Sign in to follow this  
Followers 0
Kvalsvoll

Bulding the Room2 listening room

112 posts in this topic

Just to share some thoughts around speakers and acoustics;

 

Look at what happens to the phase, when early reflections are sufficiently removed - it starts to look very much like the anechoic response:

 

post-181-0-05287000-1473477195_thumb.png

 

The IR also shows that early reflection level is low:

 

post-181-0-62657600-1473477246_thumb.png

 

But the IR also reveals there are lots of later arriving reflections, and they are quite high in level.

The interesting thing to observe here is that the later reflections has little effect on the frequency response and phase.

Both steady-state and transient response is largely determined by early reflections.

 

This means it is possible to achieve a smooth frequency response and good transient response even if the room/speaker combination is quite lively, with lots of room contribution.

 

That is exactly what a large horn speaker does - the directivity reduces early reflection level, and the sound reflected from the back of the room creates lots of later reflections.

This results in a big sound with great clarity.

 

The small F1 speakers need acoustic absorption or huuuge distance to side walls to achieve this.

As the measurements show, it certainly can be done, with the small F1 speakers in this small room, with sufficient absorption.

 

So, does this sound very dead - the IR drops to below -30dB before 0.5ms. 

No, the soundstage is larger, and perception of room information from the recording is better.

 

Share this post


Link to post
Share on other sites

Those are well controlled early reflections!  My current setup is  with screen/center channel in a corner, lots of 'front wall' absorption, and my LCR untreated 1st reflections are all behind me, except ceiling/floor.  Don't want to treat the ceiling, as eventually I want to put in floor-ceiling line arrays, which use floor/ceiling reflections to their advantage.

 

JSS

Share this post


Link to post
Share on other sites

I would still consider absorption in the ceiling, the line array uses reflection from ceiling only at very low frequencies, as the array itself is directional at higher frequencies, and the absorption does not work at lower frequencies.

 

With normal room height and a large line array the ceiling will start to have effect at around below 100hz.

 

So, why bother?

Opportunity to control and tune later reflections coming from the ceiling.

Share this post


Link to post
Share on other sites

If his ceiling and floor are perfect reflectors and the array is truly floor to ceiling, then acoustically speaking, the array will be essentially vertically infinite at all frequencies.  (Or all frequencies where the transducers are less than 1/4 wavelength apart.)  This sounds like a great strategy to me.

Share this post


Link to post
Share on other sites

It is a great idea.

 

But it does not need reflective floor-ceiling to act like a that, it only needs to be large compared to the wavelength.

For bass frequencies it is also a good idea, and now the floor-ceiling will have an effect, and any absorption will not destroy this behavior because the absorption does not work that low.

But since the floor-ceiling problem is eliminated, it may not be necessary to do anything with floor or ceiling.

 

In a practical implementation things are a little more complex, but it works well.

I built some magnestat panel speakers late in the 80's with true ribbon panels, the HF section was a very long, narrow ribbon, delivering a dipole cylindrical radiation pattern. 

Those speakers presented a very realistic soundstage, and only now do I have speakers that starts to bring back some of this sound character.

Share this post


Link to post
Share on other sites

For an 8 foot ceiling as is common in the U.S., large compared to the wavelength goes a bit higher than 100 Hz, but I agree that the effects of the floor and ceiling will diminish as you go lower.  The decision to treat or don't treat the ceiling may come down to other concerns, like managing late arriving energy.  Secondary and later early reflections could also be relevant because the direct sound-field can be scattered into the vertical dimension by objects within the room.  I kind of doubt this will be a big problem though compared to other stuff.

Share this post


Link to post
Share on other sites

I did not calculate - on the 100hz, only a very rough estimate.

Only a simulation/measurement graph can give a true picture of the situation, but a rough estimate is all we need - we understand it is not happening around 2k, it is closer to 100hz, or 200hz.

 

Interesting to see you are thinking about what happens to the sound field when there are objects around.

Just thinking out loud here, but can it be that a directional sound field is more robust for such disturbances. 

 

This is a problem when doing measurements in-room.

Do you leave the listening chair, should the chair/seating be damped.

This has huge impact on the measured response.

 

In the Room 2 I just remove the chair, because it is very easy to do.

In the Moderate Cinema I usually place some damping in the sofa, and adjusting the size, material and shape of this damping makes it possible to achieve a very wide variety of measured responses.

 

Just mentioning, in case someone really believes a +-1dB unsmoothed response is a reasonable target..

