When do you consider you have a frequency problem?

[ninja12]

I am trying to figure out when I need to deal with a frequency problem vs when I don't need to worry about it. So, how much of a null/dip has to occur before you feel you need to deal with the frequency? Is it 6db, 12db, 18db, 24db?

 

[jamie2112]

Does your system not sound right to you? Just wondering. There is no set level to eq out use your ears is my opinion. 24 db is a huge cut. I would go with 6 if that is a option.

 

[bandphan]

Try using software and a high quality mesurment mic to measure your rooms response. REW is freeware and very powerful. Alot of people are suprised to find out that once these issues are addressed, ones system which may have been thought of as good can improve greatly.

 

[mtrycrafts]

I am trying to figure out when I need to deal with a frequency problem vs when I don't need to worry about it. So, how much of a null/dip has to occur before you feel you need to deal with the frequency? Is it 6db, 12db, 18db, 24db?

Usually, a 10dB change in acoustic level is subjectively considered to be twice or 1/2 as loud. That is a pretty large change in level. At some frequencies, some consider 6dB change to be 2X or 1/2.
As suggested, try to do some measuring across the band.

 

[cbraver]

Usually, a 10dB change in acoustic level is subjectively considered to be twice or 1/2 as loud. That is a pretty large change in level. At some frequencies, some consider 6dB change to be 2X or 1/2.
As suggested, try to do some measuring across the band.

I played with it once, of course very subjectively, and found for me it was closer to 6dB it seemed half as loud/twice as loud. 10dB was very dramatic in difference for me.

 

[mtrycrafts]

I played with it once, of course very subjectively, and found for me it was closer to 6dB it seemed half as loud/twice as loud. 10dB was very dramatic in difference for me.

Yep, that is why some acoustic handbooks have one number while others have the other number, even though it had to have come from research data.
So, for the poster, with such audible differences, it can have an impact indeed.

 

[no. 5]

I am trying to figure out when I need to deal with a frequency problem vs when I don't need to worry about it. So, how much of a null/dip has to occur before you feel you need to deal with the frequency? Is it 6db, 12db, 18db, 24db?

How do you mean "deal with the frequency"?

 

[ninja12]

How do you mean "deal with the frequency"?

Well, some people indicate that they consider 10db to be acceptable. Others say that anything above a 6db cancellation is a problem that needs to be resolved. I guess it depends on the room.

 

[no. 5]

Well, some people indicate that they consider 10db to be acceptable. Others say that anything above a 6db cancellation is a problem that needs to be resolved. I guess it depends on the room.

And source material ( ), I was merely wondering if by "deal with" you meant EQ or repositioning.

Since getting good bass response involves a few factors, and it can become complected (especially if there is more than one listening position to consider), I suspect that many people would find bass response that varies no more than 12dB to be remarkably better that what they have.

As for your original question, there are a few things to remember, first, dips are less noticeable than peaks, second, the width of the peak or dip is important in regard to audibility (a very narrow and deep peak will be less audible that one that is shallow but wide), third, do not rely on a single measurement from a single position (i.e. take a few readings from slightly different points in space for one seating position).

 

[D.R. Payne]

no. 5:
1) Where in Wisconsin are you roughly?

2) When you speak of the "width" of a Mode or a Null, you are actually talking about a grouping of modes or nulls, right? ergo, a mode at 40hz is not terrible by itself but modes at 38hz, 40hz, 42hz, 44hz, and 50hz would be terrible.

Dips (nulls) are not noticable because you don't know what you're not hearing unless you also listen to the same recording in a different room/car/headphones. Almost anyone can hear "boomy" bass once they learn what it sounds like.

I like your comment about 12db being "good". Most people would absolutely balk at buying a speaker that had an average response of +/- 4db instead of the standard 3db, not realizing they could easily have a 30db room mode in the lower end of the midrange in their room...

 

[no. 5]

no. 5:
1) Where in Wisconsin are you roughly?

I think I'm a little north of you, in Door County.

2) When you speak of the "width" of a Mode or a Null, you are actually talking about a grouping of modes or nulls, right? ergo, a mode at 40hz is not terrible by itself but modes at 38hz, 40hz, 42hz, 44hz, and 50hz would be terrible.

I was referring to the Q of a peak or dip in a frequency response graph, but yes, any visible (well, audible) modes or nulls may be the result of several things stacking up.

I like your comment about 12db being "good". Most people would absolutely balk at buying a speaker that had an average response of +/- 4db instead of the standard 3db, not realizing they could easily have a 30db room mode in the lower end of the midrange in their room...

Absolutely, and it gets even worse when variations between multiple sitting positions is included.

 

[EYEdROP0]

I can hear a 4-6 db difference usually, but not through the entire hearing range, only where most of the music is (bass-midrange). Treble is a bit harder to tell on a good system.

 

[Exit]

I have seen somewhere that humans can't perceive a change in sound untill it reaches at least a 3 dB change. If you are measuring your room response and using a parametric equalizer, then I wouldn't worry about frequency response deviations less than 3 db. It is a minor point, but it is worth mentioning.

 

[avaserfi]

I have seen somewhere that humans can't perceive a change in sound untill it reaches at least a 3 dB change. If you are measuring your room response and using a parametric equalizer, then I wouldn't worry about frequency response deviations less than 3 db. It is a minor point, but it is worth mentioning.

Audibility of anomalies in frequency response are directly related to Q, frequency, source material and the direction of the anomaly (either a peak or a dip). Thus, there are times when differences as large as 10dB will not be audible and others where 1dB would be.

If one is interested in reading on the subject I recommend this paper as a starting point:

Bücklein, Roland. The Audibility of Frequency Response Irregularities. J. Audio Engineering Soc., Volume 29 No. 3. Pages 126 - 131. March 1981.

-Andrew