Hi Carl…. Classic piece of work. So much detail and lot of clarity in what you do. Awesome work. Thank You so much for your effort and time in creating these masterpiece videos on speaker- room equalisation. A million thanks for your hard work…😊😊😊
Hi Carl, Thanks for this piece of excellent work! I adopted this technique thanks to you. For the Dirac Live users that want to use this method. After dividing your target curve by your actual measurement (or average of measurements if you are EQing multiple seating positions), which creates an inverted differences curve associated to 0dB, you then need to multiply that 0dB inverted differences curve by your target curve to shift those differences around the target curve. Then you export that result as text as Carl showed and make any required adjustments at lowest/highest frequencies as Carl showed, to become your new target curve. If you try this, then the graph makes sense visually and should work for Dirac Live Users. Always use common sense and don't load a target curve that requires boost anywhere greater than about 6dB or you could damage your equipment. Good luck!
Hi Carl, Thank you for sharing the Audio Calibration Basics files! These are very nice to use with REW so I can also calibrate my atmos height speakers.
This method seems to yield good results while taking some of the sting out of the higher frequencies. Definitely looking forward to your next video detailing your newest method you get your speakers to match the target so tightly. One thing you frequently mention is how "stock" Audyssey sounds terrible because it EQs everything flat, which makes me curious if you have ever tried Dynamic EQ? It'd be interesting to find an EQ method like yours that also allows for some sort of loudness compensation at lower volumes like Dynamic EQ does.
Good question Cafiveoh, I think I could just enable dynamic EQ and the calibration should sound as good as before. I believe you can set a reference level, like -10 on the receiver would be mean no dynamic EQ. That's when it hits full peak. I believe I can do that in the MultEQ-X app. Good questions, keep em coming, and thanks for watching.
My room is really reflective of anything over 1khz so the curve I ended up with has like a -20db fall from 20khz to 1khz lol. Sounds great to me. With acoustic treatment I could probably start to increase the highs a little more but as of now my system sounds the best it has ever sounded. Otherwise my curve is the same as yours. As far as noise floor when measuring, I live in an apartment with single pane windows and thin walls right next to a busy street. I have done measurements with plenty of background noise and I have done measurements with no background noise and the measurements were almost identical. I think REW does a good job at filtering out background noise.
Midrange compensation is related to the change in directionality as the speaker transitions from the tweeter to the mid. Basically, if the speaker design has an intentional dip of about 3-4dB at the speaker's crossover point between those two drivers, you don't want to EQ that out because doing so can make the speaker sound more harsh. So Audyssey's MRC is centered at 2kHz, which is where they believe generally most speakers will transition, and where directionality issues would generally occur. Here's the thing though: If you don't see a dip there in your measurements, there's likely no reason to use MRC. Most well-designed speakers won't need it. In your case, the Elac Uni-Fi uses a concentric design where the tweeter is in the middle of the mid, which minimizes directionality issues. Your Uni-Fis transition at 2.7kHz, and I don't see any dip in that region in your measurements, so there is absolutely no circumstance where I would ever use MRC with those speakers. Even if you wanted to, Audyssey's stock MRC isn't centered at the right frequency for your particular speaker. However, for those with other speakers, if you see about a 3-4dB dip around 2kHz, there may be value in enabling MRC in MultEQ-X for that speaker. If you see a dip anywhere else in that general area, look up the specs for your speaker and see where the crossover point is for that speaker's tweeter-to-mid transition. If the dip in your measurements is centered at about that same frequency, a custom MRC filter may be a good idea. You can do this by creating a parametric peaking filter with a Q of about 2.76 at the crossover frequency at about the same dip you're seeing in your measurements. So, for instance, if you see a 4dB drop at 2.5kHz on a speaker whose specs show that 2.5kHz is the crossover point, you would create a filter in MultEQ-X at 2.5kHz, Q of 2.76, gain of -4dB. You can tweak this to shape it more like the dip you have, but this will keep any other EQ work you're doing from trying to push that intentional dip out. Caveat: Ideally, you would need to see spinorama data for your speaker to see what the directionality index at your speaker's transition point looks like to know if you really NEED some form of MRC. But generally, if you measure a dip at the speaker's tweeter-to-mid transition, EQ'ing that out could make things sound worse. If things sound harsh to you, try MRC.
