Lsc
Emo VIPs 
Posts: 3,435
|
Post by Lsc on Jan 25, 2015 9:16:02 GMT -5
I also think it has been very helpful in many ways. However, I'm in the camp that there is nothing wrong with Dirac LE.
The only think I see wrong is that the LE version doesn't provide the ability to adjust the curve...it's silly IMHO bc it's basic functionality is a standard feature for the PC version and would have solved my buddy's and the good folks issues on this forum.
I respect Emotiva's decision to sell us this "full" version but I don't agree with it. I feel the full version should have been the standard. It's no skin off my back bc my room obviously works great w the preset dirac curve but my buddy isn't as lucky.
|
|
|
|
Post by hesster on Jan 25, 2015 13:59:55 GMT -5
Does it matter if "Enhanced Bass" is selected before running Dirac? If so, on or off? Not sure why may most recent SW plot shows a curtain extending to 700Hz, and not sure it matters, but there is lots of variation after 100hz - does this matter? Prior measurements placed the curtain @ 100 hz. 
|
|
|
|
Post by socketman on Jan 25, 2015 14:29:10 GMT -5
Does it matter if "Enhanced Bass" is selected before running Dirac? If so, on or off? Not sure why may most recent SW plot shows a curtain extending to 700Hz, and not sure it matters, but there is lots of variation after 100hz - does this matter? Prior measurements placed the curtain @ 100 hz.
I believe Dirac tests all speakers full range and that is probably where a sub just cant reproduce anymore output due to natural roll off of the driver. After the Dirac correction is loaded the bass management will take over and cut off anything above your designated crossover frequency. |
|
|
|
Post by ansat on Jan 25, 2015 18:36:58 GMT -5
Does it matter if "Enhanced Bass" is selected before running Dirac? If so, on or off? Not sure why may most recent SW plot shows a curtain extending to 700Hz, and not sure it matters, but there is lots of variation after 100hz - does this matter? Prior measurements placed the curtain @ 100 hz.
I believe Dirac tests all speakers full range and that is probably where a sub just cant reproduce anymore output due to natural roll off of the driver. After the Dirac correction is loaded the bass management will take over and cut off anything above your designated crossover frequency. I think you hit part of your noise floor. Decrease the mic gain a little in dirac and set your levels normally. You want the test tones a little louder. This should fix it. Tony |
|
|
|
Post by markc on Jan 26, 2015 1:57:55 GMT -5
I think this graph actually shows why the LE version of Dirac is probably missing a lot more than target curve adjustability.
Defining the boundaries of the correction range is going to be much more useful (The pale blue central section of the graph). The LE version has Dirac selecting the limits of this range itself, based on what useable frequencies it thinks are possible, and up to 750Hz for a sub, as in this case, is silly.
For this Subwoofer of hesster's, Dirac is using a lot of filters to "optimise" the sub outside of it's useable range with a LOT of boosts over 150Hz which are simply not needed and won't be required after the bass management is applied.
This will result in a decrease to the overall volume of all channels in all likelihood, due to the positive filters being used to list to boost the sub channel between 70 and 750Hx. These filters will be in use even after bass management.
This sub should only have consideration of optimisation between 23 and 140 Hz (I think LFE .1 content is up to 120Hz, so that allows a margin) and not 23 - 750Hz.
The upgraded full Dirac will allow us to bring in the correction upper and lower limits I believe.
Similarly, for the graphs of peoples surrounds, a lot of graphs are having Dirac apply lots of boosts to the 30-60Hz range, simply because the speakers are not capable of reproducing those frequencies and are most likely run with an 80Hz cross over or thereabouts so correction from 60Hz should be enough.
|
|
|
|
Post by Mike Ronesia on Jan 26, 2015 2:27:28 GMT -5
Because the XMC sets the XO does it really matter what Dirac see's and adjusts outside of those frequencies? It can correct my LFE all the way up to 20khz, but my subs won't pass anything higher the 100hz or whatever I set it to, so why should I care if Dirac looked at it? I am asking because I want to know if there are adverse side effects from Dirac making adjustments outside of the speakers XO parameters. I can't think of any, but I'm new to this stuff.
