Search the Community

Measurements like statistics in politics can be manipulated or even ill used with best intentions. In this thread I intend to go under the hood showing the options I chosen along the way while creating convolution filters. So maybe I'll expose my goofs and you're welcome to comment them. Or maybe this will enlighten your own choices and help you take better informed decisions.

Time for some Spring cleaning in the Loft! Here, through a series of photos and screenshots, I'll chronicle temporary reworking of this room this past weekend for a final listen to the DALI Oberon 7 towers before I ship them back to Lenbrook. You'll laugh. You'll cry. It will be fun for the whole family. Step one: add some loudspeakers. Rough positioning done, based on past measurements. These are designed to fire straight down the length of the room. No toe-in Step two: Left and right distance matched to the room's center-line within 1/16th inch. I always have to pull the left speaker 1/2 inch or so forward. Walls are not true in most houses. Step three: Orchard Audio monoblocks connected and powered up. Mic in place. Ready for acoustic measurements. Note, I'm not using the REL T7i subs in this calibration. Step four: Starting the project with REW to understand our starting point. (the big green screen is tacked to wall-mounted acoustic panels. It's just there because of the shelter-at-home thing. Makes my Zoom videos a little more engaging) Step five: Review un-smoothed response at the listening position below the transition frequency. Welcome to small-room acoustics! 😕 Still, not terrible. I put the -3 dB point in-room at close to 30 Hz. The peak at 112 Hz can use some tamping down, but otherwise, very workable. The suckouts are due to room geometry (15.5 x 10.1 x 8 ft). Could move them around a bit by changing the mic position, and shifting the speakers a little, and I probably would if this was not a temporary setup. Step six: Step response is not the worst I've seen. This should be a lot better with some time-domain correction. Step seven: Distortion at measurement level (~75 dB) are at or below the noise floor. Low distortion is a feature DALI loudspeakers are famous for. Step eight: Reverberation is well-sorted in this room. 300ms is a good target for a room this size. There's probably more I could do with treatments to improve articulation at ~250 Hz, but I'm happy with this overall. Step nine: On the waterfall, that ridge at 110 Hz is going to be fun to work with. I don't recall seeing this as much with other loudspeakers in this room, so this could be a "feature" of the Oberon series' voicing...that punchy mid-bass. Step ten: Now, we get to work. This is not a bad place to start. We can see the late arrival of output from the dual 6.5" woofers in the step response graph at the bottom. Step eleven: After a bit of work with filter design, here's the predicted response after correction. The step response is looking much better (time-domain corrections are amazing). I decided not to correct the dips/suckouts, so I'm only tamping down the peaks. Amplitude chart here is unsmoothed. Step twelve: Let's now go back to REW and check our work by seeing if the measured response matches what Acourate predicted. Here, REW is playing the sweeps into a virtual sound card presented by AcourateConvolver, which is applying the FIR correction filters in real-time. REW then records the corrected response. The suckouts are still there. It's rarely a good idea to EQ these out. Attempting to do so just increases distortion, and our ears are not very sensitive to them anyway. But, the peaks are nicely tamped down in the amplitude response. Oh, and note that our -3 dB point is now easily below 27 Hz. That gets us the lowest note on a grand piano. There's also now usable sub-bass down to just a shade over 20 Hz without adding significantly to distortion. This is full-range response from a $1.5k pair of towers. Not bad. Step thirteen: The step response looks much better. This is a difference you can really hear. Transients, soundstage, localization, everything gets better when timing is good. Step fourteen: Looking at the distortion plot is a good place to to spot over-corrections. THD is below the noise floor above 26 Hz. Our ears are not sensitive to low-level distortion at low frequencies. We can see that, overall, distortion levels are lower after correction. Step fifteen: In the waterfall plot, we still have a bit of an unsightly ridge at ~90 Hz, but life looks much better at 110 Hz. Overall, bass should be very articulate and resolving with this correction. Step sixteen: Filters are loaded into a DSP Preset in Roon. Note that the REL T7i subs are not part of this correction. I'll leave them off for my evaluation, but if the Oberon 7's were sticking around, I'd definitely take time to integrate the REL's. I have yet to encounter a loudspeaker that could not benefit from a good sub-bass system. BTW, I was able to verify the corrections through Roon's convolution by saving the REW logsweep to a WAV file and then playing it back through Roon while measuring in REW. Amplitude response was identical. Step response was slightly different, but that could be due to the technique I used to play the files in Roon while trying to record in REW. Step seventeen: One last plot to illustrate a point. In this room, I usually do full-range corrections, with relatively large filter windows below the transition frequency so that lows get precise correction. Corrections to the highs are very gentle. However, even so, it's easy to end up over-correcting if you're not careful. The Fritz Carrera 7 BE loudspeakers that I usually have in here work best with a fair amount of toe-in, and I don't even have grills for them. Per the manufacturer, the DALI's want to fire straight down the length of the room with their grills on. I probably should have removed the grills for my measurements, but since I did not, I found that I had to limit the band of my amplitude corrections to 11 kHz on the top end. Frequencies above that were getting over-corrected (boosted, effectively), which is not good. Lesson learned. Here's a plot that shows how the response drops off fairly substantially above 11 kHz. I'm not sure if this is due to the grills or lack of toe-in, but they sound more balanced now with frequencies above 11 kHz (later, lowered to 6 kHz) not corrected. I'm always learning new things about this stuff. In case it's not obvious, the top lines are the amplitude response of both channels after correction relative to the target. The bottom lines show the actual inverted amplitude (correction curve) being applied via convolution. Step eighteen: Proof is in the pudding. Here's a little before/after comparison. It's subtle, even on good headphones. The effect is significant in the room and well worth the trouble. The final test: My wife and I had a great listening session last night after I finished my tweaking. I brought her up to the loft to get her opinion on the sound. She was reluctant since she didn't love the sound of the DALI Oberon 7's the last time I had them in the system (without correction). She was blown away by the sound. What started out as just spot checking with a couple of tracks expanded into several hours of exploration of her favorite recordings, including a complete piano concerto (Rach 3) that went on past midnight. 🙂 It's always nice when a bit of setup work and digital room correction pays off with results that are obvious to others. A bit about me: I've been plinking around with Acourate for 3+ years with lots of help from @mitchco's book; however, a few weeks ago, I took advantage of his April sale so that we could work together on a fresh approach to Acourate and DRC for my loft. I learned a ton through the experience, so setting up the DALI's was a fun way to test my what I learned on a different system. Feel free to share comments, suggestions, or similar experiences below. Thanks for having a look, and happy listening. -- David

