2 minutes ago, acg said:

The VPR's are completely un-skinned, and the next move is to measure the room and try to figure out what acoustic properties the VPR skins will need.  Theoretically this should all be 200Hz and higher absorption, which is way easy to provide, and some diffusion too.

Yes, I hadnt thought of that but i guess they can go behind typical absorbers and work well still or mounted on the face even if it doesnt affect the effectiveness to much.

 

3 minutes ago, acg said:

Having a room that has lower reverb at 60Hz than 5kHz is certainly unusual. 

There is nothing usual about your setup. Its extraordinary!!

2 hours ago, frednork said:

To put my cards on the table I think somewhere around 250ms or slightly more for a typical domestic room is ideal for me.  I just think it sounds better and I like to hear reverberation in the recording more than the reverberation created by my room and the longer your times get, the more your room is masking your recording. And it will mask the sound of the space in the recording. 

 

Thanks for all your thoughts, Mark. Much appreciated.

 

SQ-wise, based on the measurement of the differing treatments I have applied in my small, dedicated listening room, I would have to say I am firmly in this camp, as well.

 

The size of the room [3.65m (w) x 3.92m (l) x 2.97m (h)] , and my desire to get my speakers (Lenehan ML2+Rs) out from the front wall, and my listening position as far away from the back wall as practicable, means I am listening in a quasi near-field type arrangement, and I can certainly hear more of the recorded reverberation when RT is lower across the board.

 

I do need several diffusers to keep some liveliness in the room, though.

 

Keep up the good work.

Yes its definitely a balance thing. I think I can do better and get a more even rt60 with a little effort.

 

Just a reminder to those watching on, the reason I am putting lots of diffusion behind the speakers is due to the dipolar design. For monopoles it might not be so great but interested in your experiences. 

18 hours ago, frednork said:

Ok, another thing I was going to mention is what is a good RT60 target? Well if you read things like this forum and others you might say anything from 200ms to 600ms.  Reason is that room size and reflectivity vary and so "optimal" RT60's do vary also and just to add a final gazumper to it all, It also depends on your taste. Ok, so anything goes clearly. 

These are just some random thoughts I have on the topic.

RT60 is not a good metric for a small room (it's too long) ... but, you are pointing towards the crux of the small room accoustics problem.  It is all very important.

 

I am more a fan of shortening the listening distance, and leaving the reverberation time slightly higher......  but it is complicated because often you are fixing peaks in the energy vs time (which can be very audible) vs simply hearing "reducing reverberation generally and evenly".

 

18 hours ago, frednork said:

I havent done anything to prove what I say but am just putting it out there for you to consider it and see how that might fit or challenge your ideas on the topic. Am sure some will come in and tell me how wrong I am.

I think 400ms is fine ... and would typically tune it by having a conversastion in the room (as SL mentioned in the bit you quoted) .... but I don't think you're "wrong".   It is all very important.....  energy vs time, and the amplitude vs frequency balance of that delayed sound, it matters immensely.

 

18 hours ago, frednork said:

So I guess I should revise my previous outrageous 250ms target and say depending on the music you listen to. But still not sure about that.

I would be looking to target specific things you see in your data, as opposed to "general reduction".   I would start zeroing in on the issue below ~300Hz.

5 hours ago, davewantsmoore said:

I think 400ms is fine ... and would typically tune it by having a conversastion in the room (as SL mentioned in the bit you quoted) .... but I don't think you're "wrong".   It is all very important.....  energy vs time, and the amplitude vs frequency balance of that delayed sound, it matters immensely.

 

My theory is that reverb time is also linked to the loudness at which playback is bearable...lower reverb times enables louder listening, at least to an extent. 

On 20/02/2023 at 1:32 PM, acg said:

 

My theory is that reverb time is also linked to the loudness at which playback is bearable...lower reverb times enables louder listening, at least to an extent. 

I hadnt really thought about it as a loudness related thing but expect you are right here. volumes can go up in a well treated room without causing harshness and "whiteout". This can also reveal things at low levels in the recording if your RT's are shorter. If I had played at the volumes in the video below untreated it would sound a bit harsh and too much. The recordings were done at around 88-89dB average slow C weighting. You can hear in the vid how the sound changes as I look about and then locks in more in the listening position (as much as a crappy phone will allow anyway).  

 

On 16/02/2023 at 1:56 PM, davewantsmoore said:

Can the LP be out from the wall?

Can you do a specrtrogram with the mic out another 1.5m vs where it is now?

didnt do 1.5 m out from current as that would be 2m from wall behind speakers and they are 3.2m apart but did do forward a bit and behind a bit. Which do you like the best?

