cawgijoe
Emo VIPs 
"When you come to a fork in the road, take it." - Yogi Berra
Posts: 5,033
|
Post by cawgijoe on Feb 7, 2015 14:40:39 GMT -5
Thanks guys.
It does sound "right" to me. From highs to lows. I have not had to boost anything.
That doesn't mean I won't buy the Dirac upgrade though. Another fun thing to play with and $100 is really not that much.
|
|
KeithL
Administrator 
Posts: 10,269
|
Post by KeithL on Feb 7, 2015 15:49:33 GMT -5
It's even worse that that.... While the real audio purists are trying to get that recording to sound really right on their high-end systems, some of those mixing engineers are doing the opposite - trying to screw up the mix so it sounds good on the $20 ear buds and car radios that they think a lot of their customers will be listening on. (Luckily, at least so far, this happens a lot more with two-channel audio than with home theater - presumably because not so many people listen to Blu-Ray movies on their iPod... yet.) It's all reference vs preference. If you're talking about movies - You're trying to reproduce the results that were heard by the rerecording mixer at the dubbing stage (generally a very flat FR - and using good Genelec monitors in many cases) - what the director intended. My understanding is that although dubbing stages use the X-curve - rarely does it apply to home theater size rooms. Otherwise, you're making a curve up that isn't consistent with what was heard when it was mixed - but you like how it sounds more. For music, there really isn't a benchmark. Half the stuff produced these days is meant for iPod earbuds. For movies, I like a bump in the LFE range - but I generally put faith in the re-recording mixers doing their job correctly - and wanting to reproduce that experience. Is reference "right"? it's all relative. |
|
|
|
Post by socketman on Feb 7, 2015 16:04:27 GMT -5
It's even worse that that.... While the real audio purists are trying to get that recording to sound really right on their high-end systems, some of those mixing engineers are doing the opposite - trying to screw up the mix so it sounds good on the $20 ear buds and car radios that they think a lot of their customers will be listening on. (Luckily, at least so far, this happens a lot more with two-channel audio than with home theater - presumably because not so many people listen to Blu-Ray movies on their iPod... yet.) It's all reference vs preference. If you're talking about movies - You're trying to reproduce the results that were heard by the rerecording mixer at the dubbing stage (generally a very flat FR - and using good Genelec monitors in many cases) - what the director intended. My understanding is that although dubbing stages use the X-curve - rarely does it apply to home theater size rooms. Otherwise, you're making a curve up that isn't consistent with what was heard when it was mixed - but you like how it sounds more. For music, there really isn't a benchmark. Half the stuff produced these days is meant for iPod earbuds. For movies, I like a bump in the LFE range - but I generally put faith in the re-recording mixers doing their job correctly - and wanting to reproduce that experience. Is reference "right"? it's all relative. Then they remaster the music and put it on a sacd (if your lucky) for 50 bux so you have to buy it twice. |
|
Lsc
Emo VIPs
Posts: 3,435
|
Post by Lsc on Feb 7, 2015 16:11:28 GMT -5
But don't you agree that something could be wrong with yours and you could still be satisfied, I don't believe those two things are mutually exclusive. If that's the case, then how do we know what is "right"? Or is there such a thing? Is my Dirac, Audessey, MCACC, etc. really, truly correcting my room and sound optimally? Is it all just a compromise? So, are the folks like Viper6 and myself who seem happy with Dirac LE results just buying fools gold? I'm honestly asking. What's right is what sounds best to you. But if you want more bass and detailed highs with smoother mids, the Socketman off-axis file will get you there. It's better than the house calibration file...it's a big enough difference to notice. However, the house calibration curve is still pretty good just not as good. Also, I'm not sure if Keith or Lonnie tried Socketmans file or their own 90 degree offaxis file, but to really understand what we mean, perhaps you guys should try it out. It's very clear what you guys are saying, however I and other have already gotten better results in our opinions. The proof is in the listening...just wanting some feedback. |
|
|
|
Post by socketman on Feb 7, 2015 16:32:51 GMT -5
The graphs can tell you a lot as well. I built my own center channel/cabnet/crossover and Dirac shows its high end is excessive and Dirac pulls it down a bit much for my taste and it sounds flat. I need to add a resistor to the tweeter. Careful mic positioning and multiple tries I have reached a correction curve I can live with. There is painters tape all over the floor and couch LOL Getting the house quiet and setting the meters so it compliments all speakers took a few tries, so don't think the first run is the best run. Consistency is everything when doing multiple runs and make notes. I certainly am no expert, but a long with Tony's help I have made some positive improvements which is why I bought the XMC in the first place. This is science , and results from science do not come easy.
