X-FI soundcard owners w/ mic ~MUST SEE~

those of us with an x-fi soundcard already know that the included software comes with 'audio creation mode'

inside the program is a very important plugin everybody should be using.. 'reverb'

and i'm gonna teach you how to calibrate it.

first of all, this tutorial is for speakers that are 'toed in' and pointed directly at the listening position.
pointing the speakers in towards the listening position means you dont have to take as many steps to calibrate the reverb.

this is short and sweet.. but you need a microphone.

get yourself a program like Room EQ Wizard (or something similar that displays the spectral decay ..aka.. waterfall)

it doesnt matter if your equalizer is already calibrated or not.. because if you were gonna calibrate the equalizer, you would have already done it (or you are gonna do it in the future)

use whichever program you want, and record your sine sweep.
then pull up your time domain results.
it will be the 'RT60' tab in Room EQ Wizard.. or the waterfall in another program.

you need to look at those results and find two spots:
1. the bass range with the highest amount of delay
2. the midrange area with the highest amount of delay

the delay from the bass range goes into the 'reverb delay' knob (open up the reverb and click the 'advanced' tab)

next.. you need to subtract the midrange delay from the bass delay.
my woofers have a peak of 45ms
my midrange has a peak of 28ms

so i input 45 into the 'decay time' knob.

looking at the 'group delay' tab in REW .. it shows the delay for each frequency (much more accurate for finding the frequency)
the group delay shows why the 'RT60' tab shows an increase at the upper end of the treble frequencies.
the 'group delay' tab says the high delay stops at 17,700hz
i was using 11,500hz but i looked at the 'group delay' tab and found more accurate information.

input the treble frequency into the 'high frequency cutoff' knob.
you also need to do this again for the bass and input the frequency into the 'low frequency cutoff' knob.

then find the percentage difference between the bass delay and the midrange delay.
mine was 28 is 62% of 45
so i input the 'percentage' into the 'high frequency decay ratio' knob.
(they really have this stuff figured out and made easy)

now.. that should be compensating for the bass delay and the midrange delay.

what's next is the size of the room.
i have tried all three different room sizes:
1. from front wall to back
2. length x width
3. height x length x width

they all sound about the same (choose whichever one makes the sound more crisp and clear)

the row of knows with 'diffusion' on the left, all of those knobs need to be rotated fully clockwise.
you can adjust the 'air absorption hf' knob to change the tweeters when the temperature and/or humidity changes.

the 'reflections pan divergence' knob needs to be set to 0.0

see, the reason why we use reverb is to force those echoes from the walls to stop.
you dont need sound absorbers in a room without any carpet or fabric couch.
set the reverb right and the digital processor will do its thing to remove the standing soundwaves.
and you'll defintely hear it if you fill up your room with audio from the speakers.
the sound coming from the speakers will be more clear.
its like going from a scratched pair of glasses to a new pair of lenses.

they sell sound absorbers so you dont have to use reverb..!!!
you cant use reverb with sound aborption covering all of the walls.
reverb works by USING the walls.
it might not be perfect, but its a great improvement.
keeps the walls free to hang posters (or other things) :cool:
5 answers Last reply
More about soundcard owners must
  1. i've done some more messing around and found a solid reason to use the following knobs:
    - reflections delay
    - early reflections level
    - late reverb level

    the first thing i did was try to remove more echoes from the room.
    the air was more clear, but it wasnt REALLY clear.
    so i tried to remove the sound reflections from the rear wall.
    to do this, i knew that there was an echo.. which means more reduction is needed.
    with reverb, everything is in the time domain.
    you have to enter time values for everything to work.
    so i knew i needed to try to adjust for the time i heard heard the sound, until the time i heard the sound from the back wall.

    my first instinct was to use the same 45ms that was recorded in the impulse response.
    that helped quite a bunch.
    but then an idea dawned on me, maybe i should use the highest value seen in the group delay.

