Frequency Response With Only Two Corner Traps

[Glenn Kuras]

Not sure how much smoothing you have on but if not a lot then it is looking pretty good. Frequency response is something you want to work on, but decay time on the low end is pretty important, IMO more important. It is where the clarity of the low end comes from.

[Carcruzjo]

Thanks for the quick reply! The smoothing is set to 0 (none). I may have to buy the full version of FuzzMeasure first to test the decay time. I'm just starting to treat my room and I'm already hearing a difference. I can't wait until my order of stone wool comes in tomorrow!

[DAH]

Nice graph man! Seems like the room sounds good)

[Axiom05]

I may be wrong but the graph looks too smooth for No Smoothing. The recorded level looks about 20dB too low. Adjust the gain and output volume to get the level around -10dB and see if it looks the same. What is the background noise level of the room like?

[Jens Eklund]

A +/- 7,5 dB response from 40 to 1000 Hz without smoothing is an excellent result and suggests that you´re not in a concrete bunker with heavy solid walls (and if you are, you really need to post more info regarding your treatment). FR alone is however (as already mentioned) not the only thing that matters:

Before posting your measurement results

[Jens Eklund]

Originally Posted by Axiom05

I may be wrong but the graph looks too smooth for No Smoothing.

There’s no smoothing applied as can be seen in the higher range but the length of the FFT block looks to be quite short (16 - 32 k would be my guess) judging by the low resolution in the lows. This will have a "smoothing" effect on the graph, especially in the lows.

[DanDan]

Here's a typical FM graph with smoothing set to None.

The OP graph looks like 1/48 octave to me.
Something is not right here.

The record level looks low and the HF response suggest two speakers were used. One is the norm/default/goto.

http://www.gearslutz.com/board/studi...er-v2-1-a.html


DD

[Jens Eklund]

Originally Posted by DanDan

Here's a typical FM graph with smoothing set to None.
Attachment 317842
The OP graph looks like 1/48 octave to me.
Something is not right here.

The record level looks low and the HF response suggest two speakers were used. One is the norm/default/goto.

http://www.gearslutz.com/board/studi...er-v2-1-a.html


DD

It´s the same thing here as above; no smoothing but the FFT block length is short. Same resolution as above.

[Jens Eklund]

Set the FFT to 256k and the resolution will improve.

[DanDan]

Hi Jens. FM is a typical Mac program. We do not get to see under the bonnet.
There is no access to the underlying FFT settings.
The only obvious settings available would be length of sweep. (I recommend 10S rather than the default 1000mS) and the Window length, which also gets set automatically, but generally works fine.
Here's the same measurement I showed a while ago, but at 1/48.


One way or the other, this looks like a measurement which needs to be redone, at a higher level, one speaker etc. etc.

Carc, you can Export Impulse Response, Zip it, and post directly here. Then anyone can Import into their own software to view in any form they wish.

DD

[Jens Eklund]

Originally Posted by DanDan

Hi Jens. FM is a typical Mac program. We do not get to see under the bonnet.
There is no access to the underlying FFT settings.

Hi Dan

I would be surprised if you cannot select a larger FFT block size since the resolution shown above apparently is only 7 (or so) points between 10 and 20 Hz (so 7 give or take, p p o in this region) and although normally not an issue, I see no reason to purposely decide on such a limitation.

[DanDan]

Jens, as you can see from my Smoothed and 1/48 graphs, the thing works!
The only use of the word 'resolution' in FM, is regarding the number of slices in a Waterfall, Low 20 or High 50. This setting has of course effect on FR graph.
I can send you a FuzzMeasure by email if you would like one.
It is not very big....!

The point and shoot nature of the Mac way belies the power which is sometimes at hand. e.g. FM does appear to do sophisticated Time Domain combinations of measurements.

DD

[Jens Eklund]

16k vs 256k FFT

[DanDan]

Good points Jens. I have looked again. There are 'Window' controls in FM which seem to have a similar effect to your FFT settings. I will include that in the Primer.
Generally the default Auto works OK. However one can input whatever duration one wants. Here's a rather extreme setting but it still doesn't look like the OP's graph.


DD

[Jens Eklund]

Originally Posted by DanDan

Good points Jens. I have looked again. There are 'Window' controls in FM which seem to have a similar effect to your FFT settings. I will include that in the Primer.
Generally the default Auto works OK. However one can input whatever duration one wants. Here's a rather extreme setting but it still doesn't look like the OP's graph.
Attachment 317850

DD

"Window" usually refers to the gating times (or type) of the IR, so unless a very different terminology is used in FM, no; that´s not the setting we´re looking for here.

[DanDan]

Well rest assured Jens, I can find no direct access to the FFT length in FM.
I am not as familiar with REW, but just now I looked into it.
There is a FFT length setting on the Spectrum Analyser/RTA
But only Window settings, in the same sense as FM, on the Waterfall and Sonogram.

In any case, it is worth noting that either setting has a fairly similar effect on resolution in that important information can be rendered invisible by an unsuitable setting. I have added this to the Primer already.

DD

[Jens Eklund]

Originally Posted by DanDan

Well rest assured Jens, I can find no direct access to the FFT length in FM.
I am not as familiar with REW, but just now I looked into it.
There is a FFT length setting on the Spectrum Analyser/RTA
But only Window settings, in the same sense as FM, on the Waterfall and Sonogram.

In any case, it is worth noting that either setting has a fairly similar effect on resolution in that important information can be rendered invisible by an unsuitable setting. I have added this to the Primer already.

DD

"Window" is not what you´re looking for here. It´s the FFT block length (that determines the raw resolution of the FR graph). Have you never used different window settings or gated an impulse response?

