- Joined
- Jan 15, 2020
- Messages
- 6,873
- Likes
- 16,839
In the often occurring compromise smooth angle curves vs. smooth smooth power to opt more for the first one.
Last edited:
-
WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!
Best Room Response
- Thread starter JR4321
- Start date
wgb113
Active Member
This is the viewpoint I’ve come around to over the past several weeks experimenting with Dirac. My initial experience like the OP with full range correction was of a worse sound. For lack of better adjectives I’ll resort to using the typical HiFi terms where the space, air and detail in the mids and highs seemed to condense to a flat, thin, unnatural soundstage.
Many recommended partial correction up through the Schroeder frequency which, for my small room is 280Hz. This helped and initially I thought all was fixed. The past couple of weeks however I’ve spent more time listening through headphones and, while the soundstage doesn’t exist like it does with speakers, the imaging, space, air, detail was all there. When I switch back to speakers with Dirac off they sound more like my headphones in those areas than when Dirac is running.
Part of the allure of Dirac for me was it’s ability to tackle full-range correction as icing on the cake so to speak. While my room is small it’s dedicated so both speaker and listening positions are flexible and room treatment has been installed to help with bass issues. But not being able to utilize Dirac full range reduces a key benefit of the software - its ability to correct in the time domain which primarily is a factor above the transition frequency.
I’ve decided to dial it back a bit. I’ll likely be selling off the miniDSP and Dirac license and using a combination of REW filters with Roon to help below 300Hz. I’m going experiment without a subwoofer as well, giving up that last octave for an overall clearer sound.
Your system already has a fair but of potential equalization. Your RME ADI-2 has tone control and filters and your sub has dsp room eq. Between them you should be able to do a fair bit of adjustment. Do you have a UMIK-1 to measure in REW?
wgb113
Active Member
I do. The biggest "con" for me getting rid of Dirac is losing correction when playing vinyl. To hear those albums with correction I'd have to live with the needle-drops.
I'm using a miniDSP SHD to integrate 2.2 system (2 main speakers, 2 subwoofers) and have found that doing correction in 3 stages works best.
Stage 1 - Measure separately and work out best crossover - 12db? octave low pass and 24db/Octave in my case).
Stage 2 - Equalise main and subwoofer separately using REW with PEQ filters imported into SHD
Stage 3 - Apply Dirac (with crossover and PEQ filters active).
I've found that this works much better than doing just stages 1 and 3 - doing that seems to suck the life out of the system and narrow the image. Also, foing all 3 stages measures better.
I think that DDRC-24 has 4 PEQ filters per channel (SHD has 10) but might be worth trying before you abandon Dirac?
dasdoing
Major Contributor
without reading the topic at all, just responding OP:
fast food tastes impresive whe you prove it, but it's an ilusion. you will lose the sense for real taste because those artificial favours cofuse youre taste.
the same way boomy bass impress a lot of people, but it will prevent you from clearly hearing all the other frequencies
fast food tastes impresive whe you prove it, but it's an ilusion. you will lose the sense for real taste because those artificial favours cofuse youre taste.
the same way boomy bass impress a lot of people, but it will prevent you from clearly hearing all the other frequencies
My personal experience is that "room curve" is dependent on the particular speaker in the particular room, ergo no specific curve more suitable than others as a general rule. More controlled directive speakers will naturally have a flatter in-room response and will sound broken if you apply a -10 dB room curve to it. Whatever curve you end up with over the transition area of the room if you measure with MMM technique (or spatial averaging of many different measurement points) with a neutral speaker is the best room curve.
If you have a speaker with lots of problems, a room target is not the way to fix it because you'll likely end up with stupid relationship between direct sound and reflective sound even if it looks good on the measurements in the listening position. Buy better speakers!
If not, you should use the spinorama to fix the speaker with speaker EQ to the extent it's possible. Still sounds bad? Buy better speakers!
When it comes to Dirac my experience is that it sounds a whole lot better if you measure with the couch/sofa settings and use proper spacing in both height and width between measurement points. And measure all 9 points!
This might be different if you have no walls very close to the listening position or speakers, but I find it's a remarkable difference in sound quality compared to one-point or very closely measured points.
If you are lazy with Dirac, it might end up looking like this;
Yep, looks mighty impressive. Sounds awful. Why? Because what you measure and correct for in one point in space is not at all relevant to how the sound field looks or how our ears perceive sound. Example is with Kii Three speakers.
You know that overly smoothed, non-dynamic glassy room EQ sound? Yes, that one. That sound is the result of way too aggressive EQ and in extreme cases it looks like the red curve above.
Here's my onion;
* Measure at many points with decent spacing in both horizontal and vertical space.
* The curves you're presented with will show a clear trend in the room curve. Use that as a basis.
* The room correction will remove quite a bit of energy from the room modes and will sound lean. Compensate by raising the overall bass level.
