|
|
Post by ansat on Apr 6, 2015 9:37:05 GMT -5
How does that work? If you take a flat sub response with a roll off on the top end from a crossover, doesn't increasing the level (even in Dirac) always move the effective crossover point up? Im just trying to imagine what this would look like. If the crossover is set to say 80, the sub response should be about 3db down at 80, but if it started 6db up, doesn't that move the effective crossover point to a higher frequency? I will add it to my test list to find out for sure. Baseline eq flat, +6 sub and +6 dirac sub should be the three tests that we need to answer that. Building the dirac files now. Tony |
|
KeithL
Administrator 
Posts: 10,261
|
Post by KeithL on Apr 6, 2015 10:04:31 GMT -5
That's really a matter of semantics..... Frequency response is always expressed in terms of relative level. If I create a "12 dB/octave low pass filter at 80 Hz", then it will look something like this (with some variations depending on which filter type I use): -0.1 dB @ 20 Hz -0.2 dB @ 40 Hz -3.0 dB @ 80 Hz -15 dB @ 160 Hz -27 dB @ 320 Hz -51 dB @ 640 Hz In words, it will be "almost flat at all frequencies significantly below the crossover frequency, down 3 dB at the crossover frequency, and slope downwards at 12 dB / octave above the crossover frequency. (Remember that dB express a ratio, so all these "dB" numbers are relative to some arbitrary "0 dB". If I draw this picture on a graph with the scale in dB, changing the gain moves the entire curve up or down, but doesn't change the shape at all. (I can add or subtract the same number to all of those dB values and the curve won't change.) The answer to your question simply depends on how you define "the crossover point". If you define the crossover point as "the point where the response of the sub and main speakers are EQUAL", then moving the gain up or down will change that, because it will move the point where the two lines on the graph cross. However, if you define "the crossover point of each speaker separately as the point where its response is down 3 dB", then moving either curve up or down will change the spot where they cross, but it won't change the shape of either curve, and so the crossover point won't change. To me, if you're talking about the crossover INSIDE A SINGLE SPEAKER, then it makes sense to use the first definition, because the point where the response of the two drivers being crossed over is equal is the point where "the sound source moves from one to the other". However, when you're talking about bass management, where each speaker has its own crossover setting, which can even be set differently, then it makes the most sense to express "the crossover point of each individually". (In a normal bass management situation, you would generally WANT the crossover point of the sub to be set the same as the crossover point of the main speakers which it services, which would be "where the curves cross", but they could certainly theoretically be set differently.) In short, if you have individual control over each crossover point, then you should treat each as a separate value.... and your goal should be to make them line up. You also need to remember that the crossover point between the sub and different speakers may be different - and so the term really defines the relationship between two speakers. For example, you could set your big main speakers to have a crossover at 50 Hz, and your surrounds to have a crossover at 80 Hz. In this case, the shapes of both the high-pass and low-pass curves are different depending on whether you're talking about the mains or the surrounds. Especially when things get complicated like this, I find it much simpler to look at each curve for each speaker separately... How does that work? If you take a flat sub response with a roll off on the top end from a crossover, doesn't increasing the level (even in Dirac) always move the crossover point up? Im just trying to imagine what this would look like. If the crossover is set to say 80, the sub response should be about 3db down at 80, but if it started 6db up, doesn't that move the effective crossover point to a higher frequency? |
|
|
|
Post by nickwin on Apr 6, 2015 10:22:20 GMT -5
How does that work? If you take a flat sub response with a roll off on the top end from a crossover, doesn't increasing the level (even in Dirac) always move the effective crossover point up? Im just trying to imagine what this would look like. If the crossover is set to say 80, the sub response should be about 3db down at 80, but if it started 6db up, doesn't that move the effective crossover point to a higher frequency? I will add it to my test list to find out for sure. Baseline eq flat, +6 sub and +6 dirac sub should be the three tests that we need to answer that. Building the dirac files now. Tony Your the best Tony! Keith, I got all that, I was just asking about raising the sub level by raising the target curve vs raising the trim level post Dirac. |
|
|
|
Post by ansat on Apr 6, 2015 12:02:59 GMT -5
The results are a bit head scratching. I am waiting on a response from Dirac to ensure that I did not goof somewhere.