 

33 minutes ago, frednork said:

Which do you like the best?

Frequency responses don't show you useful information.

 

ETC, decay, specrogram.    Ideally with corrected ("nomrmalised") frequency response.... so then you can really see the difference between "treatment A" and "treatment B" when a similar SPL is delivered.

On 20/02/2023 at 1:32 PM, acg said:

My theory is that reverb time is also linked to the loudness at which playback is bearable...lower reverb times enables louder listening, at least to an extent. 

This is "true" .... but "reverb time" doesn't give us the info we need.   It is where the loud peaks occur in the ETC which is the issue.    RT is a proxy for this, in so far as if you kill those peaks, you will lower the RT measurements...... but the converse isn't true, in that you can have a long(er) RT measurement, which doesn't have conspicuous peaks in the ETC.

 

Looking at frequency responses and RTs can be fairly misleading, vs what we are actually hearing (and wanting to fix) .... and it is fairly likely that most "I like a low RT" is really just "I like to get rid of reflections in the time window and SPL where they're audible".

 

1 hour ago, frednork said:

I hadnt really thought about it as a loudness related thing but expect you are right here. volumes can go up in a well treated room without causing harshness and "whiteout".

The audiblity of reflections .... (even, after their impact has been EQed out of the directly arring sound as much as practical) ..... depends on their SPL.     One they drop below a SPL range, they become much more easily masked by other SPL.....   this creates an effect like what Ant theorised.   Say your overall SPL is 80 and the peak reflection is 55dB, it gets masked ..... but at 100dB and 75dB, the reflection is much less masked.

 

The same effect applies to (non-linear) distortion.   The new tones added ("THD", "IMD", whatever we call them) are much more easily masked when the absolute SPL is low ..... so they are more audible (even at the same % of the overall SPL) as the SPL rises.

Edited February 25, 20233 yr by davewantsmoore

6 hours ago, davewantsmoore said:

Frequency responses don't show you useful information.

 

ETC, decay, specrogram.    Ideally with corrected ("nomrmalised") frequency response.... so then you can really see the difference between "treatment A" and "treatment B" when a similar SPL is delivered.

Err, sorry, didnt pay enough attention

 

how about these

 

ETC

 

spectrogram

 

 

 

 

hard to say.

 

13 hours ago, frednork said:

spectrogram

Tick normalise to peak at each frequency (?)

Show 100ms.

12 hours ago, davewantsmoore said:

hard to say.

 

Tick normalise to peak at each frequency (?)

Show 100ms.

Sure!

I think use a 100ms window @frednorkand a longer time axis to suit.

It's out of blue and green for me... but I'd really want to be looking at this measurement with an actually corrected (as opposed to normalised) SPL.

Which is which?  😁

Before I reveal would you mind explaining why the red is least preferred?

I’ll take Red please. 
the high spl in the other two is smeared over greater time delay (reverb) for both low and mid frequencies. 
With the Red spectrogram setup you could further target narrower frequency bands with additional room treatment for mid-frequencies and possibly additional subs to even-out low frequencies. Then, depending on results,  DSP can be used to fine-tune the remaining peaks if required. 

6 hours ago, frednork said:

Before I reveal would you mind explaining why the red is least preferred?

I'm only looking in the 200 to 600 ish Hz range.... and seeing more energy fall inside what a "dirac" specragram envelope looks like .... and more engery after 25ms ..... ie. there is less energy between the edge of the dirac envelope and ~25ms.    Because we want the direct sound, then a ~25ms gap, then the reflected sound.

 

Are these 3 measurements just of the mic position moving, like you said?

1 hour ago, davewantsmoore said:

…….

Are these 3 measurements just of the mic position moving, like you said?

Where was it said that the mic was moved?  Tried to find it but missed it. 

I thought this was room treatment/furniture. 

Neuroscience studies tell us that about the 1st 10msec of sound the brain processes as “single location or direct energy” - it’s not an exact measurement. 
After that the brain interprets the data as reflected or indirect energy/sound using timing delays and pressure variations via the ears, head movements, sight etc. 
@davewantsmooreIs the Dirac spectrogram you mention based on the ~10ms of “group” energy? 


 

1 hour ago, davewantsmoore said:

I'm only looking in the 200 to 600 ish Hz range.... and seeing more energy fall inside what a "dirac" specragram envelope looks like .... and more engery after 25ms ..... ie. there is less energy between the edge of the dirac envelope and ~25ms.    Because we want the direct sound, then a ~25ms gap, then the reflected sound.

 

Are these 3 measurements just of the mic position moving, like you said?

