How to accurately measure an FR - diyAudio
Go Back   Home > Forums > Loudspeakers > Multi-Way

Multi-Way Conventional loudspeakers with crossovers

Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 6th December 2004, 08:57 PM   #1
diyAudio Member
 
Join Date: Oct 2004
Default How to accurately measure an FR

Hello

Recently, during my efforts to improve a xo, I discovered kxproject audio drivers for my soundblaster Audigy2. For those who are not aware it's a free driver aimed for EMU10K1 and EMU10K2 based soundcards that lets you exploit all the functions of the soundcard to a level where you can actualy write your own dsp's using simple machine-level language. I was only interested in the frequency splitting functions and the built in 10 band equalizer so that I could digitally split the frequency spectrum and drive directly my 2-way diy using two stereo amplifiers.

I setup my soundcard using the front speaker output for the tweeter and the rear output for the mid-woofer. The crossover point as well as the output levels were optimized using the ECM8000 microphone and the UB802 pre-amp from Behringer. I was amazed with the improved detail and clarity I got from my speakers. I started hearing things that for my system never existed before. Even the image got way better both in size and detail. I never thought that a (bad?) passive xo would have such a bad impact to the final sound. Impedance, phase, and delay issues caused by passive xo were now history! I thought that this is it, but unfortunately I got my self into another circle of problems.

The soundcard dsp lets you add notch filters of any kind as well, so I begun messing up with the frequency response to make it as flat as possible. I'm measuring the speakers in the normal listening environment. That is a small living room with thick carpets, curtains, wooden floor, concrete walls and lots of heavy furniture. I've been measuring the speakers using a lot of different combinations of microphone distance or height and everytime I got different results. For example, a near field measurement would give me pretty accurate results for the low frequencies, which after an optimization from the soundcard would sound pretty natural. But what about the mids and highs? Where do you place the mic? At tweeter or woofer height. Perhaps in between? and in what distance? What should be the maximum allowed angle between the center of the unit and the tip of the microphone?

I tried measuring at the height of the tweeter and at a distance of 0.2m using time-gated analysis and after equalizing the output to get a flat FR I found out that from the usual listening position (3.5m) the highs were attenuated a lot. On the other hand, placing the mic between the two units and at a distance of 1m would produce an FR that lacks of highs especially above 12kHz. As a result, after flattening the FR the highs would sound over-boosted. Of course measuring anything beyond 1.5m inside a normal living room is impossible due to the reflection and the fact that still today there is no way to accurately predict them.

I could be writing pages with my experiences so far but I think you can understand my point. So my question is, what can you do when you don't have an anechoic chamber, or a super-flat microphone, or a sphere-shaped living room to get a realistic FR for frequencies above 300-400Hz? How would you perform your measurements if it was to be in the same conditions? Do you feel confident that the FR you have represents what goes into your ears? Or that your supposably flat system is trully flat?

Stelios
  Reply With Quote
Old 7th December 2004, 01:15 AM   #2
claudio is offline claudio  Italy
diyAudio Member
 
claudio's Avatar
 
Join Date: Feb 2002
Location: Italy
Why not using Moller ambient curve as a reference?
Adjusting your DSP to reach that curve should be definitive.

regards

Claudio
  Reply With Quote
Old 7th December 2004, 11:02 PM   #3
diyAudio Member
 
Join Date: Jun 2004
Location: Great Britain, South
Default Re: How to accurately measure an FR

Quote:
Originally posted by maoumaou
I could be writing pages with my experiences so far but I think you can understand my point. So my question is, what can you do when you don't have an anechoic chamber, or a super-flat microphone, or a sphere-shaped living room to get a realistic FR for frequencies above 300-400Hz? How would you perform your measurements if it was to be in the same conditions? Do you feel confident that the FR you have represents what goes into your ears? Or that your supposably flat system is trully flat?
Stelios
Fun isn't it? The first time you ever plug a measuring mike into an oscilloscope (or whatever pc equivalent) indoors and *see* what the mic is 'hearing' from your speakers compared with what the amplifier is sends to the speakers, it's likely you'll be utterly appalled. From the *massive* difference between the two waveforms, and the way the difference chnages so much on moving the mike, you'd be forgiven from believing that any kind of quality sound reproduction is an impossibility. Which it would be if not for our ears/brain which do a very nice job (thank you) of making sense of and enjoying the music/speach etc they easily extrapolate from that mess of relections, refractions, standing waves, phase differences etc etc caused by the room and furnishing around you :-)

Having said all that, you can still use measured frequency responses - for instance to compare a pair speakers after one was repaired or to compare a driver in different types of enclosure etc - providing you are scrupluous about your experiment method (mics fixed in position, marks on floor to show speaker positions, no changes to room (an extra person, curtains open or shut, door closed or open etc) between tests. Then the two graphs you have are useful to make that comparison. Yes, the room has a massive effect - but it's the same massive effect on both units under test - so other differences are what they bring.