Thanks for all your content! I have followed many of your videos, and learn something new every time. Is there any reason not to use EQ functionality and "Generate measurement from filters" instead of trace arithmetic? It would presumably save some effort on manually editing the filters after export. Looks like the exported filters are more detailed when generated by EQ, but it didn't prevent import target in MultEQ-X for me. Also, what are your thoughts whether or not you are correcting for reflections in the high frequencies? I see OCA is using some different IR windows for the highest frequencies, and I know you have mentioned his channel on many occasions.
I haven't been introduced to that method of Generate measurement from filters. Can you export as text? I see OCA doing more in high frequencies, but OCA doesn't actually correction above 300hz, so it really doesn't matter. I like the way Psychoacoustic sounds, and I think it gives me the best tonality. I think you need to be an expert to draw any more beyond Psychoacoustic smoothing.
What's your opinion on not touching frequencies above the Schroedinger frequency? This is a topic often addressed in the context of room correction. In OCA's latest video, he talks about this as well.
@@hdmoviesource I can see why using VAR makes sense however, since octaves are a logarithmic scale, the lower the frequency the more notes are within a given range. E.g. from 20hz to 40hz is an octave, but also from 1000hz to 2000hz is an octave, so 20hz range for an octave at low frequencies vs 1000hz range for an octave at higher frequencies. Essentially there IS more detail in lower frequencies, so treating them more finely seems to make sense.
Yeah, I think it depends on how much control you want over bass frequencies. I'm finding that Psychoacoustic gives me the sound that I'm after. So, as long as you can get the sound that you're after with the smoothing that you're using that's all that matters.
Hi Carl… Please make a video on Speaker equalisation using Dirac ( standalone) Computer version if you have license or using a trial version. It would be immensely helpful to Dirac users ….
Quick question, are you measuring from multiple seats and using cross correlation and then vector average to eq? I saw a video showing this from oca a while ago and have been adding this to my eq process.
From a single seat. I used the cross-correlation method and it didn't sound good in my opinion. I find that single seat sounds best even for other seats.
do i get it right, that you do this method by using only ONE Target Curve (reference speaker FR) for ALL speakers; how does this work, if every speaker is 'it's own Target Curve' as far i understood!? Wouldn't it be necessary to have a target curve for each single speaker in my room!?
Thank you so much for this video!!! I'm using DIRAC on my 2.1 system and I finally got the sound I wanted after calibration. At first I was simply following the default target curves that DIRAC provides, but with those the sound felt very two dimensional. After watching your video, I traced a new response curve through the original, untreated response of the speakers in the room and used that as a target curve in DIRAC. It did exactly what I wanted, kept the original sound and coloration of the speakers, but removed most of the imperfections and tightened down the sound. The curve is nothing close to what the default curves in DIRAC are, but I love the sound it produces.
That sounds like a good idea to me Adam. Also, think about degrees and angles. I like to angle my speakers 15 degrees off-axis away from the MLP. That's the smoothest response that my speaker gives me. I may have to recheck that, but, always get each speaker to be the same degrees off-axis whichever angle you choose is something to think about.
@@hdmoviesource interesting, I’ll have to try that. That is closer than I have mine set up. Mine is about 45 degrees. I have recently rearranged according to the sound nodes. It made a bigger difference than I thought, especially with the bass.
Interesting. I find directivity to sound better like this. I've pointed the speakers directly at me once and I had no idea where sounds were coming from. LOL, so it can really make a huge difference.@@adamATOM3
@@hdmoviesource sorry I misunderstood that. I meant from MLP my speakers are in 45 degree triangle. You mean where you point your speakers? You point them a little wide of the MLP you mean?