Richard
|
|
KeithL
Administrator
Posts: 10,269
|
Post by KeithL on Feb 7, 2015 17:01:11 GMT -5
I'm really trying not to be argumentative here, but when you say things like "flat", you need to define your terms and parameters....
Anechoic measurements are easy; because there's no room involved. Put your calibrated microphone right in front of the speaker, take your measurement, and "flat is flat". (We'll ignore for now that some speakers store a little energy, so even their frequency response may be slightly different depending on how you measure it.)
However, as soon as a room gets involved, things get interesting. Rooms have messy characteristics like reverberence... which is a simple way of saying that, once you "launch" a sound into the room, it tends to bounce around for a while and die down slowly. There's a thing called "T60" - which is simply the time it takes for the loudness of a given test frequency to die down to a level of -60 dB below its original level in the room once you turn the sound off. In simplest terms, this is how you measure "how live" a room is. Unfortunately, this T60 is different at different frequencies, so you don't get a number - you get a graph of "T60 vs frequency" (both room modes and nulls, and what your walls are made of and covered with affect the T60 curve).
Now, you need to think about what this means. Let's assume I have a nice test signal; white noise. Let's also assume that I have a nice simple 20 foot square room, with the speakers on one wall, ten feet apart, and the microphone ten feet in front of the speakers. A little geometry will tell us that the microphone is about 14 feet from each speaker. (We'll ignore the obvious issues with a symmetrical room for now.)
We turn on the test signal.
At 10 milliseconds our microphone hears.... nothing (the sound hasn't had time to travel the 14 feet from the speaker to the microphone yet)
At 15 milliseconds.... now we're hearing and measuring the sound arriving directly from the speaker (it's had time to arrive, but the sound bouncing off the walls is still on the way). Actually, we might already have some early bounces from the ceiling and floor, but let's ingore those.
(Many measurement systems use various techniques like "windowing" to generate this measurement by "turning off" the microphone before the reflections reach it. The same result can be achieved by using an impulse as a test tone, and simply "not counting" anything except the early information. The only weakness to doing so is that it entirely removes the room from the measurement, which is great for accurately measuring parts like speaker drivers, but totally unrealistic in terms of relating to what you'll hear in that room with actual music.)
At 45 milliseconds.... now the sound that is reflecting off the back and side walls is reaching the microphone along with the sound coming directly from the speaker. However, since our room has a T60 that isn't the same at every frequency, that reflected sound has an altered frequency response (think of a white box, being lit by a white light bulb, sitting in a room with colored walls).
At 100 milliseconds.... now we're getting not only the direct sound from the speaker, but the initial reflections from the walls, and the reflections of those reflections, all added together (and, just like light, the frequency response of that sound is altered each time it bounces off of something). Again, we won't even mention things like cancellations and room modes, which just make it even more complictaed.
As you can see, unless your room has a T60 graph that is exactly the same at all frequencies, which isn't going to happen (unless it's an anechoic chamber, which has a T60 of "zero", or you've had some very successful room treatment done), each and every one of those measurements will be different. And they won't stop changing until you get past the longest T60 (the highest peak on that T60 graph) - which is the point where the effect of the reflections will stabilize to a steady value. For a typical large "live" room, the T60 could be as long as several SECONDS (500 milliseconds would be somewhat typical for a more or less normal live room)..
