Sign in with
Sign up | Sign in
Your question

Digital room correction

Last response: in Home Audio
Share
January 17, 2011 1:09:17 AM

i am looking for some digital room correction.
i already own a calibrated microphone and have set the equalizer with that, but i would like to use some digital room correction on top of that to further flatten my frequency response.

now i have seen some of the offerings.. so i really dont know what i am going to accomplish with this post.
coneq - $675 for the p8 version of the plugin and comes with the basic version of measurement software.
the basic measurement software doesnt allow for microphone calibration files.. a very huge problem.
the measurement software doesnt mention anything about a soundcard calibration file either.. which is apparently needed if you want FLAT (maybe the analog to digital convertor ruins the results of the digital to analog convertor.. i am not one to test each one in the circuit individually)
the basic measurement software samples at 48khz.. and that means your noise floor is gonna be higher, among other things you lose from having the DAC in a lower sampling frequency.
(all of my results show the DAC performance as far superior in higher sampling rates)
on top of all that, the cost is super expensive.

another option i was looking at: (((acourate)))
cost - $381.54 (340 euros)
i got an email explaining how to use a microphone filter file...
'Hi,

thanks for your interest in Acourate.

Please note that the demo version is just to get some feeling for Acourate but not to compute real filters.
But anyway you can import a microphone calibration file (*.txt format) with File - Read Amplitude - minphase and use it directly in the first step (macro1) of the room correction. The text file simply needs pairs of frequency (Hz) and amplitude (dB).

Best regards
Uli Brüggemann'

havent used to demo yet.. but the price is acceptable and you can use a microphone calibration file for that price.
a relatively high positive - acourate works in 24bit resolution / 96khz sampling rate
(havent opened up the software to see if there is a soundcard calibration file allowable)


another option i have read about: ARC system
this one costs $600 and comes with a microphone.
the website says the software works in 32bit with 96khz sampling rate.. but the microphone frequency response is only 20hz - 16,000hz
however.. when used with ARC, the calibrated frequency response is said to be 16hz - 20,000hz
some might see the condensor microphone as a downside since the mic preamp would have an affect on the final result.



the last one i have read about (and tried) is: Audiolense
the program allows for 32bit resolution / 96khz sampling rate
it supports ASIO
allows you to select the frequency correction width
you can change the time alignment.. i guess the filter adds whatever delay you specify.
and it also allows you to use a microphone calibration file (you can even view the microphone calibration in a frequency graph)
you can calibrate one speaker at a time or both speakers at once.
you can create whatever target curve you are capable of drawing.
its a pretty easy to use program.. cute and simple.
BUT.. i didnt notice any difference when using it.
i havent tried using the program without an already calibrated equalizer.. but will give it another go when i decide to try acourate.



i would really like to see an option that allows for soundcard calibration.
i mean, if we can have a microphone calibration file.. why stop there?



does anybody know of some cheaper options?
does anybody care to share how to add the soundcard calibration file to the mic calibration file so the entire input is calibrated before doing an analysis?
or maybe there is a digital room correction program that allows a specific place for soundcard calibration?

currently i am using foobar with the resampler at 96khz
i used the convolver plugin to load the generated correction file.

maybe some options or suggestions about using foobar with the convolver plugin?
January 19, 2011 2:27:25 AM

well today i tried the acourate program.
simply said.. the program does not allow easy/automated use of a mic calibration file.
i read the email instructions.. and read then again.. and read them again.
i cannot make sense of using a mic calibration file and i am not going to go through the menu fiddling with it until i get it right.
therefore i uninstalled that program and have since taken it off my option list.

i learned something today.
room EQ wizard allows some type of correction filter.
if you do a room measurement and go into the equalizer portion of the program, there is the option to construct a filter for generic equalizers.
there you can set your target and allow the program to make a filter to correct dips and peaks in the measured response.
you can actually see the filter correcting the measured response with opposite amplitude.
however, the resulting sound was very dull.
the bass was lacking a huge amount with my two twelve inch woofers.
the vocals stick out an amount.
and i dont remember the treble.
you can try this yourself by selecting the 'export filter impulse response wav' from the menu.