Yes, I point them a little wide of my MLP. I've found that it's really opened up the soundstage and I can actually hear when things are coming from now.
my recommendation is not using a mic, or looking at graphs and that means no targets. sit down in listening position, do real time eq by ear so it sound evenly loud to you, you can call that flat if you will. pick a loudness that you most often use and that is it. require no tools and experience. and takes into account room, speakers, hearing, audio taste, loudness curve/house curve.
Always informative, thank you for your hard work and efforts to make our systems better 🙏
Thanks Christian, I love to hear comments like this. I hope it helps and give you some ideas.
That’s Saturday night sorted, love watching your EQ videos thanks
Thanks Paul, I love this stuff, and have a great Saturday night.
Hi Carl….
Classic piece of work. So much detail and lot of clarity in what you do. Awesome work. Thank You so much for your effort and time in creating these masterpiece videos on speaker- room equalisation. A million thanks for your hard work…😊😊😊
I really appreciate it.
Hi Carl, Thanks for this piece of excellent work! I adopted this technique thanks to you. For the Dirac Live users that want to use this method. After dividing your target curve by your actual measurement (or average of measurements if you are EQing multiple seating positions), which creates an inverted differences curve associated to 0dB, you then need to multiply that 0dB inverted differences curve by your target curve to shift those differences around the target curve. Then you export that result as text as Carl showed and make any required adjustments at lowest/highest frequencies as Carl showed, to become your new target curve. If you try this, then the graph makes sense visually and should work for Dirac Live Users. Always use common sense and don't load a target curve that requires boost anywhere greater than about 6dB or you could damage your equipment. Good luck!
Nice advice, thank you. I'll try Dirac soon.
Thank you Carl.
You're the man Cheekster, thank you for all your comments.
@@hdmoviesource 👊🏻
👊
Great video, this is exactly what we needed to hear. Thank you.
You are so welcome!
Hi Carl, Thank you for sharing the Audio Calibration Basics files! These are very nice to use with REW so I can also calibrate my atmos height speakers.
Absolutely yes, thank you.
Great video Carl. Thank you Carl.
Greetings from Austria !
Thanks you Dan, I really appreciate it.
Absolutely 💯 the best way I've ever seen to calibrate speakers this tightly 👍🙏🙌
Glad it helped Christian. I've just figured out a better way, video should be coming soon.
@@hdmoviesource ohhhhh yeah 👍 👍 👍 👍
I just finished a video on timing methods, hopefully that will be out soon, but been really busy lately.
@@hdmoviesource tease lol
lol
This method seems to yield good results while taking some of the sting out of the higher frequencies. Definitely looking forward to your next video detailing your newest method you get your speakers to match the target so tightly. One thing you frequently mention is how "stock" Audyssey sounds terrible because it EQs everything flat, which makes me curious if you have ever tried Dynamic EQ? It'd be interesting to find an EQ method like yours that also allows for some sort of loudness compensation at lower volumes like Dynamic EQ does.
Good question Cafiveoh, I think I could just enable dynamic EQ and the calibration should sound as good as before. I believe you can set a reference level, like -10 on the receiver would be mean no dynamic EQ. That's when it hits full peak. I believe I can do that in the MultEQ-X app. Good questions, keep em coming, and thanks for watching.
My room is really reflective of anything over 1khz so the curve I ended up with has like a -20db fall from 20khz to 1khz lol. Sounds great to me. With acoustic treatment I could probably start to increase the highs a little more but as of now my system sounds the best it has ever sounded. Otherwise my curve is the same as yours.
As far as noise floor when measuring, I live in an apartment with single pane windows and thin walls right next to a busy street. I have done measurements with plenty of background noise and I have done measurements with no background noise and the measurements were almost identical. I think REW does a good job at filtering out background noise.
