[Mitchco]

After investigation, the woofers in my 3-way time aligned system were offset by a few samples relative to the mids and highs. Fixing that plus linearizing the drivers led to a better response:

 

[natehansen66]

What's the window length and how many cycles for the fdw?

 

[Mitchco]

Default values used for both window length and FDW.

 

[natehansen66]

Does anyone know of any literature that goes into our perception of sound and these "sliding" windows we're using? It makes sense to me, but whats optimum? Or to put that a different way what window length vs. frequency best corresponds with the way we perceive sound in a room? REW seems to default to 15 cycles, is there anything behind that number to explain the choice there?

What's interesting is in my setup the floor suckout is at 400hz. With the right window at 500ms, and the FDW at 15 cycles the suckout is visible in the FR on the left speaker but not the right. Change the FDW to 5 cycles and the suckout is apparent in both speakers. Obviously there's other reflections "filling" the floor bounce with the longer FDW window for the right speaker but not the left(lots of asymmetry in this room), but what is correct here?

I have links to a few AES papers somewhere, but before I go and spend $20 a piece on several papers I'm wondering if anyone has any suggestions? I'll see if I can find those links....

 

[3ll3d00d]

There is some commentary on this in http://www.amazon.co.uk/Auditory-Neuroscience-Making-Sense-Sound/dp/0262518023 which starts with talking about how the basilar membrane works, how it can be modelled as a set of gamma tone filter banks and how those filters have progressively shorter impulse responses as frequency rises. I am not aware of any aes papers on the subject (though no doubt they exist).

 

[natehansen66]

Thanks for the link. Now that you mention it I seem to recall Geddes posted a chart showing gamma-tone impulses and their duration not too long ago.

 

[3ll3d00d]

Thanks for the link. Now that you mention it I seem to recall Geddes posted a chart showing gamma-tone impulses and their duration not too long ago.

it's something like this

http://www.mathworks.com/matlabcent...labcentral/mlc-downloads/downloads/submissions/32212/versions/18/screenshot.jpg

which is, AIUI, basically like what FDW does

as to the choice of window length (cycles), 15 is the default in acourate IIRC for magnitude correction but 6 is commonly recommended as a good starting point for the excess phase correction. This is an example of how it depends on you're trying to do. I've experimenting with a variable cycle length to try to get a view into some LF issues, I find that playing around with it just gives you more data to try and understand what is going on. There was a suggestion once around using decay times to guide your choice of cycle count but that's the only objective suggestion I've seen.

 

[natehansen66]

I was just about to post that same link

Geddes said he calculated the gammatone filter at 70hz to be about 60ms and 50ms at 100hz. If we were to go by these numbers then it seems a variable fdw would be necessary. This is basically what I've done in the past...........just use different windows for different freq ranges. The nice thing about the fdw is I can fairly easily do low crossovers (80-300hz) with measurements taken at the lp. Due to our perception at these lower freqs the resolution is "close enough".

Here's the papers that I was talking about, cited by speakerdave aka David Smith:

A New Psychoacoustically More Correct Way of Measuring Loudspeaker Frequency Responses
Preprints 1871 (F-4) and 1963 (G-4)
Jorma Salmi

A Perceptual Criterion for Loudspeaker Evaluation
James M. Kates
JAES Vol. 32, No. 12, Dec 1984

and: Samuel Bridges
Effect of Direct Sound on Perceived Frequency Response of a Sound System
AES preprint 1644 (H-6)

I haven't checked these out for myself yet but I'll probably pick up the Salmi paper.

 

[3ll3d00d]

Geddes said he calculated the gammatone filter at 70hz to be about 60ms and 50ms at 100hz. If we were to go by these numbers then it seems a variable fdw would be necessary.

Thanks for the links.

FWIW acourate has a variable fdw, you can set the left and right and low and high cycles independently. IIRC low is something like 20Hz and high 20kHz, it then interpolates between the two points.