today i also realized why audiolense wasnt creating any real difference when i loaded the filter into the convolver.
the filter is a 'dummy filter' that is only supposed to be used to see that the convolver accepts the file and is reading/using it.
the output should be lower because the dummy filter lowers the overall amplitude.
i had the 'auto level adjust' checked in foobar and therefore wasnt hearing a difference.
but yes, the convolver was accepting the filter and the output was lower.. verifying the chain of procedures was working (except for the actual correction of the measured response).


therefore i decided to tackle the DRC program.
i have been putting it off because i am crosseyed confused when doing anything command-line based.
once i decided to hunt down some instructions and force myself to follow them.. i couldnt find any instructions for a while.
after some searching i did find this: http://www.duffroomcorrection.com/images/d/de/DRC_Guide...
it goes into detail about using DRC.
how to generate a log sweep.. how to record the log sweep.. and how to get the impulse file that is needed for DRC to generate a correction filter.
but since room EQ wizard already measured my room response.. i decided to be lazy and use what i already had.
a solid plus side to using room EQ wizard is the fact that you can use a mic calibration file AND a soundcard calibration file at the same time.
going through the steps provided in the link above.. there is no real option to use a mic calibration file and/or soundcard calibration file.

so, instead, i went back into room EQ wizard and exported a NON filtered impulse response.
then i skipped down the DRC guide to the point of loading the impulse file into DRC.
dont make the same mistake i did at first.. i couldnt find a program that would convert a .wav file to .pcm
eventually i gave in and downloaded cool edit pro 2.1

i ran into some problems trying to use DRC.
first of all, the selected file would not run because the filename had spaces in it.
then it wouldnt run through the operation because the impulse file was a stereo file and not a mono file.
after changing these two variables, i finally got the program to complete the processing and give me some filters.

the next problem i ran into was choosing the wrong created file to use in the convolver.
DRC produces two files.. and i will get back to this in a minute.
but after selecting the right file.. it was time to select which method of correction to use.
i tried normal.. extreme.. soft.. minimal.. and then finally insane.
using the wrong file.. i could clearly tell there was a difference between them.
after realizing the sound wasnt right because it had echos, i tried the other file and things were better.
a quick look in the DRC guide will show you that the file you are supposed to be using shows the impulse response far to the left.
the other file shows the impulse response in the middle of the waveform.
i then switched from soft to insane expecting to hear some distortion.
the insane file states that the file is to be used to show correction artifacts.. however, i dont hear any.
therefore i kept it in use.
i turn the convolver on/off and can rather easily tell the difference.
but remember, i calibrated my equalizer before using the digital room correction.. so my change shouldnt be vastly different.
the bass is audible now compared to the room EQ wizard's correction filter.. the treble isnt screeching at all anymore.. and the vocals arent excessively louder than everything else anymore.
my details havent improved a great deal.. but the result is refined.
and if you cant comprehend what i mean by refined, picture sound effects like stereo seperation/blending are much smoother.
subtle changes of octave blending/fading are strong and hard to miss.. but these were previously hard to hear in comparison.
again, stereo effects are softer .. but general sound f/x panning can now be heard when previously it wasnt there at all.
(i mean some songs, i could tell they were trying to do something.. but the effect was so short in duration, i couldnt tell what the point was)
now i can hear the time and effort that was put into the track with original mastering.
the frequency response is clearly more aligned.. which was the point.
it does sound like i actually lost some detail.. but i think it has a lot to do with changing the amplitude of the frequency spectrum as a whole.
i know that lower fidelity speakers will struggle if you try to force them to play a flat frequency response.. they cant do what you are asking because there arent enough 'options' in the voice coil (and/or magnet) to allow for the request.
the best way i can tell you so you can listen for yourself, is, the crisp clarity dulls down a little bit.. and depending on the speaker, there might be less overall energy being emitted from the speaker.
if you wanted a visual representation of what i am talking about.. the spikes that leave the massive chunk of amplitude in the measured frequency response .. those spikes will shrink.
and you will hear it as a dull-ing of the very subtle details like the saliva in a persons mouth .. or the details in the pick of someone stringing a guitar.

