from Hacker News

Show HN: HiFiScan, a Python app to optimize your loudspeakers

by erdewit on 9/11/22, 11:59 AM with 103 comments

• by IAmGraydon on 9/11/22, 2:51 PM

  I'm not an audiophile in the obsessive-compulsive sense, but I've been recording music in my home studio for 20 years and I know my way around it. This sort of calibration is not ideal. Not only are you measuring with a device that has an imperfect response curve, but you are also measuring the room at a single monophonic point in space. The way that sound interacts with the room and your ears is far more complex than that. Ultimately, this is a bandaid for a poorly treated room. If you're serious about getting a flat response curve from your monitoring room, you're far better off learning how to treat the room properly and how to position your monitors within the room for the best results.

• by timc3 on 9/11/22, 8:04 PM

  This type of software, is just a bandaid and really doesn't work very well (though it can work better with headphones). Properly thought out and tuned acoustic work is what is needed.

  I am lucky enough to have a spare room in my house, and set out to build a studio (an almost life-long dream) and decided that I didn't want to compromise on the acoustics and spent some time looking into the subject. In the end I built it myself with a huge amount of acoustic treatment (lost a large amount of the volume room), but more that that I enlisted the help of a professional who could do the maths and help with not just the trapping but also the panels that are needed. In the end after I built it was also tuned with DSP by the professional, has what you would normally call 4-way speakers with the subwoofers going to a higher frequency than most would consider normal and even the desk was specifically chosen to not cause a problem for the listening environment. The difference between this and something like Sonarworks (commercial software that I tried for a laugh beforehand) cannot be overstated. It's basically flat between 23hz (slightly rises at 20hz I believe) and 20Khz - we actually tuned in a more natural response curve.

  It's still a home studio because it's in my home and I don't do anything commercial with it, but it's pretty much mastering grade, all with materials that are available in a builders yard and the special sauce, someone that knew what they are doing. Not everyone has the room or space to do this, but most people can build some bass traps and something to tame first point reflections.

• by patrakov on 9/11/22, 3:10 PM

  There is an older project with better math inside: http://drc-fir.sourceforge.net/

  For starters, it doesn't try to achieve a phase-neutral response, because a phase-neutral response created in a room is only valid in one point of the room, and creates pre-echo artifacts elsewhere. In fact, it tries to separate the response of the speaker itself from the response of the room, by setting a threshold in the time domain, so that everything coming before it must be unaffected by the room. Then, everything coming before the threshold is corrected to a linear phase, while everything else is corrected to the minimum phase (thus making the second part of the filter purely causal).

  Also, they provide an argument, citing literature, that equalizing to a flat frequency response would be wrong in a room, and thus provide an option to remove excessive treble and achieve a 1dB/octave roll-off.

  Please see the details at http://drc-fir.sourceforge.net/doc/drc.html

• by qbonnard on 9/11/22, 1:20 PM

  Newbie question: how do we know we can trust the microphone?

  It sounds like a chicken-and-egg problem to equalize speakers with an equalized microphone, but maybe microphones are simpler and can be assumed to be equalized ?

• by zihotki on 9/11/22, 12:58 PM

  I wonder what are the differences between this tool and industry standard REW app - https://www.roomeqwizard.com/

• by hedgehog on 9/11/22, 3:02 PM

  This looks cool. I'm not sure if they are intending to go all the way to room correction but it can really do wonders. A good while back my music setup used filters calculated by an open source FIR tool with playback driven by an older version of Shairport (emulating an AirPort express) using BruteFIR as a convolver. Fiddly to set up but it sounded really good.

  1. http://drc-fir.sourceforge.net

  2. https://github.com/mikebrady/shairport-sync

  3. https://torger.se/anders/brutefir.html

• by ttpphd on 9/11/22, 4:02 PM

  I'm a psychoacoustician and this is not the way, very sorry to report. Others have touched on the acoustic issues already, so let me touch on the psychological ones: your perception of sound from loudspeakers doesn't just depend on the acoustic waves hitting your ears. It also depends on your personality and expectations. If you genuinely believe that doing a seance to drive out the poltergeist from your speaker set up will make the sound better, it will be difficult to convince you otherwise precisely because the acoustics did not actually perceptibly change.

