|
|
Post by ansat on Feb 6, 2015 22:20:21 GMT -5
The full version should allow you to use any calibrated mic. With shipping overseas. It might be cheaper to pick up one of their umik microphones. Tony Okay, this is the result using the CS calibration, right? Is there really a spike in your speakers of about +8db from to 12 to 20kHz before Dirac? I don't think so, I have been seeing this in the math calibration as well, I had it removed from the Ansat cal, but then CS validated it. I am curious to if this is due to the distance, reflections and capsule size. I have not had time to dig into it. Since its at the end of the highest octave and in all the cases I looked closely at, Dirac has moved the house curve up with it at the same rate, it doesn't seem to be affecting the final result. But it is curious. Tony |
|
|
|
Post by geebo on Feb 6, 2015 22:50:32 GMT -5
Okay, this is the result using the CS calibration, right? Is there really a spike in your speakers of about +8db from to 12 to 20kHz before Dirac? I don't think so, I have been seeing this in the math calibration as well, I had it removed from the Ansat cal, but then CS validated it. I am curious to if this is due to the distance, reflections and capsule size. I have not had time to dig into it. Since its at the end of the highest octave and in all the cases I looked closely at, Dirac has moved the house curve up with it at the same rate, it doesn't seem to be affecting the final result. But it is curious. Tony Visible on Lsc's CS trace as well. Curious it is. |
|
|
|
Post by bargain on Feb 6, 2015 23:21:27 GMT -5
Hi lads. Truly impressive work Tony, much respect and appreciation mate.
Out of curiosity Tony, are you using your math file or the CS file?
I've had a slight problem and sort of given up... I put the CS file into Dirac and started running measurements. The first 'sweet spot' runs fine, but the second measurement has an error EVERY time. Something like TIMEDIST_MAXCORR_PEAK_INDISTINCT
I checked mic settings, AGC off, volume 100% etc
I tried lower input, higher input, lower output, higher output, slightly different mic positions, moved furniture, slightly moved speakers...
Out of ideas.
This never happened with the default emo cal file, that's all I can put it down to.
Anyone familiar with this error and what it means i.e. what I need to change to correct it?
Thanks.
|
|
|
|
Post by socketman on Feb 6, 2015 23:27:08 GMT -5
which file do you have, get the one from tony's signature. The earlier versions had phase included and I don't think it will work with that in there. It is a circle with a line through it after each correction.
Richard
|
|
|
|
Post by bargain on Feb 6, 2015 23:51:06 GMT -5
Yea I grabbed the file from Tony's signature about 2 or 3 hours ago. Using the CS file. Didn't try the math file.
Thank you too, socketman, from down here in Aus.
|
|
|
|
Post by ansat on Feb 6, 2015 23:55:23 GMT -5
Hi lads. Truly impressive work Tony, much respect and appreciation mate. Out of curiosity Tony, are you using your math file or the CS file? I've had a slight problem and sort of given up... I put the CS file into Dirac and started running measurements. The first 'sweet spot' runs fine, but the second measurement has an error EVERY time. Something like TIMEDIST_MAXCORR_PEAK_INDISTINCT I checked mic settings, AGC off, volume 100% etc I tried lower input, higher input, lower output, higher output, slightly different mic positions, moved furniture, slightly moved speakers... Out of ideas. This never happened with the default emo cal file, that's all I can put it down to. Anyone familiar with this error and what it means i.e. what I need to change to correct it? Thanks. Bargain, I do not use either, I have another that is matched almost exactly to my microphone. with the math cal, is the average over 5 microphones, then the difference expected for 90 vs 0 on the high end. If the calibrations in my signature do not work, then delete them and let dirac load the original calibration to see if that fixes the problem. If it does them PM me. Tony |
|
|
|
Post by viper6 on Feb 7, 2015 0:04:17 GMT -5
Well, I for one find it troubling how many people still don't see the problem after seeing it illustrated with Tony's graphs and all the data. You can't take a calibration, use it on a different axis and say "it's all good, move along". But that's the response I'm reading. Am I off base here? I've been on vacation. This is a black and white issue. Either the calibration is correct or it's not. The data argues that the included calibration being wrong. 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. |
|
KeithL
Administrator 
Posts: 10,271
|
Post by KeithL on Feb 7, 2015 1:00:58 GMT -5