Tony
|
|
|
|
Post by socketman on Apr 6, 2015 12:32:55 GMT -5
Nick , what do think would be the result? In my minds eye I see it being the same difference. I am interested to see what Tony's test will show. Raising the sub level is a wholesale change which in my mind should lift all frequencies equally and after the interaction point of the sub with the main speakers.
|
|
|
|
Post by geebo on Apr 6, 2015 13:26:52 GMT -5
How does that work? If you take a flat sub response with a roll off on the top end from a crossover, doesn't increasing the level (even in Dirac) always move the effective crossover point up? Im just trying to imagine what this would look like. If the crossover is set to say 80, the sub response should be about 3db down at 80, but if it started 6db up, doesn't that move the effective crossover point to a higher frequency? If you raise the output of the sub by 6db, then the -3db point (aka crossover point) has also been raised 6db but has not moved up or down in frequency. Now where the curves intersect will be further up in frequency but that's not the crossover point by definition. |
|
|
|
Post by ansat on Apr 6, 2015 13:48:21 GMT -5
How does that work? If you take a flat sub response with a roll off on the top end from a crossover, doesn't increasing the level (even in Dirac) always move the effective crossover point up? Im just trying to imagine what this would look like. If the crossover is set to say 80, the sub response should be about 3db down at 80, but if it started 6db up, doesn't that move the effective crossover point to a higher frequency? If you raise the output of the sub by 6db, then the -3db point (aka crossover point) has also been raised 6db but has not moved up or down in frequency. Now where the curves intersect will be further up in frequency but that's not the crossover point by definition. Correct, The confusion was in my comments stating otherwise. My thought process is that emotiva does not know the level at which Dirac has its gain and the bass management would be applied at what ever Dirac shows as reference, since we are altering the reference for that speaker, I was thinking that it might not change the intersection point. But looking at the charts, I am showing a ton of oddities with the levels of the sub and I am not sure if I can trust the results yet. Tony |
|
|
|
Post by alucard on Apr 6, 2015 14:12:28 GMT -5
For what it's worth, the Harmon curve didn't sound good on my setup. Very muddled... I started by clicking Auto-Target and playing from there. I'm using Tony's subwoofer curve - I really like the punch. Can you post the curve/measurement of the sub? Looks like it starts to drop off at 150hz? |
|
|
|
Post by socketman on Apr 6, 2015 14:46:37 GMT -5
How does that work? If you take a flat sub response with a roll off on the top end from a crossover, doesn't increasing the level (even in Dirac) always move the effective crossover point up? Im just trying to imagine what this would look like. If the crossover is set to say 80, the sub response should be about 3db down at 80, but if it started 6db up, doesn't that move the effective crossover point to a higher frequency? If you raise the output of the sub by 6db, then the -3db point (aka crossover point) has also been raised 6db but has not moved up or down in frequency. Now where the curves intersect will be further up in frequency but that's not the crossover point by definition. With an active crossover you should be able to minimize this by using a steeper roll off without the inherent issues posed by a passive crossover. |
|
|
|
Post by geebo on Apr 6, 2015 15:01:26 GMT -5
If you raise the output of the sub by 6db, then the -3db point (aka crossover point) has also been raised 6db but has not moved up or down in frequency. Now where the curves intersect will be further up in frequency but that's not the crossover point by definition. With an active crossover you should be able to minimize this by using a steeper roll off without the inherent issues posed by a passive crossover. Or if the goal is to keep the same intersection point then lower the crossover point as you raise the sub's level. |
|
|
|
Post by ncred02 on Apr 6, 2015 17:08:13 GMT -5
I wish I knew what in blazes yall are taking about.
This Dirac stuff is way to confusing for us with short term memory lose ( im not joking)
It took me a long time to figure out my UMC1 and i still don't know what roll offs or bypass are. And the graphs? forget it!
But seems likes Some of you folk know ya sh....
