DSP/room correction a good idea with Naim stuff? Or bass traps?

I would do everything possible to improve room acoustics with layout, speaker position, soft furnishings, bass traps etc. before resorting to DSP. Then, if you’re not satisfied, I might consider running a sub connected to a pre out via dsp rather than using it to process the signal running to your main speakers. 

DSP will alter a bit perfect data and therefore aways degrade sound quality.

Unless you're willing to fill your room with massive quantities of sound absorption material then no bass trap will be effective below 100hz (where most room bass issues are).

The best solution is speaker and listening chair positioning and matching speaker to room, the hardest thing in this hobby to get right imho.

I've had good responses in my 4.3m x 4m room with ATC SCM19 and Harbeth 30.1 speakers. Both work well without over exciting room modes. Basically stand mount sealed or front ported speakers will work best in problem rooms.

Jason from Naim was asked at a ND555 demo.......if the door is a problem sound wise, change the door or do something to improve the door/acoustics. Whatever you do , do not mess with the sound signal........a Roy George mandate.

I’ve recently done an experiment with Dirac on a laptop. In that room, the improvement was quite impressive. It takes less than an hour to download the Dirac demo version and calibrate the room. Then you can compare by playing some music with and without Dirac and decide if the impact is positive in your room.

It would be interesting, in this thread, if anyone who has tried SPACE from their source in an otherwise predominantly NAIM system, could contribute their experiences.

I see no problem in using dsp for bass frequencies ..... I use a small velodyne dd10+ sub from output of my 552 ..... it is fantastic. The main speakers still get an unaltered full fat signal ..... the trick is to use small speakers and augment subtly ... my little pmc 20.21s sound fantastic.....this on the end of a 500 system..you must not oversize your speakers in problematic rooms. Another approach would use say an nd555 into Kii fully active full dsp speaker....there are allot of positive reviews on this revolutionary speaker.

You can also use the DSP in Roon, although as a sort of glorified digital graphic equaliser, I’m a little concerned that there will be a price to pay in sound quality. With a new Naim streamer, it would be useful as a means of assessing the effect of reducing any bass nodes caused by the room. 

Reading about the new Linn range last night, also appears their room correction in built to the unit has been updated, the journo's thought it was good at a lock in controlled demo.

Gazza posted:

Jason from Naim was asked at a ND555 demo.......if the door is a problem sound wise, change the door or do something to improve the door/acoustics. Whatever you do , do not mess with the sound signal........a Roy George mandate.

Absolutely, if you understand to what extents Naim design their implementation of a relatively  modest single low pass filter DSP function in their DAC/Streamers so as to cause minimal signal distortion and side effects, I think that engineering effort would need to be multiplied severeal times to have a cascade of filters required for room correction so as to cause minimal distortion. Each filter also adds digital noise albeit if well designed at very small levels.

if one was going to do this I suggest the the best place would be as part of the DACs inline DSP function where one has already oversampled and increased the sample word size so as to reduce the impact of arithmetic errors.

Sadly D48Rs will always reveal the bass resonance of that room unless you use quite dramatic DSP filtering (which will also induce unwanted side effects).  It will be best to use much smaller speakers and a sub that's controlled by a DSP; even then to completely eliminate the bass resonant peak, you will still need bass traps (but this way they can be a fraction of the size they would need to be to handle the output of D48s at the resonant frequency of your room).

The other thing you'll find of considerable benefit is to do is use room treatment (such as diffusers and absorbers) to deal with the reflections.  Dealing with acoustic problems in a room with acoustic solutions (i.e. eliminating the issues where they occur) will always give a better result than trying to compensate for the acoustic problems by mucking about with the electrical signal.

Some people seek to avoid bass tesonances by choosing speakers that simply don’t go low enough to be a problem - but if you value the bass in the music that clearly is not an option. Otherwis, particularly when considering existing speakers that you like and don’t want to change, you should seek to optimise speaker and listening position first, then consider room treatment, and only when you have done whatever is possible and domestically tolerable DSP may be of assistance to improve things if significant problems remain.

The balance then is any negative effect of the DSP on sound quality, which may or may not be evident to you, vs the negative effect of room issues. Reducing bass where there are peaks in response at the listening position can be quite effective, and to some extent noosting dips, but beware that boosting  can much more seriously affect system performance, and can even Damage speakers, while if a the problem is a cancellation effect it will never actually be able solved by boosting.

The reason boosting is not adviseable other than only slightly is this:  say you have a 12dB dip - by no means an extreme example: To raise the level at that frequency by 12 dB means 8 times the power, so if you are  playing at a level where crescendos use any more than one eighth of the power amp’s maximum peak power capability before clipping, boosting by 12dB would mean any peaks in the music at that frequency  would take the amp into clipping on those peaks, risking damage to speakers and possibly amp.

IB, that's a very good point about not trying to use DSP boost nulls and dips, however you underestimated the problem...
Raising the level by 12db is 16 times the power (not 8 times).