So, which one is right?
Which one would you LIKE to make flat?
Remember that, as the saying goes, you can ONLY CHOOSE ONE to make flat.
Well, you would think that would depend on which one correlates best with what we actually HEAR, right?
Unfortunately, THAT discussion has been going on for decades, with no end in sight.
Long before modern DSPs, and advanced room correction stuff like Dirac, when all you had was actual room correction, how to make "the best sounding room" was a very popular topic of discussion. Most people agreed that the best sound would be had with a room that was "part dead" and "part live" - which yielded an overall reasonable T60 - and didn't sound like an anechoic chamber or a cave. One camp was convinced that the front half of the room should be live (wood floors and panelling) while the back half should be dead. This provided for lots of "early reflections" (which make the room sound "lively" but tend not to mess up imaging), and very few "long reflections" (which our brain recognizes as echoes, and so mess up our perception of where a sound originates because they distract us). Others insisted the opposite, that those "harmless" early reflections weren't really harmless, and all that sound that diffracted around the speaker and ricocheted off the back wall before finally reaching us actually blurred the image - because, while our brains could pick out and ignore the obvious echoes, early arrival times confused the delicate mechanisms that our brain uses to locate where a source is by the phase between the arrival times.
If you read enough books on acoustics, you will see arguments made for both, as well as "live side walls and dead end walls". You will also see suggestions for mixing in diffusers, which, unlike flat walls, "break up" the reflections, which in turn alters how our brain interprets them. (Perhaps, by properly diffusing the reflections, our brain will percieve them as "live" without being confused by them.) Most studios and concert halls these days use a combination of absorbers and diffusers. And, if you haven't noticed, since we're talking about stuff that is dependent on the MECHANICAL properties of the room, it's all stuff that CAN'T BE ENTIRELY ADDRESSED BY ANY ROOM CORRECTION SYSTEM. (A room correction system can ADD extra "artificial" reflections, and it can cancel some reflections, but only to a limited degree, and for certain target locations. It cannot magically "eat" all that sound that's bouncing around and so turn a live room into a dead one. Dirac Live can correct frequency response and, unlike lesser systems, can also adjust many of the phase "discrepancies" that often contribute to poor imaging and less-than-stellar sound quality, but it CAN'T make a dead room live, or make a live room dead.... It is just one (very powerful) tool in the arsenal of room correction.
I'm seeing a sort of disturbing trend here to presume that "if we could measure things just right" or "if we could just find the perfect correction curve we could make everything perfect". and, more to the point "if I ran this and things didn't come out perfect, then it didn't work right". It reminds me a bit of the guy who was convinced that, if he could only find the perfect hack for his car's computer, his Jetta would outperform a Ferrari, and get 123 miles-per-gallon doing it.... or that guy who's sure that the 1/4" camera lens in his cell phone can take a picture just as good as my Nikon d800 if the software you use with it is good enough. I've been watching a lot of people expend a lot of time and effort on this one aspect of things, and I would like to suggest that some of you consider expanding your horizons a bit and learning about things like room acoustics, and T60, and how absorbers and diffusers affect things like imaging and frequency response. There's a whole world out there, and Dirac Live is only one (very cool and powerful) part of it.
I'm going to start up a few discussions about the OTHER aspects of room acoustics - like room treatments, and optimum speaker placement, and even optimizing room construction..... look for a few to pop up on "New Threads" and "New Posts" later today or tomorrow.