GO DUKE!!!!
|
|
|
|
Post by wilburthegoose on Apr 6, 2015 20:03:29 GMT -5
ncred2. Completely agree. It was detailed technical discussions like this that almost kept me from buying my XMC-1. But, boy I'm glad I did, even though I don't understand all of this - YET.  |
|
|
|
Post by nickwin on Apr 7, 2015 9:36:11 GMT -5
ncred2. Completely agree. It was detailed technical discussions like this that almost kept me from buying my XMC-1. But, boy I'm glad I did, even though I don't understand all of this - YET. There's really no need to worry. Unlike most of its competition the xmc1 sounds amazing right out of the box. It more than holds it's own even just as an analog pre amp. Things get a little more complex when you bring in crossovers and eq... |
|
|
|
Post by nickwin on Apr 7, 2015 9:40:20 GMT -5
Nick , what do think would be the result? In my minds eye I see it being the same difference. I am interested to see what Tony's test will show. Raising the sub level is a wholesale change which in my mind should lift all frequencies equally and after the interaction point of the sub with the main speakers. i honestly can't even begin to guess. It sounds like it's definitely not what you would expect. Any update Tony? |
|
|
|
Post by nickwin on Apr 7, 2015 10:03:17 GMT -5
If you raise the output of the sub by 6db, then the -3db point (aka crossover point) has also been raised 6db but has not moved up or down in frequency. Now where the curves intersect will be further up in frequency but that's not the crossover point by definition. Correct, The confusion was in my comments stating otherwise. My thought process is that emotiva does not know the level at which Dirac has its gain and the bass management would be applied at what ever Dirac shows as reference, since we are altering the reference for that speaker, I was thinking that it might not change the intersection point. But looking at the charts, I am showing a ton of oddities with the levels of the sub and I am not sure if I can trust the results yet. Tony Ive been wondering about the specifics of how Dirac sets gain because depending on how it's done it could make a big difference. Specifically is the gain set before or after the target curve is set? If you move the whole sub response up to +6 do you actually get +6 or does dirac then bring the gain down 6 to compensate? If your sub target has a 6db rise from 80-20hz, where does Dirac match the gain? Does it avg the sub spl so it would be set about -3 (what it would be around 40hz)? If that's the case you would have to manually boost the gain 3db to get the sub to match up with your mains spl wise at 80hz, and to get your full 6db boost at 20hz. |
|
|
|
Post by wilburthegoose on Apr 7, 2015 19:23:27 GMT -5
Well, I've gone back to the curve that shipped with Dirac LE. I've come to realize that I don't know enough yet to fiddle with the curves. I'll wait until the repeated Dirac Full crash on open is fixed, and then I'll restart my quest  |
|
|
|
Post by socketman on Apr 7, 2015 19:50:25 GMT -5
Wilbur, so far all I have done is the measurements and in Dirac full I used auto target and closed up the curtains so it doesn't fix things that don't need to be. I also lifted the bass on the sub a few db and it sound great to my ears. Don't get too down about messing with it, we are picking nits here IMHO. There is so much to learn and many levels to room acoustics and frequency manipulation. I found I have to draw the line somewhere, its much like an airplane, I know it uses the dihedral effect and lots of horsepower and that's good enough for me, I don't need to know how to build a plane to enjoy its convenience. I found the best way to learn was by using REW and asking lots of questions at Home theater shack and reading lots and relying on guys like tony and know when its as good as I can get it. This machine sounded great without Dirac so its hard to make it sound better. Just enjoy the music I say.
Richard
|
|
|
|
Post by ansat on Apr 7, 2015 20:08:39 GMT -5
Nick , what do think would be the result? In my minds eye I see it being the same difference. I am interested to see what Tony's test will show. Raising the sub level is a wholesale change which in my mind should lift all frequencies equally and after the interaction point of the sub with the main speakers. i honestly can't even begin to guess. It sounds like it's definitely not what you would expect. Any update Tony? Not yet, the answer I am waiting on is to if the info outside of the curtains us used in determining the gain applied. I seen a pretty large difference between the gain in some of my measurements. Tony |
|
|
|
Post by Thunderduck on Apr 7, 2015 21:45:31 GMT -5
Well, I've gone back to the curve that shipped with Dirac LE. I've come to realize that I don't know enough yet to fiddle with the curves. I'll wait until the repeated Dirac Full crash on open is fixed, and then I'll restart my quest I know of what you speak. Sometimes it is hard to keep my eyes from glazing over as I read some of these posts. Not because they are not valuable, but I don't have any idea what they are talking about. I am still having trouble figuring out the graphs Dirac produces, let alone some of the other things discussed around here. I am learning, but I am so far behind many of you it is not funny.