Hi, I wonder if someone could provide a step-by-step guide to achieving room correction using REW and HQ Player. I tried to find a tutorial but the information is scarce and spread all over the web... It's been years since I last used REW to measure the frequency response at the listening spot. I have a mic, REW and I'm ready to go. Thanks, Ricardo

Goal /the incumbent filters: get a faster more “agile” articulate intelligible sound while keeping big bass extension Crisis started when I switched off convolution to compare my setup + original Argo of Jamal at Alambra to the needle drop sold as dsd download (yes yes and to much praise by Acousticsounds ; meaningless comparison since I doubt the needle drop is sourced from the distorting but sometimes gloriously tubey original) : though convolved DSD was a clear winner over non convolved DSD or LP, it sounded too analog and too tamed, fat and slow. Findings Do eQ, REW is donations welcome free : try it (but you need a microphone). Doubt you get comparable results from power usb whathaveyou cables… RR1 (as in https://www.innerfidelity.com/content/warren-tenbrooks-summary-head-measurements-harman) is a mighty good target house curve Mind applied smoothing : a predicted response curve matching Bob Katz’s target (flat to 1 k, 6 dB downward to 20K ) created with “Psyschoacoustic” smoothing in REW resembles RR1 when looked at in VAR or traditional (ie 1/3 octave) smoothing mode (Bob Katz uses Acourate’s Psychoacoustic) Multiply measurements but pick the one that resembles average, not the average summation (ie in REW the Psychoacoustic profiling only applies to single measurements ; as do many options) Avoid messing with frequencies above Minimum Phase zones galore (on the left half of frequencies) (850 Hz max here, different from the theoretical 300). Lower Shelves are great : found out that RR1 paralleled the measured response in my room and that a 5 dB 600 Hz start LS puts frequencies below 600 on the RR1 target Dare switch to Manual. Automatic was not at ease, I took the time to flatten the bumps and, up to 9 dB in Minimum Phase zones, fill. Listen (Stereophile CD3 and best of Chesky 2 have been most helpful, mainly the tracks with moving point sources (acoustic clickers, cowbell…) ). I did not have at hand the mike I created the measurements with, only a iPhone + SPL app, useless below 250 and probably giving a false 3 dB bump reading at 500, but measuring would be great too. In the end listening matters more. Bottom line :girlfriend threatened to quit madman but is as amazed as I am in the end : WOW ! Goal met

Mitch Barnett of Accurate Sound tells us about his journey to become a leading expert in Digital Sound Processing (DSP). Accurate Sound is a remote service using Audiolense, Acourate and Room Eq Wizard (REW).

Have you ever heard of Equalizer APO? It is really great, it is fully costumizable parametric equalizer with unlimited numbers and OPEN SOURCE It is compatible with Windows (Dunno if Vista and 7 is Supported, But 8 and 8.1 works) If you have Windows 8 and then version 8 and under is supported for Windows 8.1 is version 8.1 and upwards (something with changed APO thingy) You can use Room Equalizer Wizard to create your Equalizer for you (Read the instructions in the Equalizer APO Documentation) Note the higher end of equalizer (10khz and up, is the dB timed up to 2 times, like increase 1dB will result in 2db) The not great is, it is used in WIndows APO, note WASAPI or ASIO likewise will not work, it is directly streamed and bypasses the APO. You can find the program at this link: https://sourceforge.net/projects/equalizerapo/

Here are some itching funnies (and ready made for REW targets) I’d like to share. First, Let’s take the same set of Impulse Responses obtained by Sweeps and average* them 3 ways in REW : Frequency Dependent Windowed RMS Average, RMS Average, FDW Vector Average. I add results from Real Time Analysis of Pink Noise measured while Moving Microphone. There have been some changes between MMM and Sweeps times but still there’s a good consistency between RMS Av and MMM results while FDW RMS Av and FDW VA yield radically different results, in my room with my speakers. Hence, you’d better know what you (or the expert and/or commercial software you would trust) are/is to pick. If I am to obtain textbook Impulse and Step Response (thanks to RePhase) I have to pick the FDW VA. If I set Phase with the FDW VA and then set Amplitude on RMS Av or MMM, Impulse and Step remain OK but not as pretty. This raises the question of what but the VA is a commercial software picking if it is to display textbook Phase related results. And that conflicts, unless What You See is Not What You Get, with the choices of targets (neither B&K nor Harman dubbed them such) such as RR1 or B&K. RR1 low end extends B&K that dates back to Pink Noise and RTA time, so no Impulse Response, no FDW, it’s Steady State that is measured… So, if one is to choose a target related to Steady State, one should better apply it on Steady State measurement… *L and R are in different environments in my room and hence measure differently below 300 but for the sake of simplicity they are here added/averaged