Share this post


Link to post
Share on other sites

Part from the dips in the lower midrange, I think the chair is the most limiting factor of this room right now, it simply is not very comfortable..

 

Saw some of the avshowroom youtube-casts yesterday, if for nothing else it is always a good laugh.

Ignorance, religion and incompetence rules in this business, making it hard to get in the market for any serious - or, is it really an opportunity, depends how you look at it.

 

Like this statement from a manufacturer rep:

"Oh, if you do the math on 24-bit audio you see why we need that powerful amplifier.."

 

I can see at least 3 flaws in this statement.

Can you spot any of them?

1 person likes this

Share this post


Link to post
Share on other sites
Like this statement from a manufacturer rep:

"Oh, if you do the math on 24-bit audio you see why we need that powerful amplifier.."

 

 

I think I'll stick with DSD audio.  How many watts is that?  See, I *knew* it was better than PCM.

Share this post


Link to post
Share on other sites

I think I'll stick with DSD audio.  How many watts is that?  See, I *knew* it was better than PCM.

 

Didn't think about that solution.

 

Anyway, I solved it by converting all my music to 8-bit, now I can still use my vintage class-A amplifier.

Share this post


Link to post
Share on other sites

Yes, about that need for a 144dB dynamic range..

 

The answer to this may not be obvious to all, and certainly difficult for anyone to read my mind to see how I am thinking.

 

First, the ? is "need for 144dB" because that is what we have with 24-bit, if we use all the dynamic range available.

Lucky for us, we don't need to, neither do we want to, nor is it possible technically. 

 

1. The "high-end" speakers will compress and distort way before reaching 144dB spl, regardless of how powerful the amplifier is. Now, if we consider a noise floor in the 20dB to 40dB range in a very quiet room, we actually need 164-184dB..

 

2. This spl in the midrange/higher frequencies will lead to instant hearing loss. We don't want that. we don't want sound louder than we perceive as reasonably pleasant, and certainly not so loud that it will cause permanent hearing damage.

 

3. No recording has a dynamic range anything close to 144dB, in fact it is physically impossible to achieve if parts of the production is made from recording voices or acoustic instruments using a microphone.

 

4. No electronics have more than around 120dB dynamic range, this is partly a physical limit, the noise level can not be lowered more.

 

 

So, has 24 bit audio any advantage, is the performance increase audible compared to say 16 bit?

 

Yes, it is audible, but it has no practical advantage in sound reproduction.

 

It is audible, if you play silent audio in 16 bit the noise floor is audible around -3dB MV.

For 24 bit, it depends on the DAC and preamplifier stages, typically noise will be audible around +8 to +12dB MV.

In more practical terms, 24 bit audio lets you hear silence a little louder.

 

When playing music, the music itself will mask the noise floor, and even in the totally silent parts the noise from the recording will totally mask the noise floor of 16 bit audio.

Share this post


Link to post
Share on other sites

Without 24-bit audio, BEQ would not be possible.  I'm glad we have it over just 16-bit on most discs.

 

JSS

Share this post


Link to post
Share on other sites

I agree that 16-bit is more than sufficient for playback media.  A theoretical quantization noise floor of -96 dBZ is substantial, and using shaped dithering, it is possible to push the noise floor much lower in terms of dBA.

 

However, for production intermediates and master storage, 24-bit or higher is preferred.  Inside a DSP, 32-bit float or higher is better.  The concern here is that every time digital media is transformed, there is loss of information.  The result of a computation often requires more bits to store precisely than are available.  These errors accumulate with each transformation and degrade the audio.  Therefore, it is usually better to do a series of computations with higher bit-depth and then convert it back to a lower bit-depth afterwards.

 

Without 24-bit audio, BEQ would not be possible.  I'm glad we have it over just 16-bit on most discs.

 

JSS

 

I don't see why this should be the case.  With BEQ, all you are usually boosting is the level of the lowest bass, where hearing is least sensitive.  Even with noise shaped dithering, the amount of quantization noise here is likely to be miniscule compared to audibility thresholds.  A more realistic problem is the fact that just about every audio sample used in a track that wasn't HPFed somewhere through the process will have a rather high ULF noise level from the source.  For big effects, this is irrelevant because the effects are well above the noise floor, but for recordings of ambience, foley, or dialog, a lack of HPF can allow a lot of unwanted bass noise onto the track, which BEQ may severely accentuate.

 

Now if you are talking about the DSP required to implement the BEQ, then 24-bit is probably not even enough.

