speaker level...

http://www.goodsoundclub.com/Forums/ShowPost.aspx?PostID=755

 stuck.wilson wrote:
any light to shed on this phenomenon?  i realize that there could be an incredibly complex bevy of factors at work here..  but i wonder if anyone else has had a civilized experience with listening this way.

Actually, Stuck. I do nor exactly understand what you are asking…

Posted by stuck.wilson on 08-19-2007
thanks for the responses doc and romy-

much food for thought-  but what i'm thinking, or at least extrapolating from what you're saying is that the possibility exists that i AM measuring what i think im' measuring.

i have a good friend who is very much taken by the western electric phenomenon-- and his take has ALWAYS been very similar to yours in some ways- that outside 60-12.5khz, theres information, but it's not as necessary for gratifying listening (pardon if i misquote)- particularly if it's badly reproduced, in which case it's just destructive to the rest.  i don't have anywhere near the funding or space or inclination for his particular affliction of historical, oversized, and outrageously priced ancient stuff--  but this experiment has shed a little light on that.

what i meant by 'civilized listening' was moreso asking what you reiterated in the link-- if you can't recreate the extremes, can you still recreate a facsimile of MUSIC with solidly reproduced fundamentals?  what am i missing in the upper 2 octaves-- is it primarily all ambient information and upper harmonics that are seldom recorded well to begin with?

i've been looking into the possibility of expanding to a 3 way system-- but these questions are primarily the initial theoretical 'legwork' to ask if the upper octaves are really a necessary evolvement, as they're oft times a lot more difficult, expensive and IRRITATING to listen to- not to mention-- in what i seem to hear- not always present in the 'proportions' of flat 20-20k when at a concert hall or playing music (thanks, mrs fletcher and munson...). and given your caution regarding electricity in a town with ancient aboveground electrical infrastructure, constant brownouts and outages..  it could be a monster set-up for investing in another unused driver in the basement!

i'm going back to remeasure just my horns today-- there are a few more options i'd like to try-- shifting my crossover point and reattenuating, and cobbing in a decent sounding silk dome tweeter that's way less efficient, but decent sounding just as an experiment.  neither is perfect nor necessarily feasible-  and i'm not sure that the bms is going to be a usable driver in the least--  but i certainly appreciate the roadsigns, gentlemen!

thanks again!

dan

Posted by Romy the Cat on 08-19-2007

Dan, I do not know where you have seen a lot of food for thought, all posts above were about nothing, as I said I really do not know what this thread is all about. If your interest is the relationship between measurable and auditable and extreme frequencies then it is a noble subject but I don't think it is what you are talking about. You were talking about general applications of RTA, but I generally have difficulties is to talk with audio people about RTA.

Generally a RTA’s are very accurate and the microphones are the one which create discrepancies.  Getting at good-quality an instrumental calibrated microphone is mandatory.  An alternative way would be to build a correction chain in mic-preamps and to calibrate against reference microphone… however a properly calibrated microphone and RTA are just very beginning of the entire RTA saga.

Audio people attribute to RTA some kind of supernatural meaning, without spending efforts toward interpreting the RTA measurements. RTA measurements themselves are completely bogus and irrelevant unless they are viewed in intelligent format. Where to measure? How to measure it? What kind signals to use? How to correlate all conditionals parameters with what RTA outputs? And of course the most important is how to or correlate the result of RTA measurements we is listening experiences of the given playback installation. I do not know how for others but it took for me approximately 5 years to measure and to listen until I begin to understand what is really important in the RTA interpretation.

Also one very-very important moment that unfortunately many audio people do not “get”. RTA shows just amplitude but does not show the reasons for that amplitude.  We, as humans of course register amplitude but we also are capable to recognize and to be influenced by them means by which the amplitude was reached. That is very imperative concept because the correct understanding of amplitude is possible ONLY in context of the methods which created that amplitude. So, answering the questions that you did not ask: the linearity of frequency response is very positive and very important thing but only if the linearity was achieved by only “proper ways”.

The Cat

Posted by op.9 on 08-19-2007
Dan, are you using the Behringer DEQ2496 for your RTA? If so, you could be experiencing the same problem I had. Once in a while its frequency range goes doolally and shows rubbish - but not quite enough rubbish to make it obvious. (I think it reads a multiple of each frequency) - Perform a hard reset and all is well until the next time...

