Note that there are many implementations of all-pass filters (and phase shifters, and time delays...)
Addressing a Null
- Thread starter gshelley
- Start date
In fact, even with the most simple second order allpass filter of form
(cz^2+bz+a) / (a+bz+cz^2) there are an infinite number of selections, even for when the roots for the denominator polynomial are inside the unit circle.
There is something else one needs to realize. If phase = 2* pi*f*t*constant, then what you have is a pure time delay of "constant" in seconds. This is sometimes misleadingly called "linear phase" when what it means is constant delay. Yep. Really.
Hi Everyone. As others have mentioned, Room Nulls are tough to deal with. For two-channel, I don't care for external subs, and I'm not too fond of sitting off-center of the room (side-to-side). So I elected to address the problem with treatment. Below is a picture of the treatments on the front, left half of the room (to give you an idea of placement). The graph shows the difference between NO bass treatment, and 5 RPG Modex Plates, plus 12 RealTraps. The panels flat on the ceiling between the speakers are 2 RPG BAD Panels. The point of first reflection are 2 RealTraps 'HF' Mondo Traps.
You can see that although the nulls are not eliminated, they have been reduced, along with the opposing peaks being smoothed out some. Audibly, it is a huge difference.
Green - No Bass Treatment; Teal - Modex Plates and RealTraps
You can see that although the nulls are not eliminated, they have been reduced, along with the opposing peaks being smoothed out some. Audibly, it is a huge difference.
Green - No Bass Treatment; Teal - Modex Plates and RealTraps
Also, here are the waterfalls from the two measurements:
Levels are the same in both measurements, other than the amplifier in the speakers had to be adjusted after placing all of the traps. So I don't have an explanation for the overall level difference (other than I don't really know what I'm doing when it comes to REW
).
Levels are the same in both measurements, other than the amplifier in the speakers had to be adjusted after placing all of the traps. So I don't have an explanation for the overall level difference (other than I don't really know what I'm doing when it comes to REW
).Attachments
-
modex_002.jpg -
nomodex_001.jpg
Last edited:
Bill,
Can you post the first graph with a scale decreasing the scale value to 5dB division? Please just change the limits in REW to 40 and 100 dB.
Can you post the first graph with a scale decreasing the scale value to 5dB division? Please just change the limits in REW to 40 and 100 dB.
It is not a question of being correct, but easier to anlyse. But I was addressing the frequency response graphs, not the waterfall one's - apologies if I was not clear.
But it seems that others settings or things have changed besides the room treatment between the measurements. This will make difficult to compare the before and after.
Yeah, I'm not sure why there is such a difference. REW doesn't seem to have a way to overlap the two graphs that I can see, post measurement.
But here is the adjusted graph:
Hi Bill, I am not experienced, like the experts here on WBF, in reading these graphs, though it seems to me that you have only two peaks in the low freq. One peak at 20-30hz and another at 40-50 or to put it another way, a 10db null at 35hz. The graphs look quite similar to other posted graphs here on WBF.
My questions are;
1) How does your system sound? Do you hear problems in the low, mid or high freq?
2)What do the graphs indicate to you and are you able to correlate them to the sound of your system?
3)Are you worried about what the graphs tell you, if your system sounds good to you?
Wendell
My questions are;
1) How does your system sound? Do you hear problems in the low, mid or high freq?
2)What do the graphs indicate to you and are you able to correlate them to the sound of your system?
3)Are you worried about what the graphs tell you, if your system sounds good to you?
Wendell
You can use the add offset to data in the Controls Window to move one of the graphs up or down. But IMHO the differences shown can not be due mainly to room treatment - my first guess would be that the system in green would sound better than the one in blue!
Unhappily the WBF experts are not chiming. We need them!
The only other thing that has change is the speaker cables. But there is a whole butt-load of treatments that have been added between the two graphs. One thing; all though the bass is graduated higher in the blue, I actually had to turn the bass down to this level, from the other, after adding the treatments. And, I get what you're saying about the green looking like it would sound better, but that is why I also posted the waterfalls... thinking there is more going on since adding the treatments, than just the response. Believe me, after all the treatments, I am a little baffled at what the measurements show before and after as well. Obviously I can't take the treatments out and re-measure (that would take a couple of days, and a whole hell of a lot of work!)
I believe it is known as a 'House Curve'. And yes, listening now, I prefer it to the way the system previously sounded. It may be interesting to note, that I had to indeed turn the volume DOWN on the woofer amps after installing the Modex Plates and RealTraps, and yet the low end is still quite a bit higher in amplitude than it was prior.
Yes, and it's a common misconception that adding bass traps reduces the level of bass frequencies. Peaks are typically 6 dB or less, but nulls can be 30 or even 40 dB deep. So adding bass traps raises the nulls more than it reduces the peaks, with the net result being more bass. At least in most rooms. In a square or cube room the peaks might dominate, so adding traps would lower the overall level more. The good news is bass traps do whatever is needed! That is, they lower peaks and raise nulls. (Bill, I know you understand this!)
--Ethan
That I do... understanding REW... not so much.
(but I'm working on it)Frankly, I did not use REW when determining where to place the Bass Traps. I used 3 different bass test tones and a Sound Level Meter to determine which traps and where.
Also, I believe the graphs show clearly that several of the nulls have been raised, post bass trapping. The stubborn null seems to be the one around 120hz.
Here’s another example along the same lines that someone posted at HT Shack last month. Three measurements – left and right corners, and center-of-wall. I don’t think I have to point out which trace is the center-of-wall, do I?
Regards,
Wayne A. Pflughaupt
Similar threads
 | Steve Williams Site Founder | Site Owner | Administrator |  | Ron Resnick Site Co-Owner | Administrator |  | Julian (The Fixer) Website Build | Marketing Managersing |