• by strainer on 9/11/22, 4:51 PM

  I have made a small webtool to help calibrate various EQs by ear. It kind-of mimics a graphic EQ in the browser which can also play tones around the EQs frequency bands, which should sound about the same loudness as their neighbors according to the ISO loudness curve. I increase or decrease my laptops EQ bands until the tones on the webtool play without obvious difference. This is sure to be an unsatisfactory process for technical purposes, and I couldn't even guarantee that I implemented the loudness curve well, but I have a lot more success using it to help tune EQ than without it.

  https://strainer.github.io/hearqualizer/

• by TacticalCoder on 9/11/22, 2:50 PM

  Isn't it basically what "DRC" does? (Digital Room Correction)

  https://en.wikipedia.org/wiki/Digital_room_correction

  I don't remember the exact order but way, way, way before the $10 K USD digital audio cable snake oil, audiophiles are going to say that DRC is the second single biggest thing that can enhance the quality of your setup (the first one being which speakers you're using and how you place them). Then source quality/amp/dac. And only way further down the line, for those who believe in voodoo, $10 K digital audio cables.

• by runeks on 9/11/22, 4:26 PM

  I’ve tried this for my speaker setup. And the problem is that the frequency response is a function of volume. For example, the louder I play music the more the bass is accentuated. I think this is because of standing waves.

  So the problem I find is that when the volume is low the bass is too low, and when the volume is high the bass is too loud. Only when I play at the same volume as the equalization was performed at do I get a good result.

• by solardev on 9/11/22, 5:30 PM

  Asking because I'm not smart enough: Is this kinda similar to what the Sonos Trueplay feature does? (Where you move your phone, and/or a mic-enabled speaker itself, around the room so that it plays and measures various frequencies to calibrate)

  https://support.sonos.com/s/article/3251?language=en_US

  https://patents.google.com/patent/EP3531714A2/

• by rcarmo on 9/11/22, 1:51 PM

  This is very nice. I also appreciate the pointers to various equalizer apps in the README, I didn’t know a couple of them.

• by anotheryou on 9/11/22, 2:27 PM

  I'm using the commercial https://www.sonarworks.com/soundid-reference and it's amazing.

  I'd say the worse your setup (especially your room) the more magic it does.

  I did it without an individually calibrated mic though (but with a decent measuring one), wonder how much better it could be.

• by 127 on 9/15/22, 8:05 AM

  So many negative comments here. I for one think this is absolutely fantastic tool to calibrate cheap television and computer speakers so they don't sound like complete crap. We're talking about almost 20dB variations on the spectrum here.

• by tomduncalf on 9/11/22, 5:48 PM

  Cool project! I recently bought a set of iLoud MTM monitor speakers which come with a special mic which they use to analyse the room and correct for it in a similar way to this.

  It makes a good difference to the sound - highly recommend the speakers if you are looking for a smallish set of monitor speakers that sound great and can be used very near field so you can use lower volumes.

• by LeSaucy on 9/11/22, 2:21 PM

  How does this compare to dirac?

• by etaioinshrdlu on 9/11/22, 9:36 PM

  I think it would be cool to make a more advanced version that corrects for many types of nonlinearities: amplifier distortion and mechanical parts resonating badly.

• by tambourine_man on 9/12/22, 3:21 AM

  > A good microphone is needed

  What qualifies as good enough?

  Is it worth trying with consumer mics like the ones built ins on phones and laptops?

• by lvl102 on 9/11/22, 8:00 PM

  I highly suggest getting flat neutral speakers first. Preferably high end studio monitors. What would be interesting is if someone can work on music-specific optimization based on a handful of inferences and ML.

• by amelius on 9/11/22, 4:32 PM

  How does this compare to just buying a good headphone?