its a problem just like a truck engine.
if you try to pull a trailer that is seen as heavy to the engine.. the acceleration will slow down.

but let me tell you.. aftering calibrating my equalizer with a calibrated microphone .. then correcting that frequency response with digital room correction ..... the soundstage has really changed into an even better blending between the two speakers.
and the reason i say 'even better' is because my soundcard comes with THX console .. which allows you to calibrate the distance and angle of each speaker.
it really helps.. especially for me because i have one speaker closer to my listening position than the other.
my receiver has the ability to change listening distance.. but it assumes both speakers are at the same distance apart and the distance is applied for both speakers (nothing individual).
besides.. i cant turn the feature off on my receiver and had to set it to its lowest setting of 3ft
then i went into the THX console and compensated for the receiver by subtracting the 3ft from the distance input.
and simply doing that with the THX console really helps the speakers perform in uniform.. but applying the DRC filter on top of it has made an envelope.
i'd say the first stage is to get the soundfield matched.. then if you can go a degree further, that would be an envelope.
i guess the THX console doesnt match all octaves exactly even .. and using the digital room correction has filled in those gaps.

i want to go back now to the two files DRC produces when you create a correction filter.
one of the files will produce an echo .. the other file doesnt.
the fact that i tried using the reverb settings of my soundcard to compensate for my room size, i am experienced enough to know that those echos can be used for something.
it wouldnt suprise me if you could use one DRC file for the front speakers .. then use the DRC file with the echo for the back speakers to create its own reverb cancellation to help fill the air with sound.
i dont know what size room those echos are made for.. but its possible that one DRC file spits out pulses , and the other DRC file absorbs those pulses in such a way to create a final result of higher complexity.

this is just imagination and possibilities speaking.. what or why is not fully known.
the ability to adjust those parameters to change the echos are also not known because i am not willing to go that in depth.
i suppose some of my lost detail could by hiding because i am not using the other file for the rear speakers to bring both sets of audio waves to a matching point.
i'm not gonna lose much sleep over it because the final sound is good.
there is so much more midrange in the audio now that i consider it an upgrade in detail simply because i hear them better, despite their presence of detail.
(simply making them present was the first step)

using DRC was as easy as putting the files needed in a folder.. then place the impulse file into the folder and renaming it to the sample impulse response file.
open up a command prompt and change the directory to the folder DRC is in.
and then typing ' drc insane-48.0.drc ' and pressing enter.

i made the mistake of letting all the log sweep generation and other tasks make the command-line use a lot longer than needed.
because the more command-lines needed .. the more you have to understand and manage on the side.

DRC is free and works an improvement.
however, i am using a mono impulse file.
i didnt record each speaker seperately.. and thus, i might not be experiencing the full experience of what DRC can do.
i have to giggle a little bit because i really dont know if the correction filter is being applied to both speakers or only one.. lol
i will probably try to record each speaker seperately tomorrow.
i know that then DRC will have to work extra to compensate for the differences between the two speakers.
i also know that my equalizer calibration was taken with both speakers at the same time.
i dont know if recording an impulse response for each speaker seperately will bring each speaker to a point of greeting the equalizer at a higher state of bliss.
will make it worse/will make it better .. all depends on the ability of DRC.

i also havent used a loopback on the soundcard to grab a timing reference for the room.. but i believe if i did, the final results would be met with more detail and higher quality results.
you can adjust resonances with the initial equalizer calibration.. but that doesnt change the phase of the frequencies to match, only the amplitude changes to match.
changing the phase will alter the need of amplitude change.. and will drastically change the characteristics of the detail coming from the speaker.
if you think your low fidelity speakers are growing dull from amplitude changes .. wait until you start changing the phase of frequencies !
mix the phase change with amplitude changes and you have more things to cook than your four burner oven can accomadate.
besides.. raising the amplitude instead of changing the phase is only going to make something louder .. and if you make something louder, you are drowning out lower volume details.

