|
|
Post by socketman on May 7, 2015 16:13:53 GMT -5
This issue just got pointed out to me via customer service. And it definitely had me in a panic for a while. I don't have the time to review all of the posts. I am focusing entirely on Tony's graphs top of page 2. They seem to point to a serious problem. The only trouble is, I can't duplicate the results. My measurement is attached. I tried to set this up so it would look as close as possible to Tony's plots. As you can plainly see, there is no "crossover crosstalk" in this plot. Either there is some pilot error on Tony's part or there is some subtle difference in setup that causes a not so subtle difference in result. I am using a freshly factory reset unit with the following changes in Dirac preset all speakers 120Hz crossover 12dB/octave sub crossover. These changes were made so that the subwoofer response from the mains would be exactly the same as the hard-wired LFE response. This has no effect on Dirac filters. Note that +10dB LFE is required by Dolby and dts - completely normal. I ran the same plots but with LFE 10dB lower and the subwoofer curves for LFE and Left Front (and all other speakers) overlapped exactly. Tony - if you wouldn't mind - could you repeat your measurements with my settings and see if you still see this issue? If not, start restoring your settings one at a time and see when the problem returns. My source is 7.1 PCM over HDMI. Measured at XMC-1 balanced outputs. What was yours? Thanks -ray Please tell me if I am reading this graph correctly Purple = full range main speaker ? Blue = LFE Red = LFE down 10db Grey = rolled off bass 12db/oct Is one of the traces either red or blue redirected bass? |
|
|
|
Post by raydyo on May 7, 2015 17:14:46 GMT -5
I was trying to stick with Tony's color scheme
following taken with 120Hz crossovers, 12dB slope
Blue is the LFE
Red and greenish-grayish is Front left
Following taken after changing fronts to large
Purple is full range fronts
Note that red and blue are identical other than expected 10dB offset - no evidence of "crossover crosstalk"
|
|
|
|
Post by socketman on May 7, 2015 21:17:12 GMT -5
I was trying to stick with Tony's color scheme following taken with 120Hz crossovers, 12dB slope Blue is the LFE Red and greenish-grayish is Front left Following taken after changing fronts to large Purple is full range fronts Note that red and blue are identical other than expected 10dB offset - no evidence of "crossover crosstalk" This is interesting , it would seem everything is working as one would expect and according to design. I wonder why Tony was able to duplicate Nicks reported issue. I was going to attempt this myself with REW but alas my memory let me forget to do it. Maybe this weekend if the weather is bad . Thanks for taking the time to drop into this thread, I know you are super busy and it is appreciated. Richard |
|
|
|
Post by millst on May 8, 2015 0:18:29 GMT -5
I ran a few tests of my own and things look more like Ray's charts. I left my crossovers at 80Hz and used the unbalanced outputs since my balanced connections are in use. The green/blue graph is of the subwoofer output. The green is the LFE channel input. The blue is front channel input. The green/purple graph is of the front output. The green is with the fronts set to large. The purple is with the fronts set to small. The Dirac filters seem to be applied per channel. I don't see any bleed through between the two charts. Attachments:
|
|
|
|
Post by ansat on May 8, 2015 6:53:52 GMT -5
This issue just got pointed out to me via customer service. And it definitely had me in a panic for a while. I don't have the time to review all of the posts. I am focusing entirely on Tony's graphs top of page 2. They seem to point to a serious problem. The only trouble is, I can't duplicate the results. My measurement is attached. I tried to set this up so it would look as close as possible to Tony's plots. As you can plainly see, there is no "crossover crosstalk" in this plot. Either there is some pilot error on Tony's part or there is some subtle difference in setup that causes a not so subtle difference in result. I am using a freshly factory reset unit with the following changes in Dirac preset all speakers 120Hz crossover 12dB/octave sub crossover. These changes were made so that the subwoofer response from the mains would be exactly the same as the hard-wired LFE response. This has no effect on Dirac filters. Note that +10dB LFE is required by Dolby and dts - completely normal. I ran the same plots but with LFE 10dB lower and the subwoofer curves for LFE and Left Front (and all other speakers) overlapped exactly. Tony - if you wouldn't mind - could you repeat your measurements with my settings and see if you still see this issue? If not, start restoring your settings one at a time and see when the problem returns. My source is 7.1 PCM over HDMI. Measured at XMC-1 balanced outputs. What was yours? Thanks -ray Ray I will re run the measurements this evening. The previous results we're measured using hdmi to send the test tones. Tones generated by rew. I use direct mode as a control to see a straight line output to ensure there is nothing being altered by the source. I switched to stereo mode to measure the post dirac results since rew can only control 1 channel at a time. All measurements taken at the balanced outputs. Tony |
