ProSoundWeb Community
Sound Reinforcement - Forums for Live Sound Professionals - Your Displayed Name Must Be Your Real Full Name To Post In The Live Sound Forums => Audio Measurement and Testing => Topic started by: peter dakin on March 14, 2012, 01:29:38 PM
-
So I'm quite new to all this black magic, but I'm currently demoing Smaart7.
In my understanding, once you've time and phased aligned your system (well as close as you can ;)), is to then us the transfer magnitude window to give you a graphical representation of your system in the venue.
I've been taking multiple measurements around the venue, to create a average.
Below is an average from the weekend. Sadly the system was in-house and controllers all locked away so I couldn't do anything, expect delays and EQ.
(http://i39.tinypic.com/6hup6v.png)
7 measurements created average, Smoothing set to 1/24 Oct, with a time average of 1second
Now to use just the graphic representation, then I'd be reaching for the EQ to take 300-400, 1.5k, and the big lump around 10k.
However using the most important tools, (my ears); the EQ, to tame these sections was a little heavily handed and I ended up balancing EQ with my ears!
However here lies my question. Using this graphical representation I'd like to learn to read it a little better and then tune a system, to create a 'blank' canvas for guest engineers to work with, as most of my work is as a system tech.
What sounds good to me, might not work for them.
As I'm brand-new to the audio measurement, I haven't had a chance to measure a good sounding system and seen the subsequent magnitude trace.
Am I on the right approach, looking at above magnitude reading what would you do? Obviously the ears are the important factor here, and no two systems sound the same.. but applying the science, what would be the next step.
Thoughts appreciated.
-
So I'm quite new to all this black magic, but I'm currently demoing Smaart7.
In my understanding, once you've time and phased aligned your system (well as close as you can ;)), is to then us the transfer magnitude window to give you a graphical representation of your system in the venue.
I've been taking multiple measurements around the venue, to create a average.
Below is an average from the weekend. Sadly the system was in-house and controllers all locked away so I couldn't do anything, expect delays and EQ.
(http://i39.tinypic.com/6hup6v.png)
7 measurements created average, Smoothing set to 1/24 Oct, with a time average of 1second
Now to use just the graphic representation, then I'd be reaching for the EQ to take 300-400, 1.5k, and the big lump around 10k.
However using the most important tools, (my ears); the EQ, to tame these sections was a little heavily handed and I ended up balancing EQ with my ears!
However here lies my question. Using this graphical representation I'd like to learn to read it a little better and then tune a system, to create a 'blank' canvas for guest engineers to work with, as most of my work is as a system tech.
What sounds good to me, might not work for them.
As I'm brand-new to the audio measurement, I haven't had a chance to measure a good sounding system and seen the subsequent magnitude trace.
Am I on the right approach, looking at above magnitude reading what would you do? Obviously the ears are the important factor here, and no two systems sound the same.. but applying the science, what would be the next step.
Thoughts appreciated.
Before reaching for any eq you should try to figure out why those dips at 900Hz and 5.5kHz are happening all over your venue. Because of the drop in coherence it may be due to a reflection. You need to look at your measurement setup to see where a reflection may be common to all measurements. If you determine that the dips are in fact a reflection issue, and it is not caused by you mic placement but by something architectural, you need to see if that can be corrected. If it is a floor reflection try doing your measurements with a ground plane mic position to eliminate floor reflections. Eq will not be able to fix it. Leaving aside those 2 dips, 10K does look high and may benefit from some eq, everything else looks pretty good, and using your ears to tune is the right approach at that point. Try to avoid narrow band filters unless you really need them. The 10k thing might need a narrow filter, but if that 5.5k dip is not there, maybe not so narrow. In a real life venue, a slight overall down trend in magnitude response in generally OK. A horizontal response curve may sound thin.
Barking up the right tree.
Mac
-
Before reaching for any eq you should try to figure out why those dips at 900Hz and 5.5kHz are happening all over your venue. Because of the drop in coherence it may be due to a reflection. You need to look at your measurement setup to see where a reflection may be common to all measurements. If you determine that the dips are in fact a reflection issue, and it is not caused by you mic placement but by something architectural, you need to see if that can be corrected. If it is a floor reflection try doing your measurements with a ground plane mic position to eliminate floor reflections. Eq will not be able to fix it. Leaving aside those 2 dips, 10K does look high and may benefit from some eq, everything else looks pretty good, and using your ears to tune is the right approach at that point. Try to avoid narrow band filters unless you really need them. The 10k thing might need a narrow filter, but if that 5.5k dip is not there, maybe not so narrow. In a real life venue, a slight overall down trend in magnitude response in generally OK. A horizontal response curve may sound thin.
Barking up the right tree.
Mac
What Mac said :-)
Particularly, the areas of poor coherence in a multi-channel measurement are suspicious.
Also, 1/24 octave resolution is a little fine for venue measurements for my taste.
Consider 1/6 octave, continuing to look for trends.
Of course this presumes everything is correct within the sound system.
-
Before reaching for any eq you should try to figure out why those dips at 900Hz and 5.5kHz are happening all over your venue. Because of the drop in coherence it may be due to a reflection. You need to look at your measurement setup to see where a reflection may be common to all measurements.
Think this would have just been the room. It was a 4000 cap room, with wood floors, upper stalls wings, circle and upper circle. The average is just for the stalls area, as separate PA was used to cover circles.
...try doing your measurements with a ground plane mic position to eliminate floor reflections
By this I take it you mean stick it as close to floor? I'm currently using a DBX omni RTA-M mic, so was always worried if too close to floor it my cloud results?
If you determine that the dips are in fact a reflection issue, and it is not caused by you mic placement but by something architectural, you need to see if that can be corrected.