currently.. i am just happy that all my octaves are being heard at the same decibel level with a much flatter line shown in the frequency response measurement thanks to the digital room correction.
i plan on going in-depth to get some phase change results from here.
that means changing the phase of a frequency to compensate for room resonances instead of decreasing the energy amplitude from the cone.
because if your cone isnt pushing out the sound hard.. you arent hearing the details hard.
i am probably hearing details all muffled from other amplitude changes.
my woofers are much better and i am pleased to hear the 12 inches of cone area.
everything before was sound pressure levels that needed to fill the room before anything substantial was obtained.
it was too loud before.. i had way too much midrange, then too much treble, before i was happy with my woofer output.
m
0
l
January 19, 2011 7:30:33 PM

it it sounds fine, and it sounds flat to your ears, theni dont see the prob of getting further.

do u do any mixing or mastering work? or a film fanatic? cos i dont think its nesseccary to go this far. plus it wont hurt to use your ears, its not too hard tbh.
m
0
l
Related resources
January 20, 2011 3:59:26 PM

MEgamer said:
it it sounds fine, and it sounds flat to your ears, theni dont see the prob of getting further.

do u do any mixing or mastering work? or a film fanatic? cos i dont think its nesseccary to go this far. plus it wont hurt to use your ears, its not too hard tbh.


i used to play around with the Ejay and Magix programs.. but this was before i got my calibration microphone.
no mixing or mastering anymore.
i'm a little bit forced to watch movies and television to pass the time, but i'm not a fanatic.
besides, i dont have any way to take the HDTV box and plug it into the computer and apply the correction filter to it before it gets sent out of the soundcard.
if i went to school for programming, i would work on a way to use the convolver on my SPDIF input so i could enjoy regular television audio much more.

for most people, using digital room correction IS necessary.
and let me explain why..
most equalizers in use today are ONE octave.

that means you have one knob to adjust as much as five different frequencies (and all the frequencies inbetween).
for example..
my 62hz slider is responsible for these five key frequencies:
60hz
70hz
80hz
90hz
100hz

the next highest slider is 125hz

another example:
my 31hz slider is responsible for all these key frequencies:
30hz
40hz
50hz


i cant possibly adjust each key frequency accurately because i dont have enough knobs on my equalizer.
there is an improvement if i do it.. but that leaves room for more improvement.

i was talking about how much better my woofers sound producing bass.
and that is because all the gaps between 31hz and 62hz have been given proper attention.
using only the knobs for 31hz and 62hz .. there could be a huge gap at 40hz that wont get fixed .. and then a rise at 50hz that wont get fixed.
i dont have a knob to fix 40hz and 50hz
that is what made the digital room correction needed.
the software can make adjustments to 40hz and 50hz
and it obviously did because i am now reminded of listening to a subwoofer in the trunk of my car where the bass notes could not be avoided.
sure, there was dips and spikes back then too.. but all of the bass frequencies were louder than the midrange and treble, so you couldnt avoid it and it trained me what to listen for while trying to get some good bass in the house (with much less wattage).

and its important to have some training from a subwoofer in a car to know what is possible in the house.
otherwise you might be tricked into the same 40hz tune and considering yourself done.
there are many many three-way speakers that stop at 40hz
even the big woofers like the 12's and 15 inch
this makes people believe they need to spend more money on a dedicated subwoofer.
its not true and the box you have the speaker in will play the biggest part.
there is lots of evidence backing up my claim.
what frequency are most bass boosters at? 40hz
there is a lot of fun to be had below 40hz
and i knew the woofers i am using are capable of playing lower than 40hz
i took some pre-made boxes that had a frequency response down to 40hz
went up to the local hardware store and picked up some three inch pvc pipe and two right-angle elbows
i cut out a hole for the three inch port.
placed the elbow in the hole i just cut and proceeded with measuring the port length .. cutting off what i needed for the proper port size/length to get the box tune that i wanted.

all i did was go to an online speaker box calculator and input the measurements of the box i had.
it was a sealed box when i first bought it.
i also replaced all of the speakers that were screwed onto the box.
i even went as far as upgrading the crossover.