 

[AudiocRaver]

...what window length vs. frequency best corresponds with the way we perceive sound in a room? REW seems to default to 15 cycles, is there anything behind that number to explain the choice there...

It is not my intention to start an argument, just food for thought...

I am thoroughly convinced that, although the "norms" given from the research are certainly "normal" and accurate and therefore useful in understanding a baseline, that some of these perceptual patterns are subject to exposure and experience, and can be trained to discriminate with finer resolution well into the areas where they are thought not to be able to. This does no mean that such an individual is exceptional or weird or is "wired funny," only that he has paid attention and learned to hear things outside the norm. FWIW.

 

[natehansen66]

It is not my intention to start an argument, just food for thought...

I am thoroughly convinced that, although the "norms" given from the research are certainly "normal" and accurate and therefore useful in understanding a baseline, that some of these perceptual patterns are subject to exposure and experience, and can be trained to discriminate with finer resolution well into the areas where they are thought not to be able to. This does no mean that such an individual is exceptional or weird or is "wired funny," only that he has paid attention and learned to hear things outside the norm. FWIW.

It's clear to me that highly trained listeners can "hear" things that the average Joe can't, but I'm not so sure that's outside of the current understanding of how we hear. There seems to be a lot of grey area in psychoacoustics and it seems much of that is interpreted differently by many. I seek to find the interpretation that correlates objective data with my subjective impression......when these things jive I feel I'm on the right track.

 

[AudiocRaver]

I seek to find the interpretation that correlates objective data with my subjective impression......when these things jive I feel I'm on the right track.

Well said.

 

[Mark Fuller]

It would be great if the RTA could run in 31 band mode for use with that common graphic equalizer design.

 

[lkwrnflknf4lk]

I think FDWs only make sense after exporting minimum phase versions of the meassurements. any delayed frequency in the original meassurement will obviously be (partly) missing in the plot.
here is the same meassurement, 15 cycles, no smoothing: green is original, red is minimumphase export, re-imported

 

[lkwrnflknf4lk]

Some measurements of my L speaker (which is a fairly small sealed speaker, MK MP150). Mic is at the usual MLP about 3m away, mic pre is an RME FireFace 800, mic is a CSL calibrated EMM6, same levels used for both sweeps & same length sweeps used (approximately... 512k in REW is 10.9s and it was a 10s long sweep in Acourate) & same freq range (10-24000Hz).

Two wavs attached, one is the acourate sweep itself (in 24 bit PCM form for import into REW) and the other is the same after a 15 cycle FDW applied.

Comparison of REW variable smoothing and acourate 15 cycle FDW below

View attachment 94442

Here is MP exported reimported 15 cycles

 

[lkwrnflknf4lk]

OK, here we go. It appears that the "default" psychoacoustic smoothing in Acourate works this way: The window width is a linear interpolation from a low frequency of 16 Hz up to Nyquist. Not sure why Nyquist, but if the Sample rate (most commonly used) is 48 kHz that would be 24 kHz, which isn't too bad as a standard.

He takes this window measure on 1/24th octave intervals.

The apparent "default" window width is specified as 15/15, which means 15 cycles at the low frequency and 15 cycles at the high frequency. These values can be changed, but I've found they produce a result that correlates well with the ear's interpretation of the response. 15 cycles would be 15 ms. at 1 kHz. To the best of what I can determine this is actually the right hand window width, and he uses a Blackman, but some experimentation is in order. Anyway, changing the left hand window size, as long as it is reasonable, does not seem to affect the displayed frequency response.

To compare REW and Acourate I took a screenshot of Acourate's psychoacoustic response of my front left speaker taken at 9 feet distance. I exported the impulse response as a 24 bit/48 khz wav (attached). I then used my usual 500 ms. Tukey with 1/6 octave smoothing to display it in REW. I then saved this as a PNG. I then brought both of these into photoshop, carefully matched the scales, extracted the Acourate curve with the magic wand tool, and overlaid them matching amplitude at 1 kHz. That's the image I've attached.

I'll give you my thoughts about these differences in my next post in this thread.

and the same for this one