The error you're quoting means that Dirac can't pick the signal it needs out of the background noise cleanly enough to analyze it. The usual fix would be to play the tones a tiny bit louder, or, if that's not practical, then eliminate or lower any background noise that may be interfering with the readings. It shouldn't be specifically linked with a calibration file, although the calibration file you use could certainly affect whether it happens in a given situation or not. Hi lads. Truly impressive work Tony, much respect and appreciation mate. Out of curiosity Tony, are you using your math file or the CS file? I've had a slight problem and sort of given up... I put the CS file into Dirac and started running measurements. The first 'sweet spot' runs fine, but the second measurement has an error EVERY time. Something like TIMEDIST_MAXCORR_PEAK_INDISTINCT I checked mic settings, AGC off, volume 100% etc I tried lower input, higher input, lower output, higher output, slightly different mic positions, moved furniture, slightly moved speakers... Out of ideas. This never happened with the default emo cal file, that's all I can put it down to. Anyone familiar with this error and what it means i.e. what I need to change to correct it? Thanks. Bargain, I do not use either, I have another that is matched almost exactly to my microphone. with the math cal, is the average over 5 microphones, then the difference expected for 90 vs 0 on the high end. If the calibrations in my signature do not work, then delete them and let dirac load the original calibration to see if that fixes the problem. If it does them PM me. Tony |
|
|
|
Post by ansat on Feb 7, 2015 1:02:43 GMT -5
I think that the big holdup is that there are quite a few of us that are chasing accuracy. Given Lonnie's statements, and running the numbers. If you combine the emotiva calibration with the house curve, the end result would be flat from 400hz to 10khz. You get a little boost from 50 to 400 and below 50hz it goes downhill pretty fast. Above 15khz it bad really quickly as well. If your content lives in the 50hz to 15khz range, I think that the results would be nice and flat. I think that the way that emotiva sent the LE out, good results are easily achievable depending on the content. I suspect that emotiva intended that the house curve be the +3 to -3 that shows in the graph. However, they determined that that curve was best using the calibration supplied to them. If they had an accurate calibration they might have decided that the house curve was -6 @ 20hz to +6 at 20khz is what sounded best (which would of been the case as the end result for both scenarios will result in the same frequency adjustment). It just would not look like what most typical house curves look like. (exaggerated graph below. Its hard to draw lines on a computer.)  For myself, (I spend way to much time in the numbers and the graphs) I saw what Dirac was reporting back and knew my speakers did not measure that way. In the end results, I also knew that the house curve was not -3 at the high end. -3 is my normal house curve on the high end. Given the house curve built into Dirac, I should of been pleased with the results. (little light on the bass for my tastes). Would I have rather seen the funny house curve above? That's hard to tell. I would likely be groaning about how it is going the wrong direction  [/p]
Tony
|
|
|
|
Post by bargain on Feb 7, 2015 1:51:29 GMT -5
Cheers guys. I'll muck around some more when I could be bothered, and thanks Keith, I'll try boosting the output a little and reduce the background noise a bit more. Until then I'm loving the sound of Dirac with the default emo cal  Like others I have upped the sub a bit and she's apples |
|
|
|
Post by ansat on Feb 7, 2015 1:55:22 GMT -5
There seems to be a little bit of a misunderstanding about the calibrated microphone we've provided you with Dirac. The calibrated microphone was designed to be used with Dirac Live and your XMC-1. It was never our intention that you should use the microphone by itself as a calibrated microphone, and we never provided you with a calibration curve for use with it. As you should be aware, the frequency response of a microphone can only be calibrated accurately by a single calibration curve for use in a specific orientation. Now, unlike other simpler room calibration programs, Dirac Live uses a variety of information, including the basic frequency response of a system and set of speakers, and other information about the timing and frequency characteristics of the room itself, to calculate its correction filters. And, since this information includes both on-axis and off-axis information, gathered simultaneously, neither curve by itself would provide an entirely accurate characterization of the response of the room. Dirac Live LE was customized to work with the microphone we provided you, and the "house curve" that is built into it, to create a set of correction filters that we believe will produce a pleasing overall sound for most people - with most systems - in most typical rooms. Since most of our customers who've used this version of Dirac Live seem to be quite pleased with the results, we've obviously achieved this goal. However, a