What I am going to ask may have been done elsewhere, there are so many threads that bring curves and targets and graphs up it is hard for me to remember where I read something (old age will get ya). Here goes:
Would it be possible for someone to post a before curve/target/whatever, explain what is wrong or why they want to change it, and then post the after results explaining what it is they did? Like the "curtains", looking at some of my Dirac graphs, the frequency plot (is that correct), will start real low down, rise sharply, level off to a certain degree, go across the page with some peaks and valleys, and then usually fall off the page at the end. Or at least something like that. I have no idea where I should, if I should, move the curtain on the left side of the graph or the right. Should I move it to where the line starts to level off? Should I set this according to what my speaker frequency response is? I have no idea. What do the peaks and valleys mean? As you can tell I have absolutely no idea what the heck I am doing and I hope you all can understand what I am talking about. When you don't know much about the topic, it is kind of hard to ask intelligent questions about it.
I really appreciate any help that can be provided in this regard. If it has been done elsewhere, please point me in the proper direction. I sometimes feel like I have been a drain on the community here, asking more than I am able to give. I try to answer those questions where I have a reasonable understanding of the answer, but sometimes feel like I am taking more than giving.
On a more cheerful note, I will be getting my replacement XMC tomorrow (the blue light of death took my current unit). I am going to offer a bit of advice to folks here. Make darn well sure that you save your configuration info saved to a thumb drive instead of just in the unit itself. And make sure you know where said thumb drive is. I think my son needed one for his computer, saw mine laying around and decided to take it. So, I will get to start from scratch on setting up my new XMC. One thing positive, I will certainly be able to find out if there are any problems with the new one.
So in closing this Great American Novel, any help with showing and describing why and how a target curve is set up would be very much appreciated by myself and I believe others here.
Thank you very much,
Steve
|
|
|
|
Post by ansat on Apr 7, 2015 22:29:24 GMT -5
Well, I've gone back to the curve that shipped with Dirac LE. I've come to realize that I don't know enough yet to fiddle with the curves. I'll wait until the repeated Dirac Full crash on open is fixed, and then I'll restart my quest I know of what you speak. Sometimes it is hard to keep my eyes from glazing over as I read some of these posts. Not because they are not valuable, but I don't have any idea what they are talking about. I am still having trouble figuring out the graphs Dirac produces, let alone some of the other things discussed around here. I am learning, but I am so far behind many of you it is not funny.
What I am going to ask may have been done elsewhere, there are so many threads that bring curves and targets and graphs up it is hard for me to remember where I read something (old age will get ya). Here goes:
Would it be possible for someone to post a before curve/target/whatever, explain what is wrong or why they want to change it, and then post the after results explaining what it is they did? Like the "curtains", looking at some of my Dirac graphs, the frequency plot (is that correct), will start real low down, rise sharply, level off to a certain degree, go across the page with some peaks and valleys, and then usually fall off the page at the end. Or at least something like that. I have no idea where I should, if I should, move the curtain on the left side of the graph or the right. Should I move it to where the line starts to level off? Should I set this according to what my speaker frequency response is? I have no idea. What do the peaks and valleys mean? As you can tell I have absolutely no idea what the heck I am doing and I hope you all can understand what I am talking about. When you don't know much about the topic, it is kind of hard to ask intelligent questions about it.
I really appreciate any help that can be provided in this regard. If it has been done elsewhere, please point me in the proper direction. I sometimes feel like I have been a drain on the community here, asking more than I am able to give. I try to answer those questions where I have a reasonable understanding of the answer, but sometimes feel like I am taking more than giving.
On a more cheerful note, I will be getting my replacement XMC tomorrow (the blue light of death took my current unit). I am going to offer a bit of advice to folks here. Make darn well sure that you save your configuration info saved to a thumb drive instead of just in the unit itself. And make sure you know where said thumb drive is. I think my son needed one for his computer, saw mine laying around and decided to take it. So, I will get to start from scratch on setting up my new XMC. One thing positive, I will certainly be able to find out if there are any problems with the new one.
So in closing this Great American Novel, any help with showing and describing why and how a target curve is set up would be very much appreciated by myself and I believe others here.
Thank you very much,
Steve
Hey Steve, if someone else cannot post anything up for ya, I can try to do somrthing this weekend. Tony |
|