    well, going all the way to the subwoofer frequencies.. i seen massive amounts of delay at about 13hz
    i know my two 12 inch woofers can output at that frequency.. so i considered the delay at the frequency.
    the delay is 115ms at 13hz
    i assume this is an 'atmospheric' delay.. meaning the entire air in the room at its maximum amount of energy (because low frequencies have lots more energy capability than midrange)

    i input 115ms into the 'reflections delay' knob and i was astounded.
    the comcast digital receiver was on a music channel.. i happened to be listening to a band named tool, the song was aenima.
    there is a knock that comes from the right speaker.
    i used the knock to assess the new delay i input.
    at 115ms the knock sounded realistic and isolated (i really thought someone was knocking on the wall at first)
    then i tried the 45ms again, and the sound appeared to grow very dull and spread out into the atmosphere.
    i didnt like it because it sounded much more dull and lifeless.
    so i put it back to 115ms and left it there, happy.

    i learned that the lowering of 'early reflections level' and 'late reverb level' made the sound.. and i mean really MADE the sound.
    it was more clear, towards the lessening of the rear wall reflections.. exactly what i wanted.

    i tried -12 and it was an improvement
    -24 was more improvement
    -36 is a dB number often used for very low attenuation.
    i left it there while i opened up the forum to post this.
    but i tried -48 because its also a common dB attenuation level.. i also tried 0.0 to see if there was any drastic change worth noting.

    and to my suprise, there was a massive difference.
    the song had changed to one of the older songs by a band named offspring. (from their self-esteem album)
    i was hearing lots of stupid adjustments, and i thought i didnt like it.
    maybe i had done something that was too much, but i tried a different song to double check.

    the other songs didnt have the same stupid adjustments.. and that leads me to the realization of 'scoring'
    i was listening to obnoxious audio mastering that is otherwise hidden because the air isnt under control.
    i turned on the dance/electronica music station and i didnt hear a single adjustment.. OKAY i thought, i am on the right direction.

    therefore, i suggest lowering both of those knobs all the way to -100
    it shaves off 'features' and allows the reverb to do its job.
    those knobs added things that get in the way and hide the scoring.

    it appears that i have know calibrated my room to a neutral.
    all abnormalities heard are those intentions of the producer mastering the audio.
    and that is exactly what we want.. a neutral canvas to allow the producers to do their thing without worrying about common room problems.

    if you dont toe in your speakers.. you have to use the knobs discussed above to measure the distance from each reflection.
    the soundwave will hit a wall if you point the speaker directly at the opposite wall.
    its a WASTE to use those knobs to try and compensate for those reflections.
    you'll find yourself running out of relection knobs.
    sure, you get the first wall reflection.. but that soundwave will then make its way to the rear wall.
    it inevitably arrives a bit late perhaps.. but most importantly, you lose the chance to use the 'reflections delay' knob for the overall atmospheric delay.
    using the knob to compensate for the atmosphere is just too much of a sacrifice.
    the knocks in the song from tool has already established that association.

    i can now turn the radio up much louder and the 'boom' from the walls is significantly reduced.

    and trust me.. my room has ceramique tile for the floor and wood paneling on the walls, with some type of plaster on the ceiling.
    the room has the potential to sound really hollow and fatiguing.

    it would appear that i am now listening to the delay imposed by the speakers (harmonic distortion/delay)

    i would also suggest you open up the 'chorus' plugin and input the following:
    - LFO rate 10hz
    - LFO depth 100%
    - feedback 0.0
    - delay 0.0

    then adjust the rack knob until there is only a VERY small difference between muting and un-muting the effect.
    i stopped at -25.9
    doing this makes the sound just a bit more lively and detailed.. a pleasant improvement compared to without it.
    it sounds like it makes the room quieter and colder.
  2. i went back to the group delay and seen 5.71hz reports 171ms delay.
    two 12's in a 12ft x 13 x 7.5ft room ... could the mic really gather such information?

    well i tried it, and the sound improved AGAIN.

    if the new dB microphones refuse to allow air in to record such air, i'll have to try the radio shack SPL meter.. and if that doesnt work, i'm gonna need another dynamic type microphone with a diagphram readily open to the wind.
  3. yea..
    i read up on how to do the reverb the old fashioned way.