No freq response graph will ever be flat in any case. I'd use my ears rather than graphs drawn on a PC to set up any sound system to best advantage in a location - with the knowledge of it's intended use.

Nice to have real oscilloscope and spectrum analyser to play around with, especially for complex active crossover spilts - and both can be very helpful in certain difficult circumstances - but never just rely on the equipment. Ears should be used as a check on instruments (esp instruments than run on PCs!) at all times :-)

tekno.mage
  Reply With Quote
Old 8th December 2004, 08:49 AM   #4
diyAudio Member
 
Join Date: Oct 2004
Quote:
Originally posted by claudio
Why not using Moller ambient curve as a reference?
Adjusting your DSP to reach that curve should be definitive.

I have no idea what a Moller ambient curve is. Any help? Thanks in advance.




tekno.maqe

You are right. I agree that it's rather impossible to do a perfect measurement. However, relying on your ears can be misleading most of the time since ears seam to adjust to different situations.

Have you ever noticed that listening to music through your system after going to a night club gives you the feeling that the sound lacks of bass? Or have you ever noticed that a specific song might sound different from day to day? I red once in a magazine that the ear (and the brain) tries to give you the impression that the sounds you hear are as real as possible even though it's only a reproduction from a non perfect system and a non perfect recording. They also said that a bad system makes you tired after 30' of music listening because of that reason.

Another thing is that the ear cannot tell you what frequencies you are listening to. That's the best advantage in measurement equipment. I'm listening to my test cd's and I hear something irritating at mid-high's. Exactly which range of frequencies should I be playing with to equalize them? Even if I blindly started adding notch filters it would get soooo complicated and confusing that in the end you get nothing but a headache

Regards

Stelios
  Reply With Quote
Old 8th December 2004, 02:36 PM   #5
claudio is offline claudio  Italy
diyAudio Member
 
claudio's Avatar
 
Join Date: Feb 2002
Location: Italy
Default Moller curve

This is the ideal in-room Moller curve (pink noise, 1/3 oct.)
Attached Images
File Type: jpg curva moller.jpg (80.7 KB, 194 views)
  Reply With Quote
Old 8th December 2004, 05:42 PM   #6
diyAudio Member
 
Join Date: Jun 2004
Location: Great Britain, South
Quote:
Originally posted by maoumaou


You are right. I agree that it's rather impossible to do a perfect measurement. However, relying on your ears can be misleading most of the time since ears seam to adjust to different situations.
Stelios
Yes, relying on *just* ears with no other points of reference would be very misleading - probably even for musicians with perfect pitch.

However, if you have the opportunity of setting up a variety of different systems in different locations, and are used to seeing audio displayed on a scope and other test equipment you'll find that you very easily train yourself to hear when some frequency ranges are 'out of balance' and can be controlled using and equaliser, active xover etc. I might not be able to tell you exactly what frequency a sine wave is, but I will know well enough what frequency range on my xover or equaliser would need adjusting to affect it. The other thing to bear in mind when using music rather than test signals is that you must use pieces you are very familiar with. I doubt anyone can tell much about a sound system (other than recognising gross faults such as clipping) if listening to an unfamiliar piece of music! I have a group of certain pieces of music I always use as a final test after setting up a system with pink noise etc - and they are chosen for two reasons - I know them all very well, they are all very different and all tend to show up certain common 'faults' with sound system setups (ie a recording of a female unaccompanied voice that shows up even the slightest amount of sibilence.)

Regarding your comment on listening to your home system after a nightclub system, (and having set up sound systems in nightclud-type venues myself) my usual reaction on hearing my own home setup afterwards is one of relief! Admitted, the cause of this is quite simple. Many nightclub systems (however well setup initially) are commonly driven into distortion by their users :-( Someone on this forum uses a signature something like "Loud is good if it's clean" - I'm all in favour of that. However, your neighbours may be a bit upset if your home system *did* reproduce bass the way a good nightclub system should :-)

You are quite right in your remark about bad systems being 'tiring' to listen to - and it gets worse the more experience of different systems you have. I simply can't tolerate listening to some bad car stereos or cheap all-in-one hifis now, but I'm sure I was quite happy listening to far worse systems when I was young!
  Reply With Quote

Reply


Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
Tweeter Zobel, or more accurately, RC network. LNeilB2 Multi-Way 42 9th June 2014 06:41 AM
How to ACCURATELY integrat subs with mains (and a little challenge heh heh) terry j Multi-Way 22 29th September 2008 10:04 PM
Speaker Workshop - using "Measure -> Passive Components" to measure ESR ? percy Multi-Way 0 12th March 2006 10:18 PM
Measuring electronics frequency response accurately + cheaply eRiCdWoNg Everything Else 1 28th April 2004 12:24 AM


New To Site? Need Help?

All times are GMT. The time now is 01:36 AM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio

Content Relevant URLs by vBSEO 3.3.2