|
|
|
|
Post by barrak on Feb 7, 2015 19:28:21 GMT -5
I'm going to start up a few discussions about the OTHER aspects of room acoustics - like room treatments, and optimum speaker placement, and even optimizing room construction..... look for a few to pop up on "New Threads" and "New Posts" later today or tomorrow. Those would be really welcome, and would be of greater value to the uninitiated if they're kept as jargon-free as possible. You have my vote on such topics as; 1) protecting speakers from room correction abuse (you'll be seeing a lot of that once clueless users start pushing their overwhelmed woofers with Dirac Full), 2) simple DIY treatments (e.g., get rid of that carrrier deck-sized coffee table), 3) boomy bass and speaker distortion detoxing, 4) highlighting the real benefits that room correction could provide (e.g., reign-in room modes and articulate vocals). |
|
Lsc
Emo VIPs
Posts: 3,435
|
Post by Lsc on Feb 8, 2015 0:42:37 GMT -5
I finally found something that helps me understand why Socketman's file sounds better to me and other folk's logic isn't crazy.
From the minidsp's manual on calibration w Dirac...
"For home theater applications, it is best to use the 90-degree calibration file as this is created specifically for the vertical microphone orientation. This file is downloaded with the suffix “_90deg” in the file name."
So, it makes sense as to why it sounds best for those of us with 5.1 or 7.1 when the 90 degree off axis calibration file is used since the mic is used in a vertical orientation.
Also, I wonder if this means that if you are only concerned with two channel only like the demo at Emofest, if using the zero or 45 degree calibration file and pointing the mic toward the speakers produce better results. Makes sense since you are only receiving sound signals from the front speakers.
|
|
|
|
Post by ansat on Feb 8, 2015 1:55:13 GMT -5
I finally found something that helps me understand why Socketman's file sounds better to me and other folk's logic isn't crazy. From the minidsp's manual on calibration w Dirac... "For home theater applications, it is best to use the 90-degree calibration file as this is created specifically for the vertical microphone orientation. This file is downloaded with the suffix “_90deg” in the file name." So, it makes sense as to why it sounds best for those of us with 5.1 or 7.1 when the 90 degree off axis calibration file is used since the mic is used in a vertical orientation. Also, I wonder if this means that if you are only concerned with two channel only like the demo at Emofest, if using the zero or 45 degree calibration file and pointing the mic toward the speakers produce better results. Makes sense since you are only receiving sound signals from the front speakers.
Lsc, the 0 degree needs to be pointed directly at the speaker. While it would be slightly more accurate pointing it towards the center of the room then up, It still would not be accurate for 2 channel listening as the microphone will be at an angle to the speaker still. But if someone out there only has a two channel setup, point it towards the center of the room, you will get less treble from that result, if its too little treble, then change it to point at 45 degrees or you can go anywhere in-between.
good or bad -- This picture shows what will be added or subtracted from diracs end result. Red Shows the areas that are louder then the standard file and blue shows areas where it would be quieter. Easy explanation would be more low end and less high end with a few trivial tweaks in the middle. However since all microphones will have some variation (usually -+1.5db, and I believe that that is the standard that emotiva is using) the gains/cuts may vary.
