Room challenge and DSP

[ahofer]

So here is my system in my weekend place. The two attached photos show the system and living area, and then turn back and show where I was taking the photo (standing on the window seat).

As you can see, nothing but reflective surfaces, including an angled, high, hardwood ceiling. And very few options for system placement. I generally sit in that Morris Chair where I get a decent image *over* the couch. Still, left speaker is next to a window.

I’m considering getting a DSP front end to tackle some of the room effects. Looking at MiniDSP and RME. Having never used DSP before, what do forum members recommend?

I’m also extremely fond of the living area as it is, so I’ve been reluctant to look at room treatments. Hopefully you see why.

Eventually those JBL L830s will be replaced with another pair of Harbeths. They make a good, clean image but don’t have the timbral accuracy or image scale of my Harbeths in my NYC apt.

I appreciate any advice you might have to offer.

PS, yes that is fairly expensive audiophile cable I bought in a moment of weakness in 1993. Don’t hurt me. I just keep using it.

 

[Wayne A. Pflughaupt]

I’d say save your money. DSP can’t correct acoustical issues (other than peak in responses caused by room modes).

Regards,
Wayne A. Pflughaupt

 

[JeffS7444]

Not an expert at these sorts of things, but even if you could only tame some peaks in the frequency response, that can mean the difference between being happy or not with the sound.

 

[Hipper]

DSP/EQ can help for sure.

The first question you need to ask though is 'where do you want the sound to be good?'

Usually with DSP you measure the sound from the speakers with a microphone placed where your ears would be. You can, as the MiniDSP people suggest, take measurements at a number of positions in the area of the main seat, and then you will get a reasonable sound in that area. But if you want it to sound good at all the seats as you have them that becomes harder if not impossible.

Also, as hinted above, your room causes two problems that can be looked at separately. The main one will be in the bass region - you will hear different bass in different parts of the room. Play some bassy music and walk round the room to hear this. DSP can help with this but is often very position dependent - getting it right for one spot does not mean it will be right in another.

The higher frequencies are the ones that are mostly reflected. What you mostly want to hear is the direct sound from the speakers, but on top of this you will get reflected sounds mostly added to the direct sound. How you hear (called psychoacoustics - the workings of the ear and brain combined) tells us how we interpret this. Microphones only hear what's there but cannot tell where the sound comes from. In this case then it is not so simple as measuring with a microphone and making the frequency response flat. Having said that, reviews I've seen of people just using DSP through the whole frequency range are generally impressed with the improvements made.

From this you can see that DSP will not be a perfect solution but I think it should, with careful use, improve your situation.

I'm just wondering though if you wouldn't be better off just moving the two speakers about - not just on the floor but higher maybe - to get the best results. Try any positions, not just the conventional triangle.

 

[Willem]

DSP room equalization is useful for filtering out room modes, but it does not really offer much above the Schroeder frequency of some 100-200 Hz, depending on the room. The hard surfaces are primarily an issue at higher frequencies, and only additional damping will improve the sound. You could think of adding curtains to dampen the reflections at higher frequencies. We have an interior with even more hard surfaces and large windows - our next sofa will not be leather but textile.

 

[Krunok]

I’d say save your money. DSP can’t correct acoustical issues (other than peak in responses caused by room modes).

Regards,
Wayne A. Pflughaupt

Actually it can, and it is very often used in all active systems with digital XO.

 

[Krunok]

DSP room equalization is useful for filtering out room modes, but it does not really offer much above the Schroeder frequency of some 100-200 Hz, depending on the room. The hard surfaces are primarily an issue at higher frequencies, and only additional damping will improve the sound. You could think of adding curtains to dampen the reflections at higher frequencies. We have an interior with even more hard surfaces and large windows - our next sofa will not be leather but textile.

Schroeder frequency is actually in the range of 200-300Hz for typical rooms. Room EQ is also used in setting target responsee curve above that range.

 

[Speedskater]

Forum member Floyd Toole often writes on this subject, both in his book and on this forum.

 

[ahofer]

There are some recognizable bass bumps, but it also a space that rings a bit when you clap.

Listening to music I find the high end a bit harsh. What can be attributed to the JBL horn tweeter and the room is obviously harder to know. We’ve thought about a curtain by the left speaker, but we are very fond of the sun and large view from those glass doors, if not the cold they bring in winter.

 

[ahofer]

Incidentally, a full floor plan and photos can be found on page 100 of the RFK assassination issue of LIFE Magazine. The house has been this way since 1968.

https://books.google.com/books/about/LIFE.html?id=r1QEAAAAMBAJ

 

[mitchco]

@ahofer I have written several articles and a book on using DSP for accurate sound reproduction. For sure it can assist. However, the first thing to do is understand your room acoustics. If you are into it/up for it, purchasing a calibrated measurement mic and using REW to take a snapshot of your room can fill you in on what is going on.

Aside from frequency response, one can use the REW ETC display to see if there any objectionable reflections in the first 20 to 40 ms of sound travel. A waterfall view will help understand the decay time of the lower frequencies. A RT60 (Topt technically) will show you the decay time of higher frequencies to determine if it is too much or too little. Your comment about the space ringing when you clap probably means too much decay in the high frequencies, but need to see measurements to really understand.

One can get very unobtrusive room treatments if required or desired. You mentioned curtains, well, you can also get acoustical curtains like: https://www.acoustic-curtains.com/curtains/acoustic-curtains/ do a great job, even closed up (I use them myself) is one example. Other passive cosmetically products allow prints as well.