If you look at the blue curve above you'll see a neutral speaker in a (very) reflective room ending up with something very closely resembling the infamous "Harman curve"
Focus on the difference between the range below 60 and to160 hz and use that as a rough guide of where you should be when you've removed the room modes and their energy. 4-5 dB up from 160 ish can be okay, but keep in mind that this will eat capacity from your speakers.
You thought your little 6 inch woofers were bad-ass? Think again. They have been carried along by the room's awesome modal powers and you've been fooled all this time!
* Oh, do try to measure with and without pillows in place in the sofa/chair just to see how much effect that stuff has around 1-2 khz on the measurements. We don't hear the sofa when we're listening, so why would we EQ it? And keep in mind that a lot of the ripples above 1 khz may very well be from the microphone stand, just ask Amir about that if you don't believe that the ripples between 3-6 khz on my measurements came from a stupid placement of the stand.
If you have a speaker with lots of problems, a room target is not the way to fix it because you'll likely end up with stupid relationship between direct sound and reflective sound even if it looks good on the measurements in the listening position. Buy better speakers!
If not, you should use the spinorama to fix the speaker with speaker EQ to the extent it's possible. Still sounds bad? Buy better speakers!
When it comes to Dirac my experience is that it sounds a whole lot better if you measure with the couch/sofa settings and use proper spacing in both height and width between measurement points. And measure all 9 points!
This might be different if you have no walls very close to the listening position or speakers, but I find it's a remarkable difference in sound quality compared to one-point or very closely measured points.
If you are lazy with Dirac, it might end up looking like this;
Yep, looks mighty impressive. Sounds awful. Why? Because what you measure and correct for in one point in space is not at all relevant to how the sound field looks or how our ears perceive sound. Example is with Kii Three speakers.
You know that overly smoothed, non-dynamic glassy room EQ sound? Yes, that one. That sound is the result of way too aggressive EQ and in extreme cases it looks like the red curve above.
Here's my onion;
* Measure at many points with decent spacing in both horizontal and vertical space.
* The curves you're presented with will show a clear trend in the room curve. Use that as a basis.
* The room correction will remove quite a bit of energy from the room modes and will sound lean. Compensate by raising the overall bass level.
If you look at the blue curve above you'll see a neutral speaker in a (very) reflective room ending up with something very closely resembling the infamous "Harman curve"
Focus on the difference between the range below 60 and to160 hz and use that as a rough guide of where you should be when you've removed the room modes and their energy. 4-5 dB up from 160 ish can be okay, but keep in mind that this will eat capacity from your speakers.
You thought your little 6 inch woofers were bad-ass? Think again. They have been carried along by the room's awesome modal powers and you've been fooled all this time!
* Oh, do try to measure with and without pillows in place in the sofa/chair just to see how much effect that stuff has around 1-2 khz on the measurements. We don't hear the sofa when we're listening, so why would we EQ it? And keep in mind that a lot of the ripples above 1 khz may very well be from the microphone stand, just ask Amir about that if you don't believe that the ripples between 3-6 khz on my measurements came from a stupid placement of the stand.
I know I'm reviving an old thread but it's the exact question I've been wondering regarding the range after the transition frequency and before about 1000Hz. Up to this point I correct the room under about 200Hz and above that I use Anechoic measurements. You can see above 200Hz the room is still affecting things quite a bit but I haven't seen any kind of study showing whether we should be EQ'ing this region or not. My own personal testing has led me to leave this region alone and only make sure I EQ with proper CTA-2034 style anechoic measurements. When I EQ these peaks from 200-800Hz the sound isn't as natural and has a thinness to the mids. I think good arguments and logic were made on both sides of this argument but I would recommend doing some comparisons when EQ'ing in this range. This is an MMM measurement with periodic pink noise so should be flattish.
dasdoing
Major Contributor
one suggestion: go to https://www.szynalski.com/tone-generator/ and play a A4 (440Hz) and than a G4 (392Hz). Are you hearing a loudness diference?
Ok I wasn't home all week so just tried it out using REW's generator, I used 285 and 370 as they were close together and the largest peak/dip combo. It "maybe" sounded a bit louder but it's really hard to tell since they are different frequencies. I then checked both frequencies with the RTA function and my Umik and the peaks were just about the same, this seems a bit odd since my measurement above is done the same way just using periodic pink noise instead of the tone generator.
bennybbbx
Addicted to Fun and Learning
measuring is only good for tendency you have in the room. what you hear is the correct thing because ears have a directivity and are not omnidirectional as a measure microphone. so ears filters room reflections at some directions less or more. it is also important that when testtones are play, they are in the middle hear. use much eq bands with high Q reduce clarity and stereo image (worse transients).
so try some EQ corrections and hear which sound better. dont trust measures too much
Similar threads
- Poll