few of our customers have managed to extract the calibration curve used by the software, and have “proven” that it isn't an "accurate off-axis calibration curve for the microphone" - which it was never claimed to be (remember that this is an internal calibration curve used by the software). At this point we’d like to confirm that the internal calibration curve used by Dirac Live LE for Emotiva is NOT an accurate off-axis response curve for our calibrated microphone, and is intended to be used ONLY by the Dirac Live software in conjunction with the provided house curve to calculate correction filters for your XMC-1. We regret if this has caused any confusion. The Upgraded version of Dirac Live - Dirac Live FULL for Emotiva - can be used with a variety of calibrated microphones. Since calibrated microphones, when purchased separately, typically come with an off-axis calibration curve (which is the most convenient orientation for use with surround-sound applications), Dirac Live Full works with off-axis calibration curves, and we will be providing an accurate off-axis calibration curve for our microphone to go with it (and any other programs you may also choose to use the microphone with). In addition, we just wanted you to know that we do recognize that not everyone is totally pleased with the correction filters as currently calculated by Dirac Live LE for Emotiva (using the microphone calibration and house curves we provided). This is why the Full version allows you to create your own correction curve (and to use whatever calibrated microphone and calibration curve you like). This will allow you to tailor both the measurements, and the correction filters, precisely to your personal preferences. Lonnie Lonnie, Thank you for digging further into this. I am excited for the full version of dirac and to see the remainder of your xmc plans unfold. Tony |
|
KeithL
Administrator
Posts: 10,271
|
Post by KeithL on Feb 7, 2015 3:09:58 GMT -5
There still seems to be a bit of confusion.... Dirac Live LE for Emotiva ships with a calibrated microphone. The two are a set; you can't use the software with any other microphone, and the microphone really isn't specifically intended to be used with any other software. The built-in Target Curve, and the built-in microphone calibration information, were designed to work together with that microphone to get just the results we wanted. Since the Microphone was provided for use only with the Emotiva versions of Dirac Live, and we did't provide it as a general purpose calibrated microphone, we didn't provide a standard calibration curve to go with it. It just so happens that, since we didn't employ any sort of security on the internal calibration file, it's there for you to read if you like. Dirac Live LE for Emotiva also wasn't intended as "a system analysis tool"; Dirac Live is a high-end speaker and room correction tool. The information provided by the graphs in Dirac Live is intended specifically to help you decide where to set things in Dirac; since there's not much you actually can set in the LE version of Dirac Live, that means that they are mostly there for informational purposes in this version; they give you a snapshot of what Dirac "hears" and what it plans to do about it. Certain of the information provided in those graphs could indeed be very useful outside of Dirac; for example, by seeing the low frequency response of your speaker in-room, you will be able to make a more informed decision about what crossover point to use with it. You should remember, though, that Dirac Live is intended PRIMARILY as a tool for controlling and configuring the Dirac Live room correction engine in the XMC-1. Dirac Live Full for Emotiva adds several important upgrades over the LE version. One improvement is that, with the Full version, you can adjust several parameters, and create and edit your own Target Curves, which makes the information in those graphs even more useful, and allows you to customize your results just the way you like them. Another is that Dirac Live Full can be used with a variety of readily available third-party calibrated microphones. This being the case, Dirac Live Full for Emotiva is configured to "expect" the standard calibration curves that come with third-party calibrated microphones. Now, as it turns out, our microphone is quite nice, and rather accurate, and we designed it to work equally well with both Dirac Live LE and Dirac Live Full for Emotiva ... and, since Dirac Live Full is designed to work with standard microphone calibration curves, we will be providing one of those to go with our microphone and Dirac Live Full. And, as an added benefit, that calibration curve will let you use your Emotiva Dirac microphone with many other measurement programs and get accurate results. ...And, no, you won't have to pay for the calibration curve  ). Incidentally, if you want a full-featured room measurement and analysis tool, we highly recommend REW (Room