I think you're right. I might be going overboard to try and lower my noise floor, but REW does a great job.
Midrange compensation is related to the change in directionality as the speaker transitions from the tweeter to the mid. Basically, if the speaker design has an intentional dip of about 3-4dB at the speaker's crossover point between those two drivers, you don't want to EQ that out because doing so can make the speaker sound more harsh. So Audyssey's MRC is centered at 2kHz, which is where they believe generally most speakers will transition, and where directionality issues would generally occur.
Here's the thing though: If you don't see a dip there in your measurements, there's likely no reason to use MRC. Most well-designed speakers won't need it. In your case, the Elac Uni-Fi uses a concentric design where the tweeter is in the middle of the mid, which minimizes directionality issues. Your Uni-Fis transition at 2.7kHz, and I don't see any dip in that region in your measurements, so there is absolutely no circumstance where I would ever use MRC with those speakers. Even if you wanted to, Audyssey's stock MRC isn't centered at the right frequency for your particular speaker.
However, for those with other speakers, if you see about a 3-4dB dip around 2kHz, there may be value in enabling MRC in MultEQ-X for that speaker. If you see a dip anywhere else in that general area, look up the specs for your speaker and see where the crossover point is for that speaker's tweeter-to-mid transition. If the dip in your measurements is centered at about that same frequency, a custom MRC filter may be a good idea. You can do this by creating a parametric peaking filter with a Q of about 2.76 at the crossover frequency at about the same dip you're seeing in your measurements. So, for instance, if you see a 4dB drop at 2.5kHz on a speaker whose specs show that 2.5kHz is the crossover point, you would create a filter in MultEQ-X at 2.5kHz, Q of 2.76, gain of -4dB. You can tweak this to shape it more like the dip you have, but this will keep any other EQ work you're doing from trying to push that intentional dip out.
Caveat: Ideally, you would need to see spinorama data for your speaker to see what the directionality index at your speaker's transition point looks like to know if you really NEED some form of MRC. But generally, if you measure a dip at the speaker's tweeter-to-mid transition, EQ'ing that out could make things sound worse. If things sound harsh to you, try MRC.
Thanks for all your content! I have followed many of your videos, and learn something new every time.
Is there any reason not to use EQ functionality and "Generate measurement from filters" instead of trace arithmetic? It would presumably save some effort on manually editing the filters after export. Looks like the exported filters are more detailed when generated by EQ, but it didn't prevent import target in MultEQ-X for me.
Also, what are your thoughts whether or not you are correcting for reflections in the high frequencies? I see OCA is using some different IR windows for the highest frequencies, and I know you have mentioned his channel on many occasions.
I haven't been introduced to that method of Generate measurement from filters. Can you export as text? I see OCA doing more in high frequencies, but OCA doesn't actually correction above 300hz, so it really doesn't matter. I like the way Psychoacoustic sounds, and I think it gives me the best tonality. I think you need to be an expert to draw any more beyond Psychoacoustic smoothing.
What's your opinion on not touching frequencies above the Schroedinger frequency? This is a topic often addressed in the context of room correction. In OCA's latest video, he talks about this as well.
I saw it, and I disagree with it. I think EQ-ing all the way sounds best. But it's a great video.
REW recommends EQ'ing with variable smoothing, what is your reasoning for using psychoacoustic smoothing for EQ?
I saw that too. I find the bass detail way too high. Where psychoacoustic is much smoother and I think realistic to the way we hear.
@@hdmoviesource I can see why using VAR makes sense however, since octaves are a logarithmic scale, the lower the frequency the more notes are within a given range. E.g. from 20hz to 40hz is an octave, but also from 1000hz to 2000hz is an octave, so 20hz range for an octave at low frequencies vs 1000hz range for an octave at higher frequencies. Essentially there IS more detail in lower frequencies, so treating them more finely seems to make sense.
