Home Theater Forum and Systems banner

1 - 12 of 12 Posts

·
Registered
Joined
·
70 Posts
Discussion Starter #1 (Edited)
Hi,

Unsure where to post these, or check if they have already been posted anyway here are some bugs I have noticed.

Preferences
I at least cannot find a way of removing the smoothing I had set in the preferences. Changing the smoothing or using Graph->Remove Smoothing still shows the graphs (and new measurements) with the preference I had when REW started.

EQ Page.
Changing the min t60 time in the resonances page invalidates the results shown - some (very long t60 times) resonances (with very long t60 times) simply disappear. You need to recalculate to get them back.

EQ Page.
I am a bit suspect of the rt60 times calculated. If you look at the attached pic you will see EQ claiming a peak db level of 78.9 dB for 27.5Hz. This does not match what you can see in the frequency response. (1/48 octave smoothing is less than the frequency resolution so that should not be relevant in my opinion). The same applies to all the peak values - they are simply too low.

Markus
 

Attachments

·
REW Author
Joined
·
6,711 Posts
I at least cannot find a way of removing the smoothing I had set in the preferences. Changing the smoothing or using Graph->Remove Smoothing still shows the graphs (and new measurements) with the preference I had when REW started.
If the Analysis preference "Allow 96 PPO log spacing" is selected REW will convert measurements to log spaced if this would reduce memory use (usually the case for full range measurements). That conversion requires the data to be smoothed to 1/48th octave, which then becomes the minimum smoothing possible for the measurement. Untick the preference if you wish to see unsmoothed full range measurements.

EQ Page.
Changing the min t60 time in the resonances page invalidates the results shown - some (very long t60 times) resonances (with very long t60 times) simply disappear. You need to recalculate to get them back.
This is actually a bug with the min SPL filter. I fixed it in my dev build a couple of weeks ago, but in the meantime put a lower value in the min SPL field.

EQ Page.
I am a bit suspect of the rt60 times calculated. If you look at the attached pic you will see EQ claiming a peak db level of 78.9 dB for 27.5Hz. This does not match what you can see in the frequency response. (1/48 octave smoothing is less than the frequency resolution so that should not be relevant in my opinion). The same applies to all the peak values - they are simply too low.
Those are the levels of the particular resonances listed, not the overall response at that frequency. The level of the frequency response at any particular frequency is the combined contributions of all the components that make up the response, individual resonances can have levels that are lower or higher than the measured level depending on their relative phases.
 

·
REW Author
Joined
·
6,711 Posts
The decay view doesn't generate with 60ms slice window 600ms
You may find that after switching between two measurements that have decay data, the window field can no longer be adjusted and further decay plots cannot be generated. A workaround is to click the "Apply default settings" button, then re-enter the values you wish to use. The underlying bug has been fixed for the next release.
 

·
Registered
Joined
·
70 Posts
Discussion Starter #5
...The level of the frequency response at any particular frequency is the combined contributions of all the components that make up the response, individual resonances can have levels that are lower or higher than the measured level depending on their relative phases.
Hmm... Ok, I am certain I don't understand. Would you care to explain - possibly in a different thread so as not to confuse the issue? I would have thought that the point of a mode/resonant frequency was that every thing was in phase. So I do not understand what can contribute to the spl other than that mode.

Markus
 

·
Registered
Joined
·
70 Posts
Discussion Starter #6
If the Analysis preference "Allow 96 PPO log spacing" is selected REW will convert measurements to log spaced if this would reduce memory use (usually the case for full range measurements).
Thanks John, this was indeed the case.

Markus
 

·
REW Author
Joined
·
6,711 Posts
Hmm... Ok, I am certain I don't understand. Would you care to explain - possibly in a different thread so as not to confuse the issue? I would have thought that the point of a mode/resonant frequency was that every thing was in phase. So I do not understand what can contribute to the spl other than that mode.
A picture may make things clearer. In the graph below I have selected the modes with peaks above 60dB. To avoid cluttering up the graph too much REW only plots the individual mode responses over a small span either side of their centre frequency, but they each make a contribution right across the band, although that contribution becomes very small far from their resonant frequencies. The measured response is the sum of the direct sound and the contributions of all the modal resonances. At any particular frequency the contribution of each mode has a magnitude and a phase, different modes will have different phases so the sum of their contributions may have a level that is smaller or greater than their individual levels at that frequency.

modes.png

I did find a mistake in the peak levels shown in the resonances table however, they do not include the mic/meter and soundcard calibration corrections, I've fixed that for the next release.
 

·
Registered
Joined
·
70 Posts
Discussion Starter #11
Hi John,

Thanks for the explanation, only it does not quite remove the confusion. If I understand you correctly would that mean that if I were to play a sinus 31Hz at whatever level was used to make that measurement, then I would see the same spl, say 85dB, and that this 31 Hz would excite the 27Hz mode which would ring at 31Hz?

Markus
 

·
REW Author
Joined
·
6,711 Posts
If I understand you correctly would that mean that if I were to play a sinus 31Hz at whatever level was used to make that measurement, then I would see the same spl, say 85dB, and that this 31 Hz would excite the 27Hz mode which would ring at 31Hz?
Pretty much. It is not right to say the mode would 'ring', but a mode at 31Hz has an effect on signals at 30Hz or 29Hz etc. The variation of that effect with frequency is shown by the blue traces in the graph I posted. The various modal resonances of a space act like a set of parallel resonant filters, the final output at the listening position is the summation of the direct signal from the source and the outputs of all those filters.
 
1 - 12 of 12 Posts
Top