|
|
|
Post by nickwin on May 8, 2015 8:36:49 GMT -5
I ran a few tests of my own and things look more like Ray's charts. I left my crossovers at 80Hz and used the unbalanced outputs since my balanced connections are in use. The green/blue graph is of the subwoofer output. The green is the LFE channel input. The blue is front channel input. The green/purple graph is of the front output. The green is with the fronts set to large. The purple is with the fronts set to small. The Dirac filters seem to be applied per channel. I don't see any bleed through between the two charts. How do you have REW connected to the XMC? HDMI? I'm just trying to think what could have made ansats and mine meaure differently. I always did a factory reset, set the speaker size and crossover and that's it, so in my case there weren't any extra settings being applied to potentially cause problems. |
|
|
|
Post by raydyo on May 8, 2015 8:45:00 GMT -5
Thanks Tony,
That's at least two difference between your testing and mine
REW (2.0) for test tones instead of Audio precision (7.1)
Stereo mode
Maybe one of those the the key to the mystery
I'm in the middle of something else right now, but I'll try to get back to this later today.
Thanks for the info
-ray
|
|
|
|
Post by raydyo on May 8, 2015 9:04:59 GMT -5
Thanks millst
XLR vs RCA - doesn't matter. No way they can be different except hardware failure.
The slight difference you see abive 80Hz are due to 80Hz crossover on Front vs fixed 120Hz on the LFE.
I'm curious about your signal source and surround mode setting.
-ray
|
|
|
|
Post by millst on May 8, 2015 10:52:18 GMT -5
Yes, I noticed the rolling off from the crossover.
I used REW from my HTPC to generate the tones and send them PCM 7.1 over HDMI. The XMC-1 was set to surround mode. I used to J-River's DSP studio to workaround the channel steering issue that Tony mentioned.
I also created a control line, but used one of the regular speaker presets with the mains set to large.
-tm
|
|
|
|
Post by raydyo on May 8, 2015 11:08:22 GMT -5
Tony,
I repeated my tests with all combinations of 2.0 vs 7.1 and direct vs stereo and still no crosseyed crossovers.
I'm at a loss to explain your graphs. I can't spot anything you did wrong but your results are sure different than mine
Start from factory default and see if you see the problem thne start adding your settings one at a time.
This is very strange....
|
|
|
|
Post by ansat on May 8, 2015 11:10:52 GMT -5
Tony, I repeated my tests with all combinations of 2.0 vs 7.1 and direct vs stereo and still no crosseyed crossovers. I'm at a loss to explain your graphs. I can't spot anything you did wrong but your results are sure different than mine Start from factory default and see if you see the problem thne start adding your settings one at a time. This is very strange.... Will do, I will call in around 1 to 1:30 CST to go over the test before I run it. Tony |
|
|
|
Post by unsound on May 8, 2015 11:21:25 GMT -5
Tony, I repeated my tests with all combinations of 2.0 vs 7.1 and direct vs stereo and still no crosseyed crossovers. I'm at a loss to explain your graphs. I can't spot anything you did wrong but your results are sure different than mine Start from factory default and see if you see the problem thne start adding your settings one at a time. This is very strange.... Thanks for chiming in here, Ray, it is very helpful. I just want to confirm that, in terms of processing, the filters are applied after the crossovers. It'll be good to get to the bottom of Tony's measurements, but at least from a processing stand-point it will be good to confirm that the XMC is designed to apply the Dirac filters after the crossovers. |
|
|
|
Post by monkumonku on May 8, 2015 12:40:07 GMT -5
Out of curiosity, and to repeat what socketman asked early on in this thread to which it doesn't seem like anyone answered, what are the audible effects of this strange routed bass behavior? Or is it all reading graphs and stuff?  |
|
|
|
Post by geebo on May 8, 2015 12:45:41 GMT -5
Out of curiosity, and to repeat what socketman asked early on in this thread to which it doesn't seem like anyone answered, what are the audible effects of this strange routed bass behavior? Or is it all reading graphs and stuff? It all sounds fantastic with mine. So I'm very interested in hearing what the outcome of this is. |
|
|
|
Post by petes on May 8, 2015 12:56:17 GMT -5