Cheers
Mark
|
|
|
|
Post by hesster on Jan 26, 2015 9:05:18 GMT -5
I greatly appreciate all of the input posted here. This time I reran Dirac following Ansat's advice on Output and Gain adjustments, did a precise job in Mic placement per the guidelines posted, and the Graphs never looked better with good curtain placement and almost ruler flat filter applications on all of the speakers. The Sub Plot below is much better, and I agree when Full is released it will allow use precise curtain placement and additional tweaks. It pays to follow all of the recommendations on this forum to get the most out of Dirac!
|
|
|
|
Post by barrak on Jan 26, 2015 11:28:34 GMT -5
Because the XMC sets the XO does it really matter what Dirac see's and adjusts outside of those frequencies? It can correct my LFE all the way up to 20khz, but my subs won't pass anything higher the 100hz or whatever I set it to, so why should I care if Dirac looked at it? I am asking because I want to know if there are adverse side effects from Dirac making adjustments outside of the speakers XO parameters. I can't think of any, but I'm new to this stuff. Cheers Mark It might matter if the to be discarded range by bass management has substantially higher measured level than the range of interest (especially for mains). Suppose for example that you have full range mains that you'd like to crossover at 80Hz to a sub. Both speaker and sub are measuring a huge room mode at 50 Hz, say +10dB over the 80+ Hz range for the mains. In this case, Dirac's output level calibration will have you level match speakers and subs in reference to the 50Hz peak, while Dirac filters will attempt to boost the 80+ Hz range by the 10dB deficit. End result: loss of overall digital Dynamic range and risk of digital clipping at high volumes. Short of giving us the option to high-pass the test tones, I doubt the full Dirac would remedy the above situation. A very risky workaround to your mains is to deliberately push their level in the calibration process to counter the anticipated said deficit. I wouldn't attempt that even with my very capable Catalyst 12C... digital clipping triggered in the deep bass can still result in full range damaging energy. I also doubt that Dirac software would allow such unbalanced measurements. On another point, full range processing for all channels can lend itself to applying 2.2 mode to bi-amped full range speakers. Two caveats, however; a) speaker crossovers should be defeatable or you'll get a deep null at the woofer/midrange xover point, and b) XMC-1 maximum xover point is 200 Hz, which is too low for most full range speaker designs. |
|
|
|
Post by markc on Jan 26, 2015 14:01:34 GMT -5
Because the XMC sets the XO does it really matter what Dirac see's and adjusts outside of those frequencies? It can correct my LFE all the way up to 20khz, but my subs won't pass anything higher the 100hz or whatever I set it to, so why should I care if Dirac looked at it? I am asking because I want to know if there are adverse side effects from Dirac making adjustments outside of the speakers XO parameters. I can't think of any, but I'm new to this stuff. Cheers Mark It does matter because Dirac is applied to the signal before bass management and pays no regard to what will later be filtered out. It will apply all the Dirac filters it calculated and then make the necessary adjustments to volume to prevent clipping BEFORE bass management as the next step.
Every signal boost within Dirac, like those seen on the subwoofer channel graph I commented on, results in the overall volume having to be attenuated to prevent clipping. Users have noticed this volume attenuation by Dirac/XMC. This includes all boosts including the desired ones as well as the unnecessary boosts for the frequencies which will be above or below a crossover and so don't need to be filtered.
This doesn't affect the noise shaping that you get within your non crossed over range but WILL affect the overall all-channels volume resulting in a lower overall volume which you then compensate for by turning up system volume as well as possibly the volume on the subwoofer.
Two other and probably unimportant thoughts are: I don't know how many filters Dirac has access to for each channel. If it is like the two Parametric EQ banks then there is a limited number. Why use them many of them making corrections to audio which will be discarded by a crossover limiting the ones available for the parts you do want?