    what i did was cause the reverb to 'double up' and create a ballooning effect by sending out the original signal, then use the delay knob to adjust the amount of time before the second signal of opposite phase was thrown out to destroy the first signal.

    works pretty good.
    much much better than listening to the walls.

    but if you want to know how to use the reverb like they used to do back in the '70s '80s and '90s

    here is what you gotta do.

    point the speaker directly at you.
    measure the distance from you to the speaker.
    measure the distance from the speaker to the you, then keep going until you hit a wall.
    you need to know what angle the tape measure (or string) hits the wall.

    then measure from that point to the next wall.. you have to use the same degree.
    if it was 50 degrees in.. its gotta be 50 degrees out.

    measure the distance from the first wall to the second wall.
    find out what angle you hit the 2nd wall with.

    now.. the decay time is special for your room, have a look at the rt60 times from the instructions in the first post.

    'reflection delay' is how long it takes for the sound to go from the speaker to you.
    take the speed of sound for the current tempurature of your room and divide it by 1,000
    that turns it into 'distance per ms'

    to get the time from speaker to listening position, you need to take the distance and convert it into meters.
    then take the distance and divide it by the speed of sound in milliseconds.
    the result is how many milliseconds it takes.
    input that into the 'reflections delay' knob.

    'early reflections delay' is for how long it takes for the sound to travel from you to the wall.
    do the math as above to find the value for the knob.

    'reverb delay' is how long it takes for the soundwave to go from the 1st wall to the 2nd wall.
    again, do the math above to find the value.

    now you need to look at the 'azimuth' knobs.
    'reflections pan' is the angle that the soundwave entered and left the 1st wall.

    'reverb pan' is the degree needed to point the direction directly toward the listening position.
    so if the soundwave entered the 2nd wall at 40 degrees, you need 50 degrees to make the soundwave leave the wall at 90 degrees.
    you are bending the direction with this knob and you need to know how much angle to add for pointing the tape measure directly at the listening position (for the 3rd bounce)

    i know this works because i could adjust the 'late reverb level' and hear the sound behind me get louder.
    there was only two speakers in front of me.
    raising the volume of the 'early reflections level' made the sound get louder on the sides.
    i put 'em on 0.0 each .. but you can adjust them according to taste.

    after all of this was done.. i went back to chorus and found it best to leave the LFO rate at 0.0
    LFO depth at 100%
    feedback at 0.0
    and delay at 0.0

    then adjust the volume knob on the mixer according to taste.
  4. i have had a look at 'ffdshow tryouts' to see these 'lots of others' you were talking about.

    seems like the only thing of importance is the convolver and delay.

    the reverb in the ffdshow cannot be setup properly without using the reverb with the xfi soundcard.
    there isnt anything for distance or angle.. which brings me to the quick conclusion that the settings are to be used for the 3rd wall reflection.

    perhaps those settings are capable of helping the audio 'interleave' in the LFO's
    because it certainly doesnt give the option to calibrate the walls.

    i found it funny that you said the creative reverb isnt 'important'
    as i have tried to use the ffdshow convolver.. i cant seem to 'port' the output of the media player to the ffdshow decoder/processor.

    audio players already have working convolvers.. there is no reason to use ffdshow's convolver for audio tracks without video.
    none of the video players have a convolver though.
    and that made me interested to work with ffdshow, simply to add digital room correction on top of my already functional (calibrated) equalizer settings.
    no reason why i should remove the + from the overall A i have already obtained.

    like.. if my car had TWO turbos, there is no reason why i should have to remove one turbo.

    see.. my situation is pretty clear.
    my microphone broke, so i cant re-calibrate my equalizer.
    the calibration i did was for 2 channel analog output.
    should be quite the same for 2 channel digital output.
    however, if i want to use the THX setup console, i have to use analog outputs.
    the optimizations dont work with digital output.
    that makes the delay filter important.. because i can time-shift the speakers again using the digital output.

