Experienced only: What have you done with room correction?

Very nice detailed explanation of frequency-dependent sound behaviour at SynAudCon here.

So-called room EQ is fascinating I think. First it helps to understand the behaviour of sound in your your room in terms of the broad division of frequency bands into specular, transitional and modal. Room EQ certainly helps with tonality across the frequency spectrum. You also want to know the reverberation behaviour of your room EQ can’t help much with that.

My room is quite dry in acoustic terms due to its construction (a Japanese style timber house) so I’m good to go with room EQ. My RT60 (an approximation, because small room behaviour doesn’t give us a diffuse field) is around 200 ms until the bottom octave where it increases to 300 ms. If you live in a concrete apartment block, those figures will be significantly higher and you’ll have to add absorption first.

I model my room in REW and Amcoustics room simulators to get a handle on Schroeder frequency (which divides specular and modal) room modes and influence of loudspeaker position. When I bought new speakers a year or so back I positioned them, measured FR at listening position, moved the speakers in 200 mm steps laterally then longitudinally, measuring at each step until the most problematic nulls were minimised. You can’t fix severe nulls with EQ so this comes first.

I use a Mac as source running Apple Music (now with hi-res and spatial audio aka Atmos) and Sonarworks software to EQ, then USB over Thunderbolt to the DAC. I like Sonarworks because it measures ~37 locations spread around the listening position to get a better model of room behaviour before running EQ. I used their mic in a Røde shock mount pistol grip handle and their sonar location method is completely fun.

My previous speakers (Audio Physic Tempo 25) were a bit warm in the mid-bass but otherwise pretty good FR-wise. Running Sonarworks EQ full-range along with the B&K 1974 curve improved mid-bass FR smoothness (expected) and resulted in even better delineation of the stereo image (which I didn’t expect) so I was very happy. AP speakers have great timbre and EQ didn’t harm that at all.

My new speakers (Audio Physic Codex) are close enough to full range so bass issues that didn’t arise before presented themselves (reinforcement at 25 Hz being the rooms lowest lontitudinal mode was beneficial but a 15 dB bump at 50 Hz from the room’s second longitudinal mode was deleterious). EQ fixes errant peaks unproblematically. Those speakers were also a bit shy in the mid bass (opposite to their smaller brethren) as room modes in the transitional zone (250-700 Hz in my case) had a negative influence and EQ coped quite ok with that, balancing mid-bass with midrange nicely. EQ also provides a bit more lower bass extension (but don’t go overboard as high levels can damage speakers).

In my experience, even with excellent speakers, careful positioning followed by judicious application of DSP has no real downside. With the smaller speakers, bass distortion (I use Fuzzmeasure for analysis) rises but doesn’t sound too bad, but headroom is an issue then. With the larger speakers, bass distortion is insignificant and higher levels are comfortable.

Btw I’m aware Acourate and Audiolense are very cool but I haven’t gone there as running things in a Windows VM doesn’t seem with the effort. If you use a Windows PC source, by all means try them.

@burtlake

Interesting discussion and the tech is tempting to try out. One question; doesn’t the DSP do it’s own A/D and D/A conversion? After spending a relatively large sum on a DAC, it seems counterproductive to have the signal reconverted twice more downstream by a much lower cost device… am I off-base here?

@erik_squires

Depends really on where you do the DSP. If you use Roon for instance, you do it all before the streamer gets the data. If you use a DSP capable streamer/DAC then no difference. If you use it only for the subwoofer (like I often do) then it’s out of the way of the main DAC/Amp chain

Running DSP software on a Mac/PC source is similar to using a streamer (which is just a dedicated computer) as a source, everything is done in the digital domain upstream of your DAC.

The current line of miniDSP boxes running Dirac (like minDSP Flex and SHD) take digital in (USB/Toslink/SPDIF coax on Flex, plus Ethernet and AES-EBU on SHD) and do the DAC job themselves. Because you send them a digital signal, there’s no redundant ADC/DAC conversion stages. There’s also a studio model of their SHD which sends digital to a downstream DAC. The SHD models use a 96 kHz internal sample rate, so a Mac/PC can theoretically do better (Dirac and Sonarworks say 192 kHz there). The Flex uses 96 kHz also but says 48 kHz with Dirac licence, that may be across the board for their stuff?

I was using Tidal and thought about MQA but that makes DSP hard, fortunately Apple decided to do hi-res (and multichannel Atmos) on Mac around that time which made for a very simple setup (and I was happy to entirely avoid the AVP/AVR world). But consequently I can’t say anything useful about sonics via the miniDSP devices.