Yes, as far as practicable, I cant account for random background noise that might be incuded but there was nothing obvious at the time. It is likely that the treatment has changed a little since the previous set but all 3 had no change except for mic position and were done within a 15 minute time window.

 

40 minutes ago, frankn said:

Where was it said that the mic was moved?  Tried to find it but missed it. 

I thought this was room treatment/furniture. 

 

It was a fair way back but didnt have time to do it then. Also I didnt mark it on the graphs bcos I know dave likes to blind view his graphs 😜 , and it makes it more fun.

On 16/02/2023 at 1:56 PM, davewantsmoore said:

Ima just goin keep derailing your thread 🤣  (until you say no, of course)

 

Can the LP be out from the wall?

Can you do a specrtrogram with the mic out another 1.5m vs where it is now?

 

Of course it probably can't for daily living .... but you can easily setup two profiles on your signal calibration.

 

Nice effect at ~150Hz in the 50 to 100ms range.... which is part of why I am interested in moving the LP.  

40 minutes ago, frankn said:

I thought this was room treatment/furniture. 

It is, but it was a request from Dave as he proposed that things may be improved by a change in listening position and considering he has been very generous with his time in this thread I dont mind doing that. And I am always interested in improvement. (especially for free!!).

 

Yes, if you move the microphone then the density, amplitude and decay times will vary.  Thank’s for confirming you moved the mic. 
Ethan Winer has discussed  measurements, small variations of microphone position (in the context of using Dirac and other DSP systems to optimise the “sound field”) and, the sometimes unintentional downsides.  
I think we agree that it’s better to treat the room as much as practical (practical is a varying amount for individuals and their particular circumstance) and “ask” less of DSP. 

http://ethanwiner.com/dirac/dirac.htm

58 minutes ago, frednork said:

and it makes it more fun.

🙂

58 minutes ago, frednork said:

It is, but it was a request from Dave as he proposed that things may be improved by a change in listening position and considering he has been very generous with his time in this thread I dont mind doing that. And I am always interested in improvement. (especially for free!!).

👍

 

1 hour ago, frankn said:

Is the Dirac spectrogram you mention based on the ~10ms of “group” energy? 

It is the spectrogram of a dirac pulse.... so, if the frequency response was flat and there were no reflections.

 

Something like, this:  https://imgur.com/7Oz2EDy (im in my office right now, so no audio tools)

44 minutes ago, frankn said:

Yes, if you move the microphone then the density, amplitude and decay times will vary.  Thank’s for confirming you moved the mic. 

Yep, that was the point of moving the microphone.   To try and help the region where the response is the worst, through LP placement.

5 hours ago, davewantsmoore said:

I'm only looking in the 200 to 600 ish Hz range.... and seeing more energy fall inside what a "dirac" specragram envelope looks like .... and more engery after 25ms ..... ie. there is less energy between the edge of the dirac envelope and ~25ms.    Because we want the direct sound, then a ~25ms gap, then the reflected sound.

 

Are these 3 measurements just of the mic position moving, like you said?

I have been looking at that spectrogram for the last 10 minutes and going to myself I have no idea what he is looking at but then suddenly I saw it. Eureka! So indeed the red one follows that dirac shape less so than the other 2. Would love to see a real life  version of this , ie how good can it get IRL?

 

So the big reveal, Red was behind the current LP by about 50 cm, The Green was the current LP and the Blue was about 50cm forward of it.

 

 

 

6 hours ago, davewantsmoore said:

I'm only looking in the 200 to 600 ish Hz range.... and seeing more energy fall inside what a "dirac" specragram envelope looks like .... and more engery after 25ms ..... ie. there is less energy between the edge of the dirac envelope and ~25ms.    Because we want the direct sound, then a ~25ms gap, then the reflected sound.

 

Are these 3 measurements just of the mic position moving, like you said?

@davewantsmoore thank you for the “dirac pulse” snapshot- its similar to a Gaussian distribution of the amplitude decay, tailored by the frequency response (or wavelength).  Which is what would be expected in a free field response. 
Can you clarify the comment “and more energy after 25msec “ ?  The Dirac pulse link doesn’t show an increase in energy after a significant delay - it shows a tapering of spl/time which would be expected (the only frequency content that increased after a significant delay appears to be a very narrow frequency range - e.g an artefact). 
Also, can you expand on why are you focusing on single octave of 300-600 Hz response?  The bandwidth an octave or two higher than this and the octave from 175-350 Hz are worse than the Red  in one of the other positions.  I’m trying to understand your preferences. 
Does this lend itself to a seperate thread or a PM so not unnecessarily cluttering  up this discussion?  
thanks.