Yeah, I think it depends on how much control you want over bass frequencies. I'm finding that Psychoacoustic gives me the sound that I'm after. So, as long as you can get the sound that you're after with the smoothing that you're using that's all that matters.
Hi Carl…
Please make a video on Speaker equalisation using Dirac ( standalone) Computer version if you have license or using a trial version. It would be immensely helpful to Dirac users ….
I may upgrade to Dirac soon.
Quick question, are you measuring from multiple seats and using cross correlation and then vector average to eq? I saw a video showing this from oca a while ago and have been adding this to my eq process.
From a single seat. I used the cross-correlation method and it didn't sound good in my opinion. I find that single seat sounds best even for other seats.
@@hdmoviesource maybe a single seat recal is on the books for tonight 🤪👍🙏💯
LOL, let me know how it goes.
Where was your measurement mic in this?
You appear to be occupying the MLP was the mic behind you or in front of you?
Correct, all at MLP.
do i get it right, that you do this method by using only ONE Target Curve (reference speaker FR) for ALL speakers; how does this work, if every speaker is 'it's own Target Curve' as far i understood!? Wouldn't it be necessary to have a target curve for each single speaker in my room!?
Yes, this is true. The point of this was to make sure all my speakers sound a particular way. So they tonally match.
Thank you so much for this video!!! I'm using DIRAC on my 2.1 system and I finally got the sound I wanted after calibration. At first I was simply following the default target curves that DIRAC provides, but with those the sound felt very two dimensional. After watching your video, I traced a new response curve through the original, untreated response of the speakers in the room and used that as a target curve in DIRAC. It did exactly what I wanted, kept the original sound and coloration of the speakers, but removed most of the imperfections and tightened down the sound. The curve is nothing close to what the default curves in DIRAC are, but I love the sound it produces.
Excellent. That's awesome, and good to hear.
What is the best way to place you speakers? I put them in a circle around me so every speaker is the same distance.
That sounds like a good idea to me Adam. Also, think about degrees and angles. I like to angle my speakers 15 degrees off-axis away from the MLP. That's the smoothest response that my speaker gives me. I may have to recheck that, but, always get each speaker to be the same degrees off-axis whichever angle you choose is something to think about.
@@hdmoviesource interesting, I’ll have to try that. That is closer than I have mine set up. Mine is about 45 degrees. I have recently rearranged according to the sound nodes. It made a bigger difference than I thought, especially with the bass.
Interesting. I find directivity to sound better like this. I've pointed the speakers directly at me once and I had no idea where sounds were coming from. LOL, so it can really make a huge difference.@@adamATOM3
@@hdmoviesource sorry I misunderstood that. I meant from MLP my speakers are in 45 degree triangle. You mean where you point your speakers? You point them a little wide of the MLP you mean?
Yes, I point them a little wide of my MLP. I've found that it's really opened up the soundstage and I can actually hear when things are coming from now.
Are you on the Magic Beans Audio beta? I'm curious to see an update on this video after using Magic Beans
I am, I just haven't had the time to use it Darren.
This makes me think, this is what Magic Beans does.
I believe magic beans is correcting for nearfield with midrange and treble, and bass at the Main Listening Position. I've heard it sounds really good.
Very long video. It is full of good info, but it was putting me to sleep.
I understand, at least you know that I can make you fall asleep Hitechrr. You should see my other videos, haha.
I'm subscribed.
Thank you.@@hitechrr
Thank you for the tutorial. It would have been more valuable though, if you could present your information in a more structured and concise way.😊
I appreciate the feedback Rairun.
my recommendation is not using a mic, or looking at graphs and that means no targets.
sit down in listening position, do real time eq by ear so it sound evenly loud to you, you can call that flat if you will.
pick a loudness that you most often use and that is it. require no tools and experience.
and takes into account room, speakers, hearing, audio taste, loudness curve/house curve.
Do you use bass and treble controls?
@@hdmoviesource i use parametric equalization
Okay.