anyways.. i didnt compensate the port being inside the box (nor did i compensate the back of the woofer being inside the box).
and a trick that i learned to make the box perform as if it were a bit bigger... add some poly-fil inside.
so i filled each box with poly-fil to try and compensate for the new ports and the woofer basket/magnet.
my tune was aimed at 28hz
but realistically, i can hear all the way down to 20hz and lower.
i really cant tell where the box loses its pressure.
and i think it has everything to do with the fact that i put the end of my port down by the woofer.
the elbow is up at the top of the box.
so when the box loses pressure.. its gradual and the only way to know for sure would be to catch a glimpse of excessive xmax added to the cone between one frequency to the next.

but anyways..
i am cheating quite a bit because from about 30hz on down, i have port noise.
that is the ability to hear the air moving in and out of the port.
without that air noise.. the lower frequencies would sound like pure bass from the cone and would be harder to hear.
but thanks to the accompanied air noise, i have audio from the woofer plus audio from the air.. which helps give me the perception of amplitude.


another very important reason for using digital room correction is the midrange.
its the same principle.
i have an equalier knob for 2,000hz and another one for 4,000hz
but no way to adjust 3,000hz using the equalizer.
there can be a lot of dips and spikes between 2,000hz and 4,000hz
so again, using digital room correction to fill in the equalizer gaps really really helps.

the treble is even worse.
one knob for 8,000hz and another knob for 16,000hz
theres 9khz 10khz 11khz 12khz 13khz 14khz 15khz between those two knobs
thats a lot of gaps.

i dont know if DRC adjust each frequency (10hz 11hz 12hz 13hz)
or if its (10hz 20hz 30hz 40hz)
but the convolver says there are 65536 samples and 131072 points for a mono impulse response file.
where the DRC program has applied a correction cant be said by me.
but i know its enough to fill in some equalizer knob gaps.

all of the creative xfi soundcards come with a one octave equalizer.
i'd say they sold lots of soundcards and therefore there are lots of people who can benefit from using digital room correction.
there are probably other soundcards that come with an equalizer.. and i'm certain that they dont have enough knobs.
yes, there are software equalizers available that have more knobs.. but they degrade the audio if you move the knobs too much.
but i would even give those equalizers a dose of digital room correction to try and fill in the gaps between those equalizer knobs.

i'm just saying, if you have a 1/3 octave equalizer.. see what digital room correction can do to help you fill in the gaps.
and if you have a 1/6 octave equalizer.. i'd find that cute and suggest the use of digital room correction again!


**edit**
i was going to take some pictures and show the following differences:
- no equalizer
- with equalizer
- with equalizer and convolver

but i ran into a problem.
the pink noise loaded into the audio player isnt as loud as the rest of the music.
and if i raise the volume on the receiver.. it skews the results because i am asking something different of the amplifier.

but this isnt the biggest cause of concern.. because if i play a lower amplitude audio file and raise the volume of the amplifier, the components should produce the same output as long as i line up the frequency response to normal listening levels.

but again.. this isnt the biggest cause of concern.
i remembered that i was lazy and left my PC fans on high.
i also have a box fan in the doorway to draw some of the cool air on the floor towards the thermometer that controls the temperature in my apartment.
(the room with the thermometer was a bit warmer than my living room.. so i decided to draw the cool air out and inform the thermometer of the cold air)

anyways..
there are peaks in my frequency response that the mic is picking up.. and the only way to alleviate the problem is to make the source of those peaks go away - or raise the volume so the frequency response measured is higher than the background noise.

i didnt care to change the background noise because i was being lazy.
but now that i think about it.. the digital room correction might be trying to decrease the volume of background noise, which would affect the audio.
so i am going to try and reduce the background noise and do all my measurements again.
otherwise i have to raise the volume and i dont want to make my neighbors angry with loud pink noise.
m
0
l
January 27, 2011 2:23:08 AM

i broke my microphone last night :pfff: 

it was on my lap and it slid off and hit the ground.
doesnt work anymore.

i only paid $30 or $50 (cant remember) for it.
but i just got it calibrated a couple months ago.

i was going to calibrate some headphones so i could show somebody the difference i could make.
i never got the chance to measure the frequency response of the headphones.
it was a dynamic microphone.. and i dont think i will find another mic that plugs directly into the soundcard.
i'm gonna have to buy a calibrated ecm8000 and preamp with phantom power.

if i could sue for time, i would..!
m
0
l
!