Once you have a picture of your rooms acoustics you can also decide if DSP will help. Here is a typical example of a before and after room correction below 500 Hz.

Typical for most room are swings as much as or more than 20 dB variation in the low frequency response below the rooms transition frequency (i.e. Schroeder). It is quite audible. The above example is using Audiolense and took about 30 minutes from measurement to listening to music with the low frequencies evened out nicely. Everything above 500 Hz is left alone.

If you do get a calibrated measurement mic and use REW, post REW's .mdat file here and we can assist you.

Good luck!

 

[Willem]

What a nice house, and great that you could keep it in the family and prevent its mutilation. I had to sell my family's similarly nice modern house because it was in a different part of the country and I could not foresee us ever using it. I am just hoping that in due time we can transfer the modern house we had designed for ourselves to one of our children.
Anyway, now about room acoustics. It looks like a pretty large space, if only because it is such an open plan design. That should help a bit with room modes, but as was already said, get yourself a UMIK-1 and measure it with REW. And if you want to find out how much of the harsh sound is due to the current speakers, all you have to do is take the Harbeths from your apartment and try if they make a difference.
One way to reduce the impact of the room is to listen in near field. This will give you more direct sound and less reflected sound. In this particular room you could, if I interpret the photographs correctly, have a pair of Harbeth P3ESRs left and right of the fireplace. Add two small subwoofers hidden away somewhere and with an Antimode 8033 dsp room eq to tame their room modes, and chances are you will be fine. Of course, if you want to fill the entire living space (and at high spl), something else will be required, although the same idea might still work with the Harbeth M30.1, even if that takes up more space and has to be further away from the wall.

 

[Neddy]

VERY Nice house.
Mine (NOT a vacation house, and not in any magazine, but a FLW student designed it in the early 90s, and is a superinsulated active/passive solar design, with 10" thick walls) has a very similar ceiling/glass/rough size layout - but is more square-ish, and carpeted, and I have very little in the way of acoustic issues (other than the usual room stuff).

Hand claps are nearly dead sounding, for example.

My house does have more angular features (less rectilinear room plan view, see comments about FLW using a 30/60/90 square exclusively for many of his later designs than yours, and a small stub wall angling off to the side of the main room does create some minor reflections from my big JBL's horns.

I also have a motorized/insulated curtain (kind of like Window Quilts) over the 20' of glass wall - oddly, it doesn't seem to make much difference up or down (to audio at least, in winter it's a life saver!)...at least wrt to the REW measurements taken so far.

I do hope to redo all my floors in wood someday - but I guess your experience confirms that it would be a step backwards in audio while a huge step forward in aesthetics
(Hah! - I even have a blue accent wall like the Hofer article showed!)

l also have a Bryston 3BNRB which has continued to work like new from the day I bought in the early 90s - but will be going back for a visit and health check soon...just for the heck of it.

Thanks for sharing!

 

[Fluffy]

Actually it can, and it is very often used in all active systems with digital XO.

Care to elaborate? I don't quite understand how you can null out things like reflections.

 

[mitchco]

Here is one example system using DSP and digital XO's. This is looking at the first 100ms of sound travel of a 3 way active system, including subs in the time domain. The display view is a step response.

Here we can pick out the individual drivers and delays, including a large low frequency room reflection from the sub(s).

After applying DSP time domain correction which a) time aligns the drivers and makes sure each driver has positive polarity and b) where the timing response is aligned to the target response (from the frequency domain) of the ideal minimum phase response of a loudspeaker at the listening position:

Again the first 100ms of sound travel (after the FIR filer delay) with all drivers time aligned and the large low frequency room reflection taken care of. I left in the target response to show how closely the system follows the target from the frequency domain.

From: https://audiophilestyle.com/ca/ca-academy/ integrating-subwoofers-with-stereo-mains-using-audiolense-r712/

 

[Fluffy]

Wouldn't this kind of compensation be only relevant to a specific point in the room? How well does it hold when moving to a slightly different location?

 

[Krunok]

Wouldn't this kind of compensation be only relevant to a specific point in the room? How well does it hold when moving to a slightly different location?

It depends on the initial measurements that serverd as an input. If you took spatial measurement, averaged over some area, like your listening sofa for example, than you will get results that covers that area. That is a preferred method, meaning you should avoid taking single spot measurements.

 

[RayDunzl]

Wouldn't this kind of compensation be only relevant to a specific point in the room? How well does it hold when moving to a slightly different location?

As an example:

Impulse Response at Listening Position (middle of the couch), both stereo speakers active:

The wiggle at 800us is a reflection off the base of the mic or the top of the couch.



And the impulse response from another seat on the couch:

It's not a fault of correction.

The same thing occurs with or without correction, as the distance to each speaker becomes different.


Step response shows similarly, off-axis:

 

[Krunok]

Care to elaborate? I don't quite understand how you can null out things like reflections.

You can correct on-axis response above transition frequency all the way up to 20kHz, and that will help if speaker's resopnse is smooth off-axis as well. If it isn't, well you're screwed, as that cannot be corrected by DSP.

 

[Krunok]

As an example:

Impulse Response at Listening Position (middle of the couch), both stereo speakers active:

View attachment 35227

And from another seat on the couch:

View attachment 35228

It's not a fault of correction.

The same thing occurs with or without correction, as the distance to each speaker becomes different.


Step response shows similarly off-axis:

View attachment 35229

Illustrating FR from a different spot with IR is not really a good example.