Post by: Patrik T on September 02, 2010, 12:46:24 PM
My latest adjustments have resulted in the following measurement (which is basically my first ever):
Should I bother a LOT more about the Hula-hula between 80-200 Hz?
My intentions are to open up a small indie mastering shop pretty soon. I've been doing things in the underground for a few years now but will try to make it all a little more ongoing and official sooner than later.
Best Regards
Patrik
Post by: AndreasN on September 03, 2010, 11:28:50 AM
Post the result here if you're unsure what it means and we can get into the particulars of reflection hunting. Ideally one graph showing the total response from 0 to some hundred milliseconds (depending on overall decay lenght in your room) and another graph showing the 20-30 first milliseconds, with level adjusted for maximum detail.
Post by: Patrik T on September 03, 2010, 02:44:46 PM
Post by: Patrik T on September 03, 2010, 02:47:30 PM
Best Regards
Patrik
Post by: AndreasN on September 04, 2010, 08:54:33 AM
Anyone else have any qualified guesses?
PS: are you happy with the modal response as seen in the waterfall plots?
Post by: Patrik T on September 04, 2010, 11:27:53 AM
This is one of the first measurements I've ever done in any room but there should be no correction going on in any way. I did no loopback-thing and just made sure that the out or in did not clip.
I'll run a few more sweeps when time allows. The reason I posted the first plot was how surprisingly even it looked. I've just been moving things around by ear before.
Will dig deeper into the waterfall when "time says OK".
Best Regards
Patrik
Post by: bruno putzeys on September 04, 2010, 11:59:24 AM
| AndreasN wrote on Sat, 04 September 2010 14:54 |
Hmmm... That was confusing. The starting portion makes me wonder. It's usually a steady noisefloor followed by an abrupt transition to the impulse at 0ms. Your graph starting climbing upwards at 50ms ahead of the impulse. Do you run room correction or something similar on the signal? Is there maybe something that introduces a lot of distortion? Wrong samplerate? Something something? |
Having written my own measurement program from scratch I can tell you it's remarkably difficult to make a sweep-based (a la Farina) method that has no pre-ringing. I cracked it but it wasn't obvious. So you're quite likely to see something like this in a sweep based test whereas it wouldn't appear in an MLS based one.
Post by: AndreasN on September 04, 2010, 01:50:39 PM
| Patrik T wrote on Sat, 04 September 2010 17:27 |
Hi I'll run a few more sweeps when time allows. The reason I posted the first plot was how surprisingly even it looked. I've just been moving things around by ear before. |
Try recording something normal, like yourself, to check that everything's alright. Make some impulsive sound to check transient response. Talk both in front and behind the mic to check omni response. Whistle to hear if there's aliasing. The sine sweep will sound weird if there's something very wrong on the output side. Have had occasions where there was alias distortion from non-SRC'ing sound cards at wrong samplerate. Sounds like a flying saucer attack.
Should be a pure tone.The graphs show a very loud measurement level and/or very low noise floor. Is that right? Do you have an SPL reading for the measurement?
(apologies if this is all way too obvious! better too much than too little..)
| bruno putzeys wrote on Sat, 04 September 2010 17:59 |
Having written my own measurement program from scratch I can tell you it's remarkably difficult to make a sweep-based (a la Farina) method that has no pre-ringing. I cracked it but it wasn't obvious. So you're quite likely to see something like this in a sweep based test whereas it wouldn't appear in an MLS based one. |
Wow! Cool stuff. Know what you mean. The sinesweep->impulse way isn't particularly clean as usually implemented. Cudos for "cracking it" (however you did it!).
50ms is a _lot_ of preringing though. Had a look a couple of different room measurements I had at hand and most of them rise from noisefloor some hundreds of microseconds ahead, to some milliseconds. One measurement was out of normal at 7 milliseconds. Nothing like the steady climb from 50ms in these graphs.
