All that is too hard for me.
I just burn a CD with white noise (NOT pink) noise and play it on my system while recording it with a microphone into the soundcard using CoolEdit '96 or whatever. Then I do a Frequency Analysis of the recording. Easy-peasy.
Make sure your microphone input on the soundcard is actually flat though.
I just burn a CD with white noise (NOT pink) noise and play it on my system while recording it with a microphone into the soundcard using CoolEdit '96 or whatever. Then I do a Frequency Analysis of the recording. Easy-peasy.
Make sure your microphone input on the soundcard is actually flat though.
If you use white noise as a test source I'd be worried about putting excess power into the tweeter. If use use an RTA then a pink noise source will give a flat response, and is much less likely to do any tweeter damage.
I use TrueRTA in its 1/24th octave mode, with an ECM8000 and a home made pre-amp/phantom power source and it works extremely well.
I use TrueRTA in its 1/24th octave mode, with an ECM8000 and a home made pre-amp/phantom power source and it works extremely well.
Ouroboros said:
there ya go. white noise is not flat...
basically, if you are willing to spend any more you can get a much quicker, more accurate system.
true RTA w/ 1/24th octave capability = $100, ECM8000 = $50, mic pre w/ phantom = $50(home brew) - $!!! sound card = $30-$80...
as usual, you get what you pay for...but that's damn good bang for the buck....
White noise is flat, but if you measure white noise with a system that measures in fractional octave bands, you'll get a response that rises at 3dB/octave.
Use white noise if you're measuring with a spectrum analyzer, but pink noise if you're measuring with an RTA.
(But then this is basic textbook stuff, so I'd assumed everyone knew this.)
I'd still recommend the $99 version of TrueRTA, 1/24th octave analysis can be quite useful.
Use white noise if you're measuring with a spectrum analyzer, but pink noise if you're measuring with an RTA.
(But then this is basic textbook stuff, so I'd assumed everyone knew this.)
I'd still recommend the $99 version of TrueRTA, 1/24th octave analysis can be quite useful.
Re: Re: Radio Shack 33-2050 and 33-4050
yes, now i get ya.
it's not a very convenient method either way, but those values vary quite a bit and the meter is kinda hard to read...it's just a lot easier for me to center around 0 vu than hop all around...i'd go 21 for the added ease
gmilitano said:
yes, now i get ya.
it's not a very convenient method either way, but those values vary quite a bit and the meter is kinda hard to read...it's just a lot easier for me to center around 0 vu than hop all around...i'd go 21 for the added ease
Re: Re: Re: Radio Shack 33-2050 and 33-4050
Agreed, but I am cheap and this is after all DIY audio!
I got an SPL meter a while back, but have yet to make any real good use out of it. I want to try some measurements with a tube and SS amp on the same speaker and compare the results. It will be somewhat crude, but may have some interesting results.
BTW - how long are the tone tracks and break between tracks on the CD?
Cheers,
Gio.
noisenyc said:
Agreed, but I am cheap and this is after all DIY audio!
I got an SPL meter a while back, but have yet to make any real good use out of it. I want to try some measurements with a tube and SS amp on the same speaker and compare the results. It will be somewhat crude, but may have some interesting results.
BTW - how long are the tone tracks and break between tracks on the CD?
Cheers,
Gio.
Circlotron said:
Took =me= a little while to figure it out at first.
Pink for Pieces of spectrum.
Whitle for Whole spectrum.
Not quite.
Basically, white noise has equal energy per Hertz, so if you measure it with a moving filter of constant bandwidth (300Hz, 1kHz or whatever), you will measure it as flat. This is how spectrum analysers measure a signal.
Pink noise has equal energy per octave (or fractional octave), so if you measure this with a constant bandwidth filter, (like an SA), then you'll see a response that falls of at 3dB per octave. On the other hand, if you measure pink noise with an RTA, which measures in octave chunks of the spectrum (or 1/3 or 1/12th octave as required: ie, the bandwidth of the filter increases as the frequency rises), then you measure pink noise as flat.
Your pictures show an FFT-based spectrum analyzer, which will show the same results (more or less) as a conventional filter-based SA, therefore it shows white noise as flat. You are quite correct in saying that if you used this noise as the source for an analyser that measures in 1/3rd octave bands (or whatever fraction of an octave you want to use), then it will show a 3dB per octave (or to be pedantic, a 10dB per decade) rise in response. To show a flat response on such an analyser when measuring a flat system (like a perfect amplifier, or the mythical 'straight wire with gain') then you need to use pink noise as a source.
There's nothing to stop you using wideband pink noise as a source to measure the whole 20Hz to 20kHz in one go if you have a suitable wideband pink noise source and an RTA that covers the whole band (like TrueRTA does).
There's nothing to stop you using wideband pink noise as a source to measure the whole 20Hz to 20kHz in one go if you have a suitable wideband pink noise source and an RTA that covers the whole band (like TrueRTA does).
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.