Posted by stuck.wilson on 08-19-2007

romy-

 i have resisted using RTA for the very reasons you're citing-- learning to read the visual data, as well as methods of how to get it accurately, are as equally time consuming as any other pursuit in audio.   anyone can be a reader of meters-- the scientist is a whole other level of sophistication.  i am working through the bugs of being the meter reader-  however, i come to it with the understanding that the top octaves of RTA are seldom very trustable without very solid gear.    the microphone  (an akg 451) is an industry standard-- the mic pre is a mackie mixer-- nothing out of the ordinary.. but certainly respectable- the program is trueRTA- and as to it's accuracy, i don't know. (thanks for the suggestion op.9-- no behringer though!)  hearing that RTA is often an accurate tool,  that my particular driver has the capacity to go shrill with bad power, that in this application, expectations of exceeding 12.5k aren't really reasonable, and that maintaining full frequency evenly is tough for any system- not only a speaker system-- all of those strike a resonant chord validating what i'm experiencing-- so none of that is nothing insofar as 'food for thought' is concerned!  and you are speaking to EXACTLY what i'm getting at in terms of the auditability vs. measurability of frequency extremes, in an oblique manner- hence asking-  is 'flat 20-20k' a myth. 

 i ask this question using RTA as one of the quirks in my experience butting heads with 'common knowledge':  is it really necessary to have HF at the same amplitude as fundamentals to seem 'natural', or much like in concert halls, does the frequency response at your seat change from, say, the conductor's position due to proximity-- so that at some seats in the house, there is less HF information?  BECAUSE-- what i am hearing (which i trust more than my measuring)  sounds distinct, in scale,  and natural to my ears-- but only when my measurements (taken in multiple positions- both from mono and stereo sources)  read as non-linear 20-20k- or more precisely- downwardly linear past 2k at about 6db/octave. 

no doubt, it's again, the lack of really good systems as a yardstick- and the relative vaccuum of people experimenting in this vein that i know personally.  distinguishing between a 'very refined upper midrange and HF' at the proper amplitude and one that's turned down as to be less audible isn't a distinction i have the listening ability to make, presently.   but asking the question as to the proper SCALE of that frequency range seems an important part of the question, to me. it's difficult to ANSWER..i'd have to hear it, obviously.  but regarding the proper use of the tool-- i think you're right-- i don't attribute magic to it.. maybe i'm using the software incorrectly!

only thing that will really answer my questions will undoubtably be more dicking around and listening- and getting the 'benefit' of adding the top octave- right or wrong- to see what it does to the sound! 

regardless-- thanks for the sounding board to keep the process intellectually honest.

d


Posted by Paul S on 08-19-2007
I don't suppose I get this, either.

Why would you ever on the conceptual level opt to NOT have "Full Range" and a "flat frequency response" that covers the whole audible spectrum?

Sure, this is easier said (or thought of) than done, it's "complicated", etc., etc; but why would FR NOT be the ideal?

RTA is just a tool, like the equipment we want to test with it.

If you are saying you are measuring stuff that your playback source/system is not "really" contributing, then you can either re-work the RTA or plug away until you learn what that means in practical terms.  I seriously doubt if "extra HF" you are "measuring" is music...

I'm guessing that no system output above 2k is quite audible as an abberation that cannot be "balanced" with LF, any more than you can make a table radio (or older console) sound "realistic" enough to do serious music.

Meanwhile, how do problems with test equipment change The Goal, in the abstract sense, anyway?

Add range up to FR, for Pete's sake.

And good luck with it ;>Wink

Best regards,
Paul S

Posted by stuck.wilson on 08-19-2007

no- there's no reason why it wouldn't make sense NOT to give 'er a crack-- at worst, i wouldn't like the results!

i guess the difficult part really is sorting out measurement vs. hearing-  tricky at best!  trying to assemble sense from possibly faulty measurements leads in alot of directions-- what needs to happen, i s'pose, is very careful calibration, which i don't have currently..   and that in itself would render all these questions moot, because i'd have a standard to measure by outside of my ear!   my apologies to all for a half formed question out of process!

meanwhiles-- there IS output well over 2k for certain- that's the knee of where it begins to decrease by about 6db/octave.  it's extrapolating whether it's a measurement error i'm dealing with or that the de facto standard isn't necessarily perfect that i'm getting at. 

onwards!

d.

Posted by drdna on 08-19-2007

Fundamentally, you must examine the RTA with a system with a known frequency response to know if it is accurate, or you must obtain a second RTA from a friend which is known to be accurate, to compare against your own RTA.

However it is always a matter of interpreting the RTA results.  One goal is to have a flat frequncy response, and we are inclined toward this by the RTA.  But this has only a little to do with sound since frequency response is only one of many factors going into accurate musical reproduction. 

Just don't fall into the trap that there must always be some correlation.  You can have a flat frequency response and still have bad sound, therefore bad sound doesn't necessarily correlate to frequency response abnormalities.

Likewise, improving frequency response will not necessarily improve the sound, since it may simultaneously make something else worse.

Adrian


Page 1 of 1 (10 items)