Equalization Wizard) - which you can download FOR FREE here: www.roomeqwizard.com/However, remember that, even with the same microphone, and the same calibration curve, the results you get when testing speakers in real rooms are influenced at least as much by the measurement methodology of the software you're using as thay are by the microphone and its calibration, so don't necessarily expect the graphs to always come out exactly the same when you use different measurement software. (For one example, some software has settings that allow you change the weighting between how much direct sound and reflected sound affect the result; and this setting, in turn, will determine how much effect the "liveness" of the room has on your results.) The REW manual has some very nice explanations about a lot of this - definitely recommended reading.  The combination of the microphone and THAT calibration file is specific.... We're not planning on changing the microphone itself.... and the new calibration file to use it with Dirac Full for Emotiva will be free.... Really.... Jeeez.....  Will the calibration file that is provided be a correct one that matches the recommended placement axis and be flat? Or is that a "really.....geez" question? I wonder how many people are going to use the full version with a real trustworthy calibration (CSL or others) and just be done with it. I don't understand why Emotiva is still beating around the bush. At the absolute very least say "hey, we did a house curve..... So don't expect flat" never mind the axis data looking all wrong |
|
|
|
Post by Jim on Feb 7, 2015 6:53:10 GMT -5
So, the takeaway seems to be... If you have low bass and exaggerated HF, it has to be mic placement.... Because it can't be the calibration file that has a downward curve on the low end and a upwards curve on the high side?
If the file with LE isn't really a calibration file (or just a cal file with a little house curve), then by that same logic you won't be able to use a flat calibration with the full version..... (because calibrations are generally only measured for only a single axis at a time)
All the talk about measurement tools is irrelevant. XTZ, REW, etc - is the only way to verify results. I don't think any of us is confusing an EQ solution with a FR measurement solution. We know what Dirac is.
|
|
|
|
Post by weigle2 on Feb 7, 2015 8:13:57 GMT -5
Thank you, Tony and Socketman for all your hard work on getting not only the bass back in my system, but also smoothing out the mid and high end. I used Tony's MathCal file and I could not believe the difference it made in smoothing out everything. The missing bass punch is back, especially on the very low end and mid's and high's never sounded better.
THANK YOU!
|
|
|
|
Post by rogersch on Feb 7, 2015 8:25:19 GMT -5
Dirac Live Full for Emotiva adds several important upgrades over the LE version. One improvement is that, with the Full version, you can adjust several parameters, and create and edit your own Target Curves, which makes the information in those graphs even more useful, and allows you to customize your results just the way you like them. Another is that Dirac Live Full can be used with a variety of readily available third-party calibrated microphones. This being the case, Dirac Live Full for Emotiva is configured to "expect" the standard calibration curves that come with third-party calibrated microphones. Just make the FULL version available for sale and I'll buy it immediately! |
|
|
|
Post by LuisV on Feb 7, 2015 8:37:20 GMT -5
Yea, agreed... when will it be available. BTW Mac version as well please...
|
|
|
|
Post by hesster on Feb 7, 2015 8:43:46 GMT -5
Yesterday I ran the CS Cal File twice and selected the best looking plots of the 2. I removed Tony's earlier furnished cal file to a thumb drive (which was an improvement in my system), and replaced it in the Apps folder so there was no confusion as to what to read in the set-up. AGC was turned off, mic level at max in windows, gain in the Emo set-up just below -24db, test signals average around -12db but not too high in the green bars or I get clipping with my woofer when the Mic position is closest to it.. Turned off heat and ceiling fan to get the room as quiet as possible. Mic positions carefully placed per recommendations. Following these guidelines work best for me to provide optimum results. The results are the best yet - for me and my system a vast improvement over the EMO cal. Not to degrade the EMO cal as it probably works just fine across many systems, but just stating that the latest Cal provides improvement in my system that can be easily heard. Still interested in the Full version to allow tweaks.
|
|
klinemj
Emo VIPs 
Official Emofest Scribe
Posts: 15,090
|
Post by klinemj on Feb 7, 2015 9:02:06 GMT -5
So, if my count is right, we have 3-4 saying the new cal file from Tony and only 1 saying "meh...no big deal". Great job Tony! If I had an XMC, I would be trying your file.
