TOPIC:

Initially, I thought the notch was very audible, but it seems my new setup just needed a slight house curve, compared to the previous. That had a lot to say. Good idea with the sub. I have decided not to sell my old gear, so I will test that.

What type of measurements could be interesting with regards to your last sentence, just sweetspot, both speakers, no correction? (Not each channels separately)

With one sub there is not much you can do. Your mains are fixed. Run them full range.
Put the sub in the corner near your listening place. Hope for the best.
If you had three subs, you could follow Geddes. One sub in a corner, one on one opposite wall but not inthe middle of it. The third as far as possible from the previous two. Avoid symmetries. Geddes did his PhD in acoustics in that area. I experimented a lot with bass and find his solution better than Welti's (Harman Kardon). The nice thing is you can usually skip fine tuning subs positions. Geddes works best most of the time. Although Earl himself recommends experimenting as all rooms are different. Oh,all subs are on one mono channel. No DRC.

lintoft wrote:

What type of measurements could be interesting with regards to your last sentence, just sweetspot, both speakers, no correction? (Not each channels separately)

Bass is often said to be "mixed mono" (altho it isn't always), so it may be that with both speakers the null is less pronounced.

Again, sorry if what I describe below is me really working against the principles of Dirac, but I am puzzled by the results:

I ran some test tunes, and it was obvious that the null was at/just in front of sweetspot. If I were to scoot all the way forward, to a near field position, the response is quite nice and flat.

There is no smoothing applied to any of the curves below. Also, all measurements is one point only, directly in sweetspot.

Anyway, I pulled the speakers 40cm out from the front wall (this can not be a permanent position, but I could do this when wanting to get optimal sound) and, when sitting directly in sweetspot, aside from the nasty bump at 65, resutls were not too bad (no correction):



I ran a Dirac measurement, one position only, and optimized. After optimization, this was the curve (Dirac correction only):



I do not have the separate right/left measurements with me right now, but they were quite different, the left speaker being way louder to 70hz, then dipping more at 100hz, while the right has a dip at 150hz.

I tried to compensate using PEQ only, and the results seem better (this was done very quick and dirty, not much testing) (PEQ correction only):



I have seen that in some cases it is recommended to disconnect left/right in Dirac etc. Could this be one of these cases, and how would I go about doing that? Again, I am happy compensating for sweetspot, or a very narrow area only.

I know it’s hard to not get trapped in the chase for the perfect target... but any measurement method that uses only one point is useless.
Especially for Dirac, using it with one point only is a big waste.

Can you re-do some REW measurements pre-any EQ with at least 8 points?

Thanks. I will do that. Should I then average the results, and should the measurement points be similar to what I would do in Dirac?

Also, when you say that it is a big waste. Is it a waste because Dirac is able to do much more, to a larger area, or is it useless because I will get results as I show above? (seemingly producing worse results than with PEQ only).

lintoft wrote:

Thanks. I will do that. Should I then average the results, and should the measurement points be similar to what I would do in Dirac?

An average would show you something closer to what you’ll hear, but you could also keep all independent sweeps to illustrate the « variability » of the results. So 8/9 sweeps + the average for ex on a screenshot, per L/R.
And using the same points as for Dirac would be a good start yes, or close, no need to use the exacts same point by the mm.

lintoft wrote:

Also, when you say that it is a big waste. Is it a waste because Dirac is able to do much more, to a larger area, or is it useless because I will get results as I show above? (seemingly producing worse results than with PEQ only).

Yes using Dirac used with one point will only fix this point, and probably make all others worst.
It can do more and better because it will determine where to adjust or not depending of the same «  variability » in measurements.
John sure knows more than me here but it's the idea, the more you feed it the more it can detect where it can boost or cut and where it’s either needed, not needed or just impossible (the sweet spot still being important because it drives the time alignment).
And personally, I had the worst results using a very small area for Dirac measurements.

You could also use REW with all these points to make the system already pretty solid before Dirac (with sub placement/crossover, EQ etc).

deleted

Thanks.

This is a very simple, two full range speakers, setup. (I tried my sub to fill out the dip, but no luck). As mentioned, a near field positions gives a pretty flat response, so I think the speakers themselves perform well, the peak/dip is all due to the small room and sub-optimal position.

One point measurements in-room are generally pointless--unless you hold your head in a vice while listening. To fight a null with one sub put that sub where the null occurs. Then you might get another null elsewhere.

Well, I do not have my head in a vice, but I don't think I typically move my head more than 5-10cm from sweetspot when working and listening to music.

Anyway, I did average a larger number for measurements from REW, and as you say, these are fairly flat. I still don't really understand how many non-flat measurements can improve the sound I hear at one single point, but the sound is good, so I guess I am happy.

Google "soundandvision speaker measurement 101" by Brent Butterworth.
It's a good primer.

I skimmed through the article, and I understand, that to properly measure a speaker, it does not make sense to measure it at a single point in a standard room.

But, I am not really interested in measuring my speakers, I just want to get as good room correction as possible at sweetspot. And, if that is my goal, I don't understand how an averaged result helps me.

If you do correction with Dirac , averaged multispot measurements is exactly how it works. If Dirac measured only one spot it could actually do more damage than good.
To start, there is no such thing as a "sweet SPOT" for humans because we have two ears spaced apart. Each ear gets a slightly different sound that in rooms consists of direct waves and multiple reflections. Even for one seat Dirac recommends multispot measurement. It tries to build a model of sound field near the listener(s) not only in the frequency domain but also in time domain. Dirac's filters are a compromise trying to help with least damage possible. They obey Floyd Toole's warning: " ...making in-situ measurements and manipulating the input signal so that the room curve matches a predetermined target shape, imperfections in (unspecified) loudspeakers and (unspecified) rooms are measured and repaired (...) is an enticing marketing story". ( AES paper "The measurement and calibration of sound reproduction systems")

And one more quote from that Toole's AES paper:
" It is a bold assertion that a single steady-state measurement in a room- a room curve-
can reliably anticipate human response to a complex sound field. Such measurements take no account of the direction or timing of reflections within the sound field. (...) Human listeners respond to these cues, in some detail, and they exhibit skills in separating room sound from the timbral identity of loudspeakers and in adapting to different circumstances.(...) Not everything measured is perceptually important, nor can our reaction to such sound fields be constant, we adapt."