Regards,
Andreas Nordenstam
Post by: bruno putzeys on September 04, 2010, 02:16:28 PM
@Patrik
On another note, the response measurement in the OP shows a very wide dip centered around 3kHz, and about 6dB deep. This worries me a bit because if this is actually there this is also going to affect your EQ'ing judgment. You should check with a gated measurement (just using the anechoic portion of a measurement done on one speaker) if this effect is just your room averaging out the power response or whether it's visible in the on-axis response of your monitors too. If it is only the power response that's got this dip it's not as bad (though still worth being careful about when making EQ decisions) but if the on-axis response has the same dip you should consider correcting it.
Post by: Patrik T on September 04, 2010, 06:35:03 PM
| bruno putzeys wrote on Sat, 04 September 2010 20:16 |
On another note, the response measurement in the OP shows a very wide dip centered around 3kHz, and about 6dB deep. This worries me a bit because if this is actually there this is also going to affect your EQ'ing judgment. |
Hi Bruno, thank you for your input.
I spotted this too and thought it could have to do with the speakers crossover which is stated to be at 2 Khz.
I'm not that worried since it is not a common pattern here to add dB's of wide bells in that area, but I am aware that there doesn't always need to be a pattern like that to highlight errors either.
| Quote: |
You should check with a gated measurement (just using the anechoic portion of a measurement done on one speaker) if this effect is just your room averaging out the power response or whether it's visible in the on-axis response of your monitors too |
This measurement was with one speaker at my rough position of my head. I'll do some more sweeps with more careful attenton to the mic's placement.
While seated at a position where everything is reachable I can still move my head towards this-and-that direction without finding that things are altering considerably much.
| Quote: |
If it is only the power response that's got this dip it's not as bad (though still worth being careful about when making EQ decisions) but if the on-axis response has the same dip you should consider correcting it. |
Correcting by what means, if that should be the case?
Best Regards
Patrik
Post by: AndreasN on September 05, 2010, 07:28:34 AM
| bruno putzeys wrote on Sat, 04 September 2010 20:16 |
The trouble is that the pre-ringing is at low frequencies, which is why it is that long. The sweep method uses an exponential sweep, which by necessity has to start somewhere (you can't start an exponential sweep at DC). The impulse you get after deconvolution is therefore high-pass filtered, and not necessarily minimum-phase. Hence the pre-ringing, unless you modify either the stimulus or the deconvolution kernel to add just the right amount of phase shift to make the high-pass portion minimum phase. |
Don't know what's going on here, but it looks way odd compared to the usual sort of results from the same measurement methodology.
Here's a normal test result using sweep as the excitation signal: http://www.gearslutz.com/board/attachments/studio-building-a coustics/166868d1270651761-lede-room-haas-trigger-haas2.png
The flat frequency response doesn't look particularly real either. My guess is that it's a response as measured across a portion of time that includes a lot of noise. The inclusion of the noise smooths the overall graph.
| bruno putzeys wrote on Sat, 04 September 2010 20:16 |
You should check with a gated measurement (just using the anechoic portion of a measurement done on one speaker) if this effect is just your room averaging out the power response or whether it's visible in the on-axis response of your monitors too. |
How does one get an even remotely close to accurate "anechoic" response without resorting to time domain spectrometry?
Post by: Patrik T on September 05, 2010, 08:23:44 AM
| AndreasN wrote on Sun, 05 September 2010 13:28 |
The flat frequency response doesn't look particularly real either. My guess is that it's a response as measured across a portion of time that includes a lot of noise. The inclusion of the noise smooths the overall graph. |
If this would be the case, what does that say about noise? I doubt that mic pre or mic noise would flatten out what acoustic treatment can't when measuring and then looking at a graph like this, but who knows...
When I look at my zoomed out ETC and try to compare it to yours I miss seeing where your floor is at. Do you have a more out-zoomed ETC?
The room was dead silent at the sweeping and the sweep sounded truly clear to my ear.
I will sweep a little more tomorrow and see what happens.