Tony |
|
|
|
Post by viper6 on Feb 8, 2015 13:52:18 GMT -5
It is really quite simple. For many of us (and given the number of XMC-1 units in the field, and the number of owners who have posted about the issue, I think "many" can reasonably be defined as "the vast majority") this is a non issue. I ran Dirac with the original cal file. I did NOT lose my bass, I did NOT get shrieking highs, and my system sounds better than ever. Better detail, clearer more defined voices, any way you care to describe the quality of the sound my system reproduces is better, with no discernable negatives. While I am not closed to the idea of trying a different cal file, that is due to simple curiosity, not to any dissatisfaction with my results. It all boils down to how it sounds. That is what matters to me. Not what the graphs look like, or what the data from someone else's system looks like. I am not here to tell anyone else that they should be happy with the way their system sounds if they are not satisfied, but at the same time I don't need any one telling me something is wrong with mine when I AM satisfied. But don't you agree that something could be wrong with yours and you could still be satisfied, I don't believe those two things are mutually exclusive. What I know is that my system is dramatically better with Dirac Live LE. I know that there is "always" something else to be done to achieve yet another incremental improvement is quality. I know that the better the sound gets, the more difficult and expensive it becomes to achieve that improvement. I know that with enough effort (and the "right" mindset) it is possible to find fault with anything, or anyone. After all, when you get right down to it, and taking everything into account, Scarlett Johansson is actually a DOG. Right? And I know that focusing on what "might" be wrong is NOT conducive to enjoying all the things that are right. |
|
Lsc
Emo VIPs
Posts: 3,435
|
Post by Lsc on Feb 8, 2015 15:47:36 GMT -5
I finally found something that helps me understand why Socketman's file sounds better to me and other folk's logic isn't crazy. From the minidsp's manual on calibration w Dirac... "For home theater applications, it is best to use the 90-degree calibration file as this is created specifically for the vertical microphone orientation. This file is downloaded with the suffix “_90deg” in the file name." So, it makes sense as to why it sounds best for those of us with 5.1 or 7.1 when the 90 degree off axis calibration file is used since the mic is used in a vertical orientation. Also, I wonder if this means that if you are only concerned with two channel only like the demo at Emofest, if using the zero or 45 degree calibration file and pointing the mic toward the speakers produce better results. Makes sense since you are only receiving sound signals from the front speakers.
Lsc, the 0 degree needs to be pointed directly at the speaker. While it would be slightly more accurate pointing it towards the center of the room then up, It still would not be accurate for 2 channel listening as the microphone will be at an angle to the speaker still. But if someone out there only has a two channel setup, point it towards the center of the room, you will get less treble from that result, if its too little treble, then change it to point at 45 degrees or you can go anywhere in-between.
good or bad -- This picture shows what will be added or subtracted from diracs end result. Red Shows the areas that are louder then the standard file and blue shows areas where it would be quieter. Easy explanation would be more low end and less high end with a few trivial tweaks in the middle. However since all microphones will have some variation (usually -+1.5db, and I believe that that is the standard that emotiva is using) the gains/cuts may vary.
Tony How are you arriving at just how much is being corrected? I think I'm missing something here. |
|
|
|
Post by ansat on Feb 8, 2015 15:53:35 GMT -5
Lsc, the 0 degree needs to be pointed directly at the speaker. While it would be slightly more accurate pointing it towards the center of the room then up, It still would not be accurate for 2 channel listening as the microphone will be at an angle to the speaker still. But if someone out there only has a two channel setup, point it towards the center of the room, you will get less treble from that result, if its too little treble, then change it to point at 45 degrees or you can go anywhere in-between.
good or bad -- This picture shows what will be added or subtracted from diracs end result. Red Shows the areas that are louder then the standard file and blue shows areas where it would be quieter. Easy explanation would be more low end and less high end with a few trivial tweaks in the middle. However since all microphones will have some variation (usually -+1.5db, and I believe that that is the standard that emotiva is using) the gains/cuts may vary.