From Flavio (Dirac) and Keith's posts, Dirac also make "decisions" and "judgements" on what it thinks represents room variations and positional variations (which is why it used the 8 other microphone positions to compare to the all important first position to "decide" what is a true variance) so it makes sense to get it to discard the unimportant information and focus on the bit you need.
|
|
|
|
Post by Gary Cook on Jan 26, 2015 14:16:20 GMT -5
Did we miss the bit where hesster made a set up mistake, had the mic gain too high and the volume too low? Hence the ambient noise was heard by DIRAC as being from the speaker involved. It wouldn't matter if the window was wide or narrow, picking up ambient noise is going to skew the initial measurements and hence compromise the results. Once the set up mistake was corrected (following Tony's advice) the measurements, including the window, were entirely appropriate. Manual adjustments are not going to correct a fundamental set up mistake. In fact manually changing the window could have resulted in a sub optimal result. For me the lesson learned was to take the time and make the effort to get the set up right.
Cheers
Gary
|
|
|
|
Post by geebo on Jan 26, 2015 14:58:26 GMT -5
I think this graph actually shows why the LE version of Dirac is probably missing a lot more than target curve adjustability.Defining the boundaries of the correction range is going to be much more useful (The pale blue central section of the graph). The LE version has Dirac selecting the limits of this range itself, based on what useable frequencies it thinks are possible, and up to 750Hz for a sub, as in this case, is silly. For this Subwoofer of hesster's, Dirac is using a lot of filters to "optimise" the sub outside of it's useable range with a LOT of boosts over 150Hz which are simply not needed and won't be required after the bass management is applied. This will result in a decrease to the overall volume of all channels in all likelihood, due to the positive filters being used to list to boost the sub channel between 70 and 750Hx. These filters will be in use even after bass management. This sub should only have consideration of optimisation between 23 and 140 Hz (I think LFE .1 content is up to 120Hz, so that allows a margin) and not 23 - 750Hz. The upgraded full Dirac will allow us to bring in the correction upper and lower limits I believe. Similarly, for the graphs of peoples surrounds, a lot of graphs are having Dirac apply lots of boosts to the 30-60Hz range, simply because the speakers are not capable of reproducing those frequencies and are most likely run with an 80Hz cross over or thereabouts so correction from 60Hz should be enough. Adjusting the lower and upper limits of the target curve IS part of adjusting the target curve. |
|
|
|
Post by ansat on Jan 26, 2015 18:12:31 GMT -5
Did we miss the bit where hesster made a set up mistake, had the mic gain too high and the volume too low? Hence the ambient noise was heard by DIRAC as being from the speaker involved. It wouldn't matter if the window was wide or narrow, picking up ambient noise is going to skew the initial measurements and hence compromise the results. Once the set up mistake was corrected (following Tony's advice) the measurements, including the window, were entirely appropriate. Manual adjustments are not going to correct a fundamental set up mistake. In fact manually changing the window could have resulted in a sub optimal result. For me the lesson learned was to take the time and make the effort to get the set up right. Cheers Gary Gary, It would depend on how quiet your room is. For my room, I need to set the mic gain below the scale, and then set the output gain as normal. The quieter the room, the lower the SPL for the test sweeps, the lower the spl for the test sweeps, the more the noise floor can impact the results. Lately I have been setting the output volume of the test sweeps to 75db on the RS meter then setting the mic gain until the output volume is in the green. Tony |
|
|
|
Post by thompson12 on Jan 26, 2015 21:20:35 GMT -5
I got a new EMM mic everything looks and sounds much better Thanks Old mic  New mic  |
|
|
|
Post by socketman on Jan 26, 2015 22:28:40 GMT -5
Wow that must have sounded horrible. Glad its all sorted out for you now. How did you figure out you needed the new mic? |
|
|
|
Post by thompson12 on Jan 26, 2015 23:26:35 GMT -5
Wow that must have sounded horrible. Glad its all sorted out for you now. How did you figure out you needed the new mic? For the sound with the old mic it was shrill and lacking in bass but the separation and sound stage was much better then without Dirac. Answer to your question Tony Figured it out he helped a lot, he got me up and running in REW. I bought a mic and the Behringer uca222 back when Andrew did that sub demo I downloaded REW and never figured it out. A few steps and tips from Tony and I was up and running in REW and comparing my UMM-6 and the EMM mics changing and loading cal files was a breeze, Now all I have to do is figure out how to read/interpret the graphs |
|
|
|
Post by socketman on Jan 26, 2015 23:36:16 GMT -5
Yes Tony certainly knows his stuff and has been a great asset . More info coming later this week .