[DanDan]

Have you never used different window settings or gated an impulse response?

I doubt the bona fide's of the question but obviously I have.
Do you use, are you familiar with REW?
Again, I can easily send you the manual for both programs if you wish.
ARTA has been well recommended to me, apart from it's Waterfalls. However, it seems best not to assume the other programs have equivalent control set.
Not surprising, as ARTA appears to be more comprehensive, and orientated towards the scientific minded.

DD

[Jens Eklund]

Originally Posted by DanDan

I doubt the bona fide's of the question but obviously I have.

Well, then you should know the difference between FFT block length and windowing durations and types.

[DanDan]

As I said I doubt the bona fides of your questions and comments.
As I said FM does not provide access to the FFT length, and REW puts it pretty much in the background. As such I have little to do with it unless I am using the REW Spectrum Analyser and RTA. Which I do and rather like.
I simply don't have a PC or ARTA and am in general much more of a car driver than a designer.
I won't comment on what I do or don't know nor quote qualifications.
But given the inferences, I will suggest that Jens Eklund is quite young.
I will further say that attempts to show intellectual superiority or even training, i.e. The use of Academia as a weapon of insult, is abhorrent.
We are all too familiar with that practice from the days of SAC.

DD

[Jens Eklund]

Originally Posted by DanDan

As I said I doubt the bona fides of your questions and comments.
As I said FM does not provide access to the FFT length, and REW puts it pretty much in the background. As such I have little to do with it unless I am using the REW Spectrum Analyser and RTA. Which I do and rather like.
I simply don't have a PC or ARTA and am in general much more of a car driver than a designer.
I won't comment on what I do or don't know nor quote qualifications.
But given the inferences, I will suggest that Jens Eklund is quite young.
I will further say that attempts to show intellectual superiority or even training, i.e. The use of Academia as a weapon of insult, is abhorrent.
We are all too familiar with that practice from the days of SAC.

DD

I´m not going to bother report this post but I really think you need to get your act together and think about your actions on this forum.

Good night.

[Carcruzjo]

Wow didn't mean to set off the heated discussion. Ill post more results when I install my other 4 bass traps in the room. Jens the only part of the room that is concrete is the floor with a thin layer of short carpet over it. All four walls are covered with 1/4 inch carpet underlay and I have a makeshift cloud over my listening position. Also I recently added a couple of qrd diffusers to the back wall but I use them more for recording purposes as they are mobile and on the floor (not level with my listening position). But when they're not in use I usually leave them on the back wall

[Jens Eklund]

Originally Posted by Carcruzjo

Jens the only part of the room that is concrete is the floor

Well, my guess was accurate then:

Originally Posted by Jens Eklund

A +/- 7,5 dB response from 40 to 1000 Hz without smoothing is an excellent result and suggests that you´re not in a concrete bunker with heavy solid walls

But:

Originally Posted by Carcruzjo

... with a thin layer of short carpet over it. All four walls are covered with 1/4 inch carpet underlay

Sticky with links?

4m x 5m x 2.5m room - modal resonance problem




... and increase the sweep length in FM when you take measurements and the resolution will increase (in the lows if unsmoothed).

[DanDan]

... and increase the sweep length in FM when you take measurements and the resolution will increase (in the lows if unsmoothed).

I would have thought so too, but I see no change in LF (or HF) resolution between a 1S and 10S Sweep. Down at 100mS the resolution decreases a bit. In FM, the Signal to Noise Ratio increases with longer or multiple sweeps.
I recommend 10S, why hurry?

DD

[Jens Eklund]

Originally Posted by DanDan

I would have thought so too, but I see no change in LF (or HF) resolution between a 1S and 10S Sweep. Down at 100mS the resolution decreases a bit. In FM, the Signal to Noise Ratio increases with longer or multiple sweeps.
I recommend 10S, why hurry?

DD

Hm ... that’s weird. I would think that since you cannot select the FFT block size in FM, at least using a longer sweep would enable FM to automatically use a longer FFT block.

So the resolution in FM is fixed at a relatively short FFT block size it appears. Well, as you say; at least the S/N ration will improve with longer sweeps so it´s still a good practice to use between 5-15 seconds sweeps (and multiple takes to further improve S/N).

[DanDan]

Mac stuff is like that in general Jens. Clever things happening under the bonnet, the user gets and easy ride. But I wouldn't assume any compromises, the resolution only diminished when I went to 100mS and then only slightly.
FM is quite a thing, but I am very frustrated with a couple of issues. I would call them bugs. If we were capable of a civil discussion I would consult you about the Windowing one in particular....;-)

DD

[Jens Eklund]

Originally Posted by DanDan

Mac stuff is like that in general Jens. Clever things happening under the bonnet, the user gets and easy ride. But I wouldn't assume any compromises, the resolution only diminished when I went to 100mS and then only slightly.
FM is quite a thing, but I am very frustrated with a couple of issues. I would call them bugs. If we were capable of a civil discussion I would consult you about the Windowing one in particular....;-)

DD

Well, how about installing Win and get access to ARTA, REW and others?

[DanDan]

I have several friends who are equally conversant with PC and Mac.
They have very strong preference.
REW works very well on Macs now. I actually prefer it to FM, ssshhhh......
Well, to be accurate, I prefer measuring on FM, which works properly with my interfaces, but REW is much better at viewing and analysing.
FM really needs some TLC.
Good night.

DD

[Carcruzjo]

So I finally got around to installing my superchunk style traps. I used Roxul Comfortbatt R-30. I've only installed two so far in the back corners and I can definitely hear a remarkable difference in the low end can't wait to fill up the front corners! Here are some fuzz tests I did after installing the traps