    i purchased the dolby digital/dts encoder for the soundcard, because i wanted rear speaker output for television and movies and video games.
    if i use analog inputs on the receiver, the built-in equalizer disables.
    i used that EQ as part of the calibration process.
    without it, the amount of bass gain is exceptionally low (non existant even)
    i also used the bass boost in the THX setup console to boost 28hz
    so i have a combination of EQ on the receiver, EQ on the soundcard, EQ on the THX setup console, additional EQ from impulse response loaded into 'SIR impulse response processor' (VST plugin.. used with foobar or any audio player with VST plugin support).
    switching to digital loses the 28hz bass boost and time alignment.
    switching to multi-channel analog loses the EQ on the receiver.

    my plan to finalize things is to get an equalizer and place it between the receiver and the soundcard.
    the receiver doesnt support the new high definition surround sound formats.
    the soundcard doesnt support the new formats either.
    but if i use software to decode the audio, i can send each channel of audio to the soundcard in PCM format.. then output each channel with an analog cable to the amplifier.

    i wont be using the soundcard processor to decode the audio anymore.
    but as long as my core2duo doesnt reach 100% processing level.. it shouldnt be a problem.

    the microphone is probably $85
    the microphone preamp with phantom power (emu 0404) is probably $200
    the equalizer that goes between the soundcard and receiver is probably $100 - $300
    - i think i am gonna try to use an audio control DQS for the equalizer though.
    it gives me 30 bands for each channel.. and that is perfect for abnormal speaker locations.
    for instance, one of the rear speakers is close to a side wall.. the other is not.
    therefore, the one closest to the wall is going to need a different calibration.

    i dont use a center channel or rear center channel.. but if i decide to use one, i will be able to adjust the equalizer for that specific speaker too.
    plus.. there is individual volume control, and that can be useful when the center channel is 2ft away from me as the other speakers are 4-6ft away.

    what sucks about the reverb..
    both speakers have to be equal distance from the listening position.
    both speakers have to hit the wall at the same angle (usually true if the speakers are at the same distance from the listening position.)

    chorus can shift the LFO from right to left with one feedback knob.
    and chorus can shift the LFO from front to back with a second feedback knob.
    maybe the first knob shifts the LFO from front to back, i dont know because i only calibrated one speaker.

    of course, if i really wanted to, i could connect the other front speaker to the rear channel and use the reverb again.
    it would provide the empty slate needed to calibrate the other walls that are different angles/distances.
    but listening to surround sound formats from television, movies, or video games is then completely out of the question.

    i believe the convolver that comes with 'ffdshow tryouts' requires the impulse response to be already inversed (i dont know because i havent gotten the program to work)
    but there wasnt any documentation saying corrections where being applied.
    convolvers generally apply the impulse response to the audio, and that means the impulse response file needs to be inversed prior to being loaded into the convolver.
    otherwise, if you add the impulse response file without inversing it.. the output will be the ringing of the room twice as bad.

    i havent tried to inverse the file in any programs yet.
    but the idea is to make the impulse response completely opposite.
    that doesnt mean simply flip the phase.
    its like, you gotta apply all of sound space to the file and remove the impulse response to deaden the ringing affect.
    but the easier way to do it is to simply inverse the impulse response to the appropriate shape.

    of course, that would give the digital room correction software a run for its money.
    simply inversing the impulse response wont be enough to charge money for the program.
    instead, they will have to inverse the impulse response.. then add time delay.. adjust the phase across the entire frequency response.. and finally, apply an equalization of the new dB output.
    it would boil down to what method they use to add the final equalizer calibration.. what method they use to excite the room to gather decay delays.. and what to do when a speaker distorts badly because it refuses to play with all of the different phase changes.

    i would think you need a simple graph that shows all of the phase changes needed for perfection.. then give the ability to toggle each change on and off, so that you could turn off the problematic changes that corrupt the speaker's motor.
    that way, people can get as many changes as the speaker supports.. and if they want to be closer to perfection, they will have to buy higher quality speakers.
Ask a new question

Read More

Sound Cards Audio