Share this post


Link to post
Share on other sites

Internal calc in dsp is always higher resolution, and if you do processing - like bass-eq - it is more likely that you end up in situations where noise is introduced if 16 bit is used.

 

And we all like 24-bit, overkill never hurts, at least not when there is no penalty.

 

Working on the room today, must get rid of the low level and cancellations in the lower mid, but.. not easy.

So, ended up listening to Gary Willis, to verify if more floor absorption is necessary - now flat from around 300hz and up.

Sounds good, but it does not fix the cancellations, and it is not practical and it does not look good.

I think it is not necessary.

Share this post


Link to post
Share on other sites

I don't see why this should be the case.  With BEQ, all you are usually boosting is the level of the lowest bass, where hearing is least sensitive.  Even with noise shaped dithering, the amount of quantization noise here is likely to be miniscule compared to audibility thresholds.  A more realistic problem is the fact that just about every audio sample used in a track that wasn't HPFed somewhere through the process will have a rather high ULF noise level from the source.  For big effects, this is irrelevant because the effects are well above the noise floor, but for recordings of ambience, foley, or dialog, a lack of HPF can allow a lot of unwanted bass noise onto the track, which BEQ may severely accentuate.

 

Now if you are talking about the DSP required to implement the BEQ, then 24-bit is probably not even enough.

 

This is due to the fact that BEQ reduces the entire track in level before boosting LF, typically by 7dB.  All of that data is lost, but it typically is within the noise floor.  24 vs 16 bit allows more volume 'steps' for a given dBFS SPL level keeping the lower SPL bits of the soundtrack a little more out of the noise floor than 16 bit.  

 

JSS 

Share this post


Link to post
Share on other sites

This is due to the fact that BEQ reduces the entire track in level before boosting LF, typically by 7dB.  All of that data is lost, but it typically is within the noise floor.  24 vs 16 bit allows more volume 'steps' for a given dBFS SPL level keeping the lower SPL bits of the soundtrack a little more out of the noise floor than 16 bit.  

 

JSS 

 

I think I understand what you are saying.  Your DSP is a through digital processor.  Realize though that the content you play can be 16-bit without a problem.  Your BEQ processor should  up-convert it to a higher bit depth before doing any processing including the 7 dB attenuation you speak of.  It should also ideally output at least 24-bits to the AVR.  However, I'm not sure if you would notice much difference if it were 16-bits instead.  Even with 7 dB attenuation, I believe it's unlikely for any relevant content to be lost to the quantization noise floor arising from representation of the signal with only 16-bits, but it will be a lot more sensitive to gain matching.  Using 24-bit is definitely better, given the choice.

Share this post


Link to post
Share on other sites

The F2 speakers back in Room2.

 

After temporary fixing the hf with modifications to the horn, the hf is better.

And moving the speakers to the same position as the previous F1 position, improves the sound.

 

post-181-0-48428200-1475971751_thumb.png

 

Still, the low-mid issues remain, and of course the 6" drivers in the subwoofers limits performance, horn loading still operates within physical limits - no magic "extra octave" or "20dB extra dynamic range", but the session listening to Gary Willis at +6dB earlier today was a very good experience.

 

Tactile in the bass is not good enough, but the overall experience is tolerable.

The sound stage extends nice in the fore-aft direction, separation and tonal contrast between instruments are quite good.

 

There is this huge soundstage extending from behind my seating towards well beyond the front wall.

And the contrast between very precise and exactly located instruments up front, to the spacious effects and sounds that fill the whole room.

 

The F1 presents a flatter soundstage with less distribution in depth.

 

The differences has nothing to do with dynamic capacities - both have headroom to spare in this small room.

It has to do with radiation characteristics.

Frequency response is not very different.

Radiation determines behavior in the time domain, and this is what causes the speakers to be perceived as sounding different.

 

The differences are independent of loudness, listening at a reduced volume still reveals the same sound characters.

Share this post


Link to post
Share on other sites

I have used this room to test a new series of compact horn subwoofers for some time now.

 

Several new designs and many configurations, from very small subwoofer systems up to something that approaches the real thing.

 

Part from the fun in testing and listening to all-new designs, here are some more general thoughts around bass in small rooms:

 

- Subwoofers do not sound the same. Even with same type - horn - and the same driver, the tuning affects sound character.

- Calibrating for a perfect and smooth frequency response is not the most important to get decent bass, time related problems are far worse. 

- Tactile feel is different from different subwoofers. Placement and calibration also affects this, but the subwoofer itself has huge significance. 

- A very small subwoofer can have a sound similar to a larger system if they are designed properly and the volume is kept reasonably low.