What you say about using positioning and acoustic treatment of the room before final adjustment with a DSP is also very valid.

In terms of using small speakers, there is another option to loosing bass performance.  Use small speakers that don't excite the resonant modes of the room, then fill in the bass using a sub controlled by a DSP that reduces and evens out the resonant bass frequencies.  This way you get the advantage of DSP based control of the room modes without having the DSP filtering causing artefacts at the rest of the audio spectrum, particularly in the mid range and lower treble where the ear/brain combination are particularly sensitive to tiny changes in audio quality.

Wouldn't DSP applied to the sub feed introduce a small delay? if so, wouldn't an identical delay need to be applied to the main output to keep the bass signal in phase with the main speaker output?

Huge posted:

Raising the level by 12db is 16 times the power (not 8 times).

.

Indeed you’re right, Huge - head must have been cloudy as the maths isn’t exactly challenging!

Allan Probin posted:

Wouldn't DSP applied to the sub feed introduce a small delay? if so, wouldn't an identical delay need to be applied to the main output to keep the bass signal in phase with the main speaker output?

There's a greater delay caused by the group delay of the Class D amp in the sub (typically about 2.5-4ms delay).

The answer to this is simple...
Position the sub closer to the listening position (typically about 1-1.5m closer to the listener than the main speakers are will be about right).

Allan Probin posted:

Wouldn't DSP applied to the sub feed introduce a small delay? if so, wouldn't an identical delay need to be applied to the main output to keep the bass signal in phase with the main speaker output?

I have an N-Sat + Velodyne sub combination fed by my Nova and there is no audible delay. It works very well in my 5mx5mx2m room (less than ideal!) with no treament. Bass is excellent and I do sometimes play dub reggae really LOUD!

Pev, I'm thinking possible delays in the range of a few milli-seconds which wouldn't be immediately audible as a delay as such but would be more a loss of coherency and timing. I would have a concern that a bass fundamental would end up out of phase with its harmonics.

Another way of applying DSP for those who want to tri-amp (pr bi-amp) their speakers is to use a digital active crossover with capability beyond simply splitting the signal - which can be used to tailor response in-room, regardless of whether peaks or (small) dips in response are caused by the speaker or the room.

Yes, it's the coherency through the sub's crossover region and the timing that are affected and audible (rather than the phase of the harmonics, which is a much less audible factor as the phase relationship of the harmonics of any sound source naturally varies with the distance to that source).

That's why having the sub 1-1.5m closer to the listener works better than having it co-located with the main speakers.

Huge posted:

the phase relationship of the harmonics of any sound source naturally varies with the distance to that source).

Huge, that's interesting, I haven't come across that before. It seems more intuitive to me that for a point source the phase relationship of the harmonics with the fundamental would be fixed, regardless of distance. Would it be possible to give a simple example with numbers?

OK no problem

Take a sound with a wavelength of 10m and consider the 2nd harmonic (which has a wavelength of 5m)

At source the waves are in phase,
at 2.5m the 10m wave is at +90°, the 5m wave is at +180°
at 5m the 10m wave is at +180°, the 5m wave is back to +0°
at 7.5m the 10m wave is at +270°, the 5m wave is at +180°
at 10m both waves are back to +0°

(All phases expressed relative to the source, at the same time.)

How do room correction measure the response?  If you measure the integrated response and adjust after this then you change the direct/early reflection sound. (which is not affected by the room).

Or is the room correction boxes smarter than I think they are?

I have a diffusor behind the normal listening position aand absorbents in back corners.

Huge posted:

OK no problem

Take a sound with a wavelength of 10m and consider the 2nd harmonic (which has a wavelength of 5m)

At source the waves are in phase,
at 2.5m the 10m wave is at +90°, the 5m wave is at +180°
at 5m the 10m wave is at +180°, the 5m wave is back to +0°
at 7.5m the 10m wave is at +270°, the 5m wave is at +180°
at 10m both waves are back to +0°

(All phases expressed relative to the source, at the same time.)

Ok, Got it! Thanks very much

Huge posted:

Sadly D48Rs will always reveal the bass resonance of that room

yeah, these speakers. Gotta chew what you say.

So, you think this stuff is kinda called out for cos of these 2 babies?  A few pieces like diffusers you mentioned. But the bass traps especially important?

(I've got IsoAcoutics Gaia IIs for the spkr feet.)

Thanks

Allan Probin posted:

Wouldn't DSP applied to the sub feed introduce a small delay? if so, wouldn't an identical delay need to be applied to the main output to keep the bass signal in phase with the main speaker output?

Although it will depend on digital filter design and type the extent of the delay, which will be more typically manifest as a phase error. So yes if using digital filters in this way, you are effectively creating a digital cross over and one should use the same filter kernel sizes (taps) or equivalent for both the low pass and high pass filters.