Mark
|
|
|
|
Post by geebo on Feb 7, 2015 9:05:49 GMT -5
So, the takeaway seems to be... If you have low bass and exaggerated HF, it has to be mic placement.... Because it can't be the calibration file that has a downward curve on the low end and a upwards curve on the high side? If the file with LE isn't really a calibration file (or just a cal file with a little house curve), then by that same logic you won't be able to use a flat calibration with the full version..... (because calibrations are generally only measured for only a single axis at a time) All the talk about measurement tools is irrelevant. XTZ, REW, etc - is the only way to verify results. I don't think any of us is confusing an EQ solution with a FR measurement solution. We know what Dirac is. Jim, I get no high frequency exaggeration with Dirac and the default cal files. I have verified this with OmniMic which comes with it's own individually calibrated microphone and specific custom test signals. The only thing for me that I adjust is the low end response which although reasonably flat is not to my taste mostly because of the chosen house curve. Many people prefer a house curve with an increase of somewhere between 6 and 10 dB beginning just below 100Hz. The Dirac LE house curve is much less than that. REW has a good write up on house curves. Changing the cal file and experimenting is one way to change that to your individual taste in lieu of adjustable target (ie house) curves. Here is an early Dirac result of mine with OmniMic. I don't consider the top end overly bright and have done nothing the change it nor do I want to. It sounds great the way it is. Here is some good info from REW regarding house curves. www.hometheatershack.com/forums/rew-forum/96-house-curve-what-why-you-need-how-do.htmlwww.hometheatershack.com/forums/rew-forum/6818-minimal-eq-target-levels-hard-knee-house-curve-long.html#post55660www.hometheatershack.com/forums/rew-forum/2462-house-curve-options.htmlAnd my early Dirac run.  |
|
|
|
Post by Jim on Feb 7, 2015 9:14:03 GMT -5
So, the takeaway seems to be... If you have low bass and exaggerated HF, it has to be mic placement.... Because it can't be the calibration file that has a downward curve on the low end and a upwards curve on the high side? If the file with LE isn't really a calibration file (or just a cal file with a little house curve), then by that same logic you won't be able to use a flat calibration with the full version..... (because calibrations are generally only measured for only a single axis at a time) All the talk about measurement tools is irrelevant. XTZ, REW, etc - is the only way to verify results. I don't think any of us is confusing an EQ solution with a FR measurement solution. We know what Dirac is. Jim, I get no high frequency exaggeration with Dirac and the default cal files. I have verified this with OmniMic which comes with it's own individually calibrated microphone and specific custom test signals. The only thing for me that I adjust is the low end response which although reasonably flat is not to my taste mostly because of the chosen house curve. Many people prefer a house curve with an increase of somewhere between 6 and 10 dB beginning just below 100Hz. The Dirac LE house curve is much less than that. REW has a good write up on house curves. Changing the cal file and experimenting is one way to change that to your individual taste. Here is an early Dirac result of mine with OmniMic. I don't consider the top end overly bright and have done nothing the change it nor do I want to. It sounds great the way it is. Here is some good info from REW regarding house curves. www.hometheatershack.com/forums/rew-forum/96-house-curve-what-why-you-need-how-do.htmlwww.hometheatershack.com/forums/rew-forum/6818-minimal-eq-target-levels-hard-knee-house-curve-long.html#post55660www.hometheatershack.com/forums/rew-forum/2462-house-curve-options.htmlAnd my early Dirac run. View AttachmentInteresting that your results are as flat as they are - despite what Tony's graphs show what the predicted response would be. Thanks for posting those links - I'm very familiar with all of those threads. I used to be very active over at the shack. I learned a lot when programming a Behringer Feedback Destroyer and later a MiniDSP to try to tweak my subs. Those are great threads about house curves. However, I believe that the stock calibration file rolls off the low content - not increases it. I don't understand why - when you'd think if you were building a house calibration - you'd put a small boost. Would be curious to see a comparison of the original cal vs a flat on from CSL - and see what Omnimic shows. I understand if you're happy with the original cal - so no worries. I think Omnimic is a great tool. |
|
[/p]
Like others I have upped the sub a bit and she's apples 
).