|
|
|
|
Post by Gary Cook on Jan 27, 2015 1:20:48 GMT -5
Did we miss the bit where hesster made a set up mistake, had the mic gain too high and the volume too low? Hence the ambient noise was heard by DIRAC as being from the speaker involved. It wouldn't matter if the window was wide or narrow, picking up ambient noise is going to skew the initial measurements and hence compromise the results. Once the set up mistake was corrected (following Tony's advice) the measurements, including the window, were entirely appropriate. Manual adjustments are not going to correct a fundamental set up mistake. In fact manually changing the window could have resulted in a sub optimal result. For me the lesson learned was to take the time and make the effort to get the set up right. Gary, It would depend on how quiet your room is. For my room, I need to set the mic gain below the scale, and then set the output gain as normal. The quieter the room, the lower the SPL for the test sweeps, the lower the spl for the test sweeps, the more the noise floor can impact the results. Lately I have been setting the output volume of the test sweeps to 75db on the RS meter then setting the mic gain until the output volume is in the green. Tony Completely agree Tony, no XMC-1 so I haven't played with DIRAC as yet, but I've found that test sweeps at normal listening levels work best for me. Too low and the ambient noise can be an issue whereas too loud and clipping becomes a problem. Cheers Gary |
|
|
|
Post by hesster on Jan 27, 2015 7:29:00 GMT -5
Gary Cook males a good point, but let me clarify the sensitivity of what I experienced as a "set-up mistake" in my previous runs that led to the suboptimal Woofer Plot. Dirac (in the manual and video) calls for setting the Mic gain just below 24db, and test gain signals right around an average of -12db, not to exceed the "green" area. Seems straightforward, especially if one can get all of the speakers with different efficiencies to fall around -12db - one would assume this provides a little flexibility in set-up. On my previous runs, I did exactly that - JUST below -24db, and an "average" of -12db on my test signals. My reason is that my Sides are much louder during the tests and had clipped at higher test gain signals during certain Mic positions in previous tests (when Mic is closer to the Left or Right Side Speaker). SO - I set the gain lower at around ~20db, and increased the test signals not to exceed at the higher end of the green bars, and so they did not fall below -12db on the average, with the Side Surrounds being the limiting factor not to exceed the green bars. Crossed my fingers and hoped I did not get any test clips, and came close, but everything passed. In addition, the Mic positions on previous tests were very close to the recommendations, but not "precise", so I took pains to be exact. Obviously, this last setup was successful and improved, but I wanted to share my experience and just how these test adjustments led to the results. Point is - I had followed the Dirac video guidelines, but implemented Tony's recommendations to reduce Mic gain and increase test signal gain which led to these very important differences not only in the Sub, but in the other speaker Plots also. Listening sessions yesterday confirmed the improvement. |
|
Lsc
Emo VIPs
Posts: 3,435
|
Post by Lsc on Jan 27, 2015 12:21:53 GMT -5
I agree that it's not a setup mistake and concur that it's in the manual.
This may be why my buddy's calibration was so low in volume as well. The ambient noise in my house is just higher bc I live by a busy street and my measurements were with a higher mic volume.
Darn it. I got my replacement F208 midrange and need to rerun Dirac. Gotta cross my fingers in hopes that it'll be as good as my last go around.
|
|
|
|
Post by Priapulus on Jan 27, 2015 13:28:01 GMT -5
Well, I just did another Dirac run, and it's the best one yet. Music now sound much better, than without Dirac. It doesn't seem as bright as previous runs. The deep bass still seems weak, but I'm going to work on evening out the room response with treatments first.
I've discovered it's important to keep checking the AGC; something keeps turning it back on. And Tony's recommendation to reduce Mic gain, and increase test signal gain, also helped.
Just think, most owners of the XMC probably aren't on this list and don't get all this great advice about Dirac. A bunch of people, not quite pleased with Dirac, and don't know why. I certainly would have abandoned it, with my initial results. It was all the advice here that helped me get it better.
Sincerely,
/b
|
|