Tony How are you arriving at just how much is being corrected? I think I'm missing something here. This is the difference in the cal files. Its emotiva cal plus the curve subtracted from the cross spectrum cal (this gives the post dirac target). Since the cs file gives flat input, that would mean the post dirac target would be +3 to -3 from 30hz th 20khz when using the cs file. I just laid them on the same screen and colored them in. Tony |
|
klinemj
Emo VIPs
Official Emofest Scribe
Posts: 15,090
Member is Online
|
Post by klinemj on Feb 8, 2015 16:32:38 GMT -5
After all, when you get right down to it, and taking everything into account, Scarlett Johansson is actually a DOG. Right? Well, if SJ is a dog, then consider me a dog lover.  But to build on the point from rswood that got you to post this - it seems simple and harmless to try Tony's proposed improvement. What have you got to lose in trying it? Mark |
|
|
|
Post by socketman on Feb 8, 2015 17:42:26 GMT -5
This morning I did a DIRAC re-run. I slightly changed the measurement positions (a bit closer to the main spot) and I reduced the MIC sensitivity and hence increased the test tone output a bit more compared to previous measurements. Here are the results (only showing the front speakers). ORIGINAL EMOTIVA CAL FILE:  CROSS SPECTRUM CAL FILE:  I can conclude that the result of the CS CAL file resembles more to original output level of the speakers, especially in the higher frequencies. First listening results is that I liked the sound of DIRAC with CS CAL file a bit more. There is some more deep bass and the higher frequencies are bit less prominent. For the time being I'll keep the calibration with the CS CAL file. Just I bit strange is the small "dip" around 600 Hz in the spectrum before correction. You can see that the CS CAL file data has a slight change in sensitivity in that area. Strange. Anxiously waiting for the DIRAC Full version. Probably I'll sent my EMM-1 mic to Cross Spectrum for calibration (if shipping costs to and from USA - NL are not to crazy.). Again very big thanks to everybody who have contributed to make this happen. Tell us a bit about your room, do you have any treatments etc how big is the room. It seems that your DefTecs suit that room very well, how do they sound . Richard |
|
|
|
Post by rswood on Feb 8, 2015 19:46:24 GMT -5
But don't you agree that something could be wrong with yours and you could still be satisfied, I don't believe those two things are mutually exclusive. What I know is that my system is dramatically better with Dirac Live LE. I know that there is "always" something else to be done to achieve yet another incremental improvement is quality. I know that the better the sound gets, the more difficult and expensive it becomes to achieve that improvement. I know that with enough effort (and the "right" mindset) it is possible to find fault with anything, or anyone. After all, when you get right down to it, and taking everything into account, Scarlett Johansson is actually a DOG. Right? And I know that focusing on what "might" be wrong is NOT conducive to enjoying all the things that are right. Every day hot girls get dumped by some guy, who thinks some other girl is hotter. If what you have makes you happy that is all that matters. |
|
|
|
Post by viper6 on Feb 8, 2015 20:46:07 GMT -5
After all, when you get right down to it, and taking everything into account, Scarlett Johansson is actually a DOG. Right? Well, if SJ is a dog, then consider me a dog lover. But to build on the point from rswood that got you to post this - it seems simple and harmless to try Tony's proposed improvement. What have you got to lose in trying it? Mark Well, when the the Mac interface is released I might. In the meantime my thirty day trial of VMware Fusion has expired, I am MORE than satisfied with the results I achieved from Dirac, don't have a compelling need for Fusion, and would rather save the $79 cost for a Fusion license for the price of Dirac Full, if I should decide I want it. |
|
klinemj
Emo VIPs
Official Emofest Scribe
Posts: 15,090
Member is Online
|
Post by klinemj on Feb 8, 2015 20:55:54 GMT -5
That is a good reason to wait.
Mark
|
|
Lsc
Emo VIPs
Posts: 3,435
|
Post by Lsc on Feb 8, 2015 21:35:15 GMT -5
That is a good reason to wait. Mark Lol! The poor Mac users. Hopefully they'll get that done in 2015  . My buddy is still waiting for it as well... |
|
|
|
Post by socketman on Feb 20, 2015 19:01:00 GMT -5
Dirac is now working excellent for me. I made a big mistake when building my crossover, I needed a 24uf capacitor but only had a 22uf and a 2uf. Even though I knew better I connected them in series which results in a 2uf capacitance rather then a 24uf which caused the midrange to be non existent. This in turn confused Dirac causing it to do some funky things to my sound. I had sent REW files to Tony and he commented back that something was funky with my center speaker but just didn't know WTF I had done . I can now move on to some bass traps and move my subs around to get rid of my nasty 45hz hump. If I had not gotten the XMC-1 and messed with dirac I would have never known what the hell was going on. Color me happy.
Richard
|
|
|
|
Post by wilburthegoose on Feb 22, 2015 7:06:58 GMT -5
|
|
. My buddy is still waiting for it as well...