Out of curiosity, and to repeat what socketman asked early on in this thread to which it doesn't seem like anyone answered, what are the audible effects of this strange routed bass behavior? Or is it all reading graphs and stuff? Audible effect would depend on what you filters happened to be - how different they were between your fronts and sub. If the fronts had a large gain applied at a freq just below the crossover, if this gain-ed signal was then shifted to the sub, you'd be sending a boosted signal to you sub. If the filters were similar, you wouldn't notice much of course. This all depends on whether this is actually a real problem of course :-) We now think that filters are applied after crossover as you'd want (according to a phone call with Emo referenced above) so this may actually be an odd issue. |
|
|
|
Post by monkumonku on May 8, 2015 12:59:35 GMT -5
Out of curiosity, and to repeat what socketman asked early on in this thread to which it doesn't seem like anyone answered, what are the audible effects of this strange routed bass behavior? Or is it all reading graphs and stuff? Audible effect would depend on what you filters happened to be - how different they were between your fronts and sub. If the fronts had a large gain applied at a freq just below the crossover, if this gain-ed signal was then shifted to the sub, you'd be sending a boosted signal to you sub. If the filters were similar, you wouldn't notice much of course. This all depends on whether this is actually a real problem of course :-) We now think that filters are applied after crossover as you'd want (according to a phone call with Emo referenced above) so this may actually be an odd issue. Well what I meant was what was the actual audible effect for those who posted that they had these problems in the thread. Was it so audible that it was disturbing or are we just talking theory here. |
|
|
|
Post by raydyo on May 8, 2015 16:31:44 GMT -5
Tony, I repeated my tests with all combinations of 2.0 vs 7.1 and direct vs stereo and still no crosseyed crossovers. I'm at a loss to explain your graphs. I can't spot anything you did wrong but your results are sure different than mine Start from factory default and see if you see the problem thne start adding your settings one at a time. This is very strange.... Thanks for chiming in here, Ray, it is very helpful. I just want to confirm that, in terms of processing, the filters are applied after the crossovers. It'll be good to get to the bottom of Tony's measurements, but at least from a processing stand-point it will be good to confirm that the XMC is designed to apply the Dirac filters after the crossovers. Confirmed - Dirac is second last processing step, delays (distance) is last. Then on to the analog world. The graphs that Tony and other have supplied seem impossible. Yet they exist and I don't think they were made by hand just t screw with my head. |
|
|
|
Post by ansat on May 8, 2015 16:47:09 GMT -5
Thanks for chiming in here, Ray, it is very helpful. I just want to confirm that, in terms of processing, the filters are applied after the crossovers. It'll be good to get to the bottom of Tony's measurements, but at least from a processing stand-point it will be good to confirm that the XMC is designed to apply the Dirac filters after the crossovers. Confirmed - Dirac is second last processing step, delays (distance) is last. Then on to the analog world. The graphs that Tony and other have supplied seem impossible. Yet they exist and I don't think they were made by hand just t screw with my head. Talked to Ray -- gathering more info, I will post the findings once I have everything. Tony |
|
|
|
Post by ansat on May 8, 2015 20:03:45 GMT -5
Some interesting results this go around.  YELLOW = FULL RANGE FRONT (Stayed the same in all 3 test sets) LIGHT GREEN = LFE in LFE out with factory restore and settings reloaded. DARK GREEN = Left in Left out with factory restore and settings reloaded. RED = LFE in LFE out before factory restore (was not able to get this result again after the factory reset) So I ran three tests. I measured full range left speaker, LFE, xover left speaker and xover sub for each of the three tests. The first test was my settings before a factory restore. The second was after a factory restore. The third was after a restore of my settings. On the initial results I was able to get some odd behavior from the LFE channel, but after the factory reset, I could not get the same odd results even when I restored my settings. I attempted to move the levels, enhanced bass, Reference stereo and changing the crossover. No matter what I changed, the reloaded settings and the factory restore results are now the same and do not exhibit the same trouble seen before the restore. Sorry raydyo, I have no method of reproducing the results any further since the restore. Tony All measurements - my first set & all 3 tests today. All measurements |
|
|
|
Post by socketman on May 8, 2015 20:48:57 GMT -5
Tony can you detect any audible difference. Things seem to be working correctly after the factory restore but still would be nice to know what caused the anomalies. I wonder if Nick has restored his machine.
|
|