Best Regards
Patrik
Post by: bruno putzeys on September 05, 2010, 08:37:46 AM
The Y scale in the plots you refer to is stretched out wider. Anyhow, my main point is that measurement programs don't all do exactly the same thing. My sweep test doesn't have pre-ringing, and Farina's later paper mentions other means of reducing the problem.
Getting the anechoic response simply means chopping off everything starting from the first reflection. In a real room this means you get only a few ms to work with, so the measurement is only meaningful from a few 100Hz upward. But it would settle the question whether the anechoic on-axis response has a dip too.
@Patrik
Although even the best crossover filters are likely to have dips in the power response, only the worst would have them in the on-axis sum. I know PMC and they are normally much flatter than this so it is worth putting some attention into finding what's going on. By correction I meant using an EQ but finding the root cause is going to be more effective.
Post by: Patrik T on September 05, 2010, 08:54:41 AM
| bruno putzeys wrote on Sun, 05 September 2010 14:37 |
I know PMC and they are normally much flatter than this so it is worth putting some attention into finding what's going on. By correction I meant using an EQ but finding the root cause is going to be more effective. |
Thank you Bruno!
I do not like the concept compensational EQ for monitoring, so let's just hope that it's either the measurement that's a little off or something that can be adjusted by acoustic means.
To my ear things sound good.
Best Regards
Patrik
Post by: AndreasN on September 07, 2010, 09:20:47 AM
| bruno putzeys wrote on Sun, 05 September 2010 14:37 |
The Y scale in the plots you refer to is stretched out wider. Anyhow, my main point is that measurement programs don't all do exactly the same thing. My sweep test doesn't have pre-ringing, and Farina's later paper mentions other means of reducing the problem. |
Get what you're saying, just pointing out it's a bit too much in this case. At least compared to every other measurement I've ever seen. Except the ones where there's been some gross error included in the calculation!
| bruno putzeys wrote on Sun, 05 September 2010 14:37 |
Getting the anechoic response simply means chopping off everything starting from the first reflection. In a real room this means you get only a few ms to work with, so the measurement is only meaningful from a few 100Hz upward. But it would settle the question whether the anechoic on-axis response has a dip too. |
If it only was as simple as being simple.
It could give a decent hint yes, but it's pretty hard to get an "anechoic" response from a normal measurement in a normal room. You probably know this better than me; think it's still worth mentioning for the casual reader that the response as seen in the frequency domain is the (convolved) sum of the original response and the effect of window function. Normal windows have all their brute force going on in magnitude only, with zero regard to phase. This works fine if one is only interested in the magnitude response in the FR. When phase is an issue, as it is in acoustics, this may throw off the observed results.
Post by: bruno putzeys on September 07, 2010, 11:07:38 AM
A few ms of echo-free response is quite sufficient to get magnitude and phase data with a resolution of a few 100Hz which is rather sufficient to verify if there is a wideband dip at 3kHz.
Post by: Patrik T on September 07, 2010, 11:12:49 AM
The 3 kHz things is still there. I swapped speakers, no difference. I measure with an "argh-cough-argh" ECM8000 through converters and preamps from the Lavry Blue line. Could it be the cheapo microphone?
Other than that I've realized that I will think about putting some lo-mid Pacman (or fatman) at between 3-3.8 meters away, there is place there for absorbers. Seems the only way to eat the 114 Hz is to go further away.
Edit: Just realized it is 1/12 octave smoothing here, but what the heck...
Best Rgeards
Patrik
Post by: Patrik T on September 07, 2010, 11:43:50 AM
Post by: Patrik T on September 07, 2010, 11:47:05 AM
Post by: Patrik T on September 08, 2010, 03:11:58 PM
Post by: Patrik T on September 08, 2010, 03:16:26 PM
Regards
Patrik
Post by: Patrik T on September 09, 2010, 06:31:33 AM
The AKG indicated the 114 Hz thing though, and my forthcoming adjustments will be to ease down that one, if possible. It is really stubborn.
Best Regards
Patrik