[vid]

contents shifted

 

[stv014]

Quote:

The graphs are noisy indeed (I take it you mean mostly in the lower frequencies?), although I'm only using 1/10 octave smoothing in HOLM as well.

 
I did not use any smoothing, and with the 20-bit sequence it can be reasonably clean. Although it also depends on how much noise is there in the recording in the first place, but it was not particularly low noise for me.
 
Quote:

Originally Posted by vid /img/forum/go_quote.gif
 
Something I've been wondering – can I just record xxxx seconds of the MLS signal and semi-randomly choose an xxx-second clip out of that recording, or do I need to very carefully isolate the entire recorded sequence with millisecond precision in the clip that I feed into HOLM? With white noise you of course don't need to input the entire recording since the sound is random to begin with, but how does it work with MLS?

 
I am not sure how HOLM works, but I guess it should be given the complete original signal (i.e. the 1048575 samples of MLS), and the recorded version of it (I do not know if it is allowed to have leading and/or trailing silence), and it would work out the impulse response from that information. Analyzing a random section of the sequence is not really better than using white noise, in this case the result will also include noise from the test signal.
But if you upload a recorded WAV file with the sequence looped a couple of times, I will check (with the convolution method) how good it can actually be.
 

 

[vid]

contents shifted

 

[vid]

contents shifted

 

[stv014]

Quote:

Well, apparently I should've thought through the act of looping the signal a bit more thoroughly than to simply queue up the signal file four times in the playlist, as the number of signal samples in the recording doesn't seem to quite match up to 4 * 1048575 like it should. I don't know why. The signal file itself is 1048575 samples, though, so I should probably just make another wav where I reproduce the signal x number of times.

 
It is apparently not a looping problem, the signal has been resampled; did you use a different sound card for recording than for playback ? Here is a FLAC format impulse response I created from the recording. I sped it up by a factor of 1.0067893 to correct the length (without that, the result would be completely wrong, but it still may not be perfect), although I am not sure if there were also any gaps between the sequences, but probably not. The frequency response of the middle impulse looks like this after being cropped to 16384 samples:
 

 

[vid]

A-ha, the sample rate! See, we're slowly working it out (hopefully). Thanks.

 

[stv014]

Quote:

Originally Posted by vid /img/forum/go_quote.gif
 
I completely forgot about the sample rate, although I'm not sure which rate I should be using with MLS to begin with (and considering the generator doesn't output a header to nudge me).

 
Any sample rate can be used, as long as it is used consistently. Although if the DAC/ADC is limited to supporting only certain sample rate(s) in hardware, then those are preferred to avoid resampling.
 
Quote:

Originally Posted by vid /img/forum/go_quote.gif
 
I can see that the low frequencies have some more noise compared to yours, but then again, anything below 40 Hz is no man's land, considering that both the recording and the signal have been highpassed at 40 Hz.

 
I do not think the highpass filtering should be necessary, especially for playback, unless the noise is really very high. When used correctly, the MLS should reduce it significantly. On the DT770 graph I posted, I used no filtering or smoothing at all (other than cropping the extracted IR to a more reasonable length from 1048575 samples (>10s at 96 kHz)), and I did record with a noisy microphone.
 

 

[vid]

Thanks a lot again, stv014. You've helped out quite a bit.

 

[RexAeterna]

those are only the EP version though. i don't know why it's not mentioned in the graph. the MP and LP use different passive drivers and bit different driver enclosure and drivers cause some sextetts has the reflective domes and some don't and the different cases has different amount of holes behind the driver like my one lp had a white casing with 3 small holes in the back and my other LP's has a red driver casing with 3 large holes in the back. each sextett varies between one another.

 

[vid]

contents shifted

 

[RexAeterna]

there is slight difference with the EP,MP,LP and there is even difference between each model as well cause akg like reusing parts so even for example one MP is different from another MP due to possible different driver casings but vary a bit if at all maybe. i don't know since all i owned in the past are lp models and i have NOS AKG 240 sextett lp that i completely rewired with 18awg wiring and terminated with 4-pin xlr connector for speaker output use and of course i made adapters as well to be used off headphone jacks, balanced 1/4'' outputs and even pre-outs. my sextetts has all new solder joints and completely refurbished as well.

i heard the EP being most bass heavy version and LP being the least but i kinda doubt that somewhat cause my lp's can slam hard and give you headaches with it's deep bass extension and powerful impact. i found the sextetts though are very output impedance dependant as well(well mine are) cause they work best off close to 0ohm output impedance off of speaker outputs if you want a very tight controlled bass impact. seem higher the output impedance the bass will seem like it's getting stronger and will warm up the sound bit but it's really cause of the loosen up mid-bass going on from high output impedance. maybe the 250 is the same way and output impedance dependent? have you tried? i know my 240DF is as well output impedance dependent as well and loves speaker outputs.

 

[vid]

contents shifted

 

[vid]

contents shifted

 

[vid]

contents shifted

 

[RexAeterna]

fun fact: all akg headphones before the 240DF and 240 monitors are free-field equalized(basically anything with the sextetts drivers). the 240DF was the first diffused-field equalized headphone AKG ever done. after the 240DF AKG continues to use Diffused-field. all the K4x, k5x, k6x, k7x series are all diffused-field. the 240 studio and 27x and 24x series are diffused-field as well but none of them are not done the same fashion as the 240DF. sennheiser was actually gonna go same route as the 240DF for the hd600,650 and even hd800 but decided to do their own modified type of diffused-field and not fellow the IRT standards. the k601 and k701 are just modified k4x and k5x series with different pads and different density foam discs around the drivers. just saying, if anyone didn't know.

also have you tested my theory on output impedance? it makes a difference for some reason. and a nice difference too with tighter bass response and clarity. i haven't heard my sextetts is while but do remember there response on multiple sources and bettering the k601 in every way still with far more extension,air and midrange clarity.