As the venue has predominately chamber/orchestral music in the venue, and has been "tuned" for said music, I think sadly they'll won't be doing anything in the future for the nasty rock 'n roll music that insists on playing there! ;)
Barking up the right tree.
Mac
Excellent. Woof woof! Thanks very much for the replies... Dips might just have been on the "locked-away" controller??
-
Consider 1/6 octave, continuing to look for trends.
cheers for the tip
-
cheers for the tip
Take your measurements again in a ground plane manner as Mac suggested. You can put a few coils in the cable and place the barrel of the mic on it, with the tip of the capsule touching the ground at an angle. This will eliminate any floor bounce reflections.
I'm having a hard time believing those dips exist everywhere, unless there is a polarity problem but even then, those frequencies would normally be unlikely candidates for crossover points.
Once that is addressed, you have something to work with. The good news is it appears to be voiced in a usable manner, i.e. rolling off in the HF. Yes, 10K needs work based on the measurement.
-
Take your measurements again in a ground plane manner as Mac suggested. You can put a few coils in the cable and place the barrel of the mic on it, with the tip of the capsule touching the ground at an angle. This will eliminate any floor bounce reflections.
I'm having a hard time believing those dips exist everywhere, unless there is a polarity problem but even then, those frequencies would normally be unlikely candidates for crossover points.
Once that is addressed, you have something to work with. The good news is it appears to be voiced in a usable manner, i.e. rolling off in the HF. Yes, 10K needs work based on the measurement.
Venue's seated, so might cause more issues trying to avoid reflections. Not too worried about this venue, it'll be quite a while till I'm there again. Sorry I wasn't meaning x-over points. Without being able to see the actual processor, I was assuming they might have some "interesting" EQ in place.
With this venue there are two areas covered by balconies, fed by the outer reaches of the arcs. These will probably cause havoc with the coherence, thus bringing the average down. But as they are covered by the main system, have to be taken into account.
-
Venue's seated, so might cause more issues trying to avoid reflections. Not too worried about this venue, it'll be quite a while till I'm there again. Sorry I wasn't meaning x-over points. Without being able to see the actual processor, I was assuming they might have some "interesting" EQ in place.
With this venue there are two areas covered by balconies, fed by the outer reaches of the arcs. These will probably cause havoc with the coherence, thus bringing the average down. But as they are covered by the main system, have to be taken into account.
I think you're not hearing what we're saying. Having your mic not be in a ground plane causes reflections off the floor to interfere with the measurement. Do a ground plane measurement. Use a sheet of plywood over the seats, or a console case lid, anything bigger than say 4'x4'. The tip of the mic should be as close to the surface as you can get, with as much of the flat surface out in front of the mic as possible. I doubt those dips are in the processor. For them to have been in measurements throughout the venue leads me to believe they are part of your measurement technique. Do not use a mic on a stand for venue magnitude response measurements.
Mac
-
Take your measurements again in a ground plane manner as Mac suggested. You can put a few coils in the cable and place the barrel of the mic on it, with the tip of the capsule touching the ground at an angle. This will eliminate any floor bounce reflections.
Doug....
I think this way of stating it may be confusing the OP. I would suggest that it does not "eliminate" any reflections, rather it makes any reflections so close to being direct sound as to render them negligible. Semantics, perhaps, but it could be confusing.
In that line, could one employ a PCC mic for measurement purposes?
-
Venue's seated, so might cause more issues trying to avoid reflections. Not too worried about this venue, it'll be quite a while till I'm there again. Sorry I wasn't meaning x-over points. Without being able to see the actual processor, I was assuming they might have some "interesting" EQ in place.
...
It is easy to discern a dip caused by EQ from a dip caused by cancellation:
A dip by EQ can be filled with EQ, so if you suspect a voicing problem, bring up 1K by about 6dB, if it comes back roughly 6dB, its EQ,
if it comes back only 1dB or less its cancellation
and for the record:
a graph without scaling is worthless
Uwe
Sorry, if this sounds rude, its to late, I should be sleeping
-
I think you're not hearing what we're saying. Having your mic not be in a ground plane causes reflections off the floor to interfere with the measurement. Do a ground plane measurement. Use a sheet of plywood over the seats, or a console case lid, anything bigger than say 4'x4'. The tip of the mic should be as close to the surface as you can get, with as much of the flat surface out in front of the mic as possible. I doubt those dips are in the processor. For them to have been in measurements throughout the venue leads me to believe they are part of your measurement technique. Do not use a mic on a stand for venue magnitude response measurements.
Mac
+1. The screen shot has the tell-tale signs of an invalid measurement.
-
It is easy to discern a dip caused by EQ from a dip caused by cancellation:
A dip by EQ can be filled with EQ, so if you suspect a voicing problem, bring up 1K by about 6dB, if it comes back roughly 6dB, its EQ,
if it comes back only 1dB or less its cancellation
and for the record:
a graph without scaling is worthless
Uwe
Sorry, if this sounds rude, its to late, I should be sleeping
Boosting the eq is also my quick check.
I am not sure the ground plane is always better. If I am doing a quick pre-show check in a theater with installed seats, the ear level measurement is often cleaner. I guess I could use a box cover or something to create a ground plan on top of the seats but that is just one more thing to haul around.
-
In that line, could one employ a PCC mic for measurement purposes?
No. PCC microphones would be disqualified for measurements because (first of all) they are cardioid, which would skew phase (and frequency) reponse considerably and (secondly) they are not really as close to pressure zone/ground plane as we would like. The capsule is oriented 90-degrees to the ground plane, so there will be a small, but real, bounce. This, of course, would only skew HF's.