- Subwoofers both up front and in the back (4 or more) gives much more flexibility for calibration, but is difficult to set up properly, it is easy to end up with worse performance compared to 2 up front only, and just forget doing this without measurements. 

- When you think you have heard it all, and think there is no more to achieve, you set up a new design and realize you were wrong.

Share this post


Link to post
Share on other sites

Much of this is already known - "it is known".

 

We know delay and crossover must be right, we have an idea for how to do this, we know the frequency response must be reasonable.

 

The interesting parts start after all this is roughly adjusted.

 

Such as the character of the bass - a big system sounds better, even at low volume. How is that - certainly more than capacity comes to play here, because the difference is present at low volume, when a small 8" sealed sub should sound similar - according to simple theory. It does not. 

 

The difference is transient response and sound field properties - how the bass couples to the air. A large radiating surface couples better, and creates a sound field with more power and tactile feel, for the same spl. We have also experienced that a ported box has more impact than a sealed.

 

In the new subwoofers I created a larger radiating surface, with high velocity output. It is narrow in one direction - giving the high velocity - and long in the other direction - acoustically large radiator. The horn loading makes it possible to make this in a practical subwoofer.

 

A horn is different from a direct radiator, in that the sound has a higher velocity at the output, and kind of already has turned into a sound wave inside the horn path. The port in a ported box works in a similar way, the difference is in the tuning and that the horn port outputs sound across a much wider frequency range. The horn output works better even when the radiating area is smaller. This means it is possible to make a smaller subwoofer with the large sound.

 

The smallest of the new designs measures 22cm x 22cm x 68cm (divide by 30 to get feet). Very small, and output capacity is also limited.

 

At reasonable listening levels, you get that lively, realistic punchy character of the bass. Which of course makes you want more, so you turn it up, but that doesn't work so well, because you soon reach the capacity limit, the bass does not sound distorted or bad, but you don't get more of that addictive, powerful punch.

 

Jøkleba - album Nu Jøk!, track Popcorn (Can be found on Tidal).

 

When the bass drum comes in, it hits with a massive, powerful impact. Of course, at low volume you don't experience a hard, physical impact, but you can hear it is there. On the larger horns it certainly is very real and present when you turn up the volume, because they have more than 10dB more output, and wider frequency range.

 

One of the largest surprises was the V6030 unit - a larger horn tuned to 30hz, for massive output. Here, the bass drum on Popcorn is very dynamic and powerful, and it does not fade off when you turn it up - loud.

Share this post


Link to post
Share on other sites

I apparently have to build that sound velocity meter....I went from two large horns to a multiple sealed build, and my experiences are a little different, but I never measured the sound velocity differences.

 

The horn-loaded system (for same SPL), had slightly more impact, but the sealed system was nearly as impactful especially once the plinth bass traps were in place, and I moved all 8 of the sealed subs to the front wall.  When I ran the sealed setup with front wall subs + nearfields, the ULF was better, but the 'slam' was not as good.   Currently my room is a huge compromise, and I run a corner-screen with lots of behind-screen/front corner absorption and the subs in an arc with the center of the arc just behind the MLP.

 

I need to build that sound velocity meter and measure. 

 

JSS 

Share this post


Link to post
Share on other sites

8 sealed units makes a much larger radiator compared to 2 horns, it also depends on the horn, the drivers. Differences in capacity will also affect preception of the bass, and often at lower volume levels than you first assume.

 

What you definitely can measure with the v meter is the direction of the sound - when everything is perfect, the direction will be along the normal axis from the front speaker system. At low freq the level has significance, but higher up in the mid and upper bass, intensity may be more important. v measurement can give you an idea of intensity because the intensity will be highest in a directional sound field. 

 

Timing and phase response also has significance.

 

If I understand you correct - adding bass traps improved, and moving all sub units to the front wall improved.

 

The bass traps improves everything, and especially the time domain, so that makes sense.

 

Having all subs up front, close to the main speakers, should make it easier to integrate properly. And easier to achieve a good transient response. Subs all over the place has some very real problems with achieving a good time impulse response, and at the same time preserve sound field velocity and intensity.

Share this post


Link to post
Share on other sites

I need to build that sound velocity meter and measure. 

 

Sound velocity meter? I'd love to measure sound velocity as well as pressure. Tell me more please. :)

Share this post


Link to post
Share on other sites

Sound velocity meter? I'd love to measure sound velocity as well as pressure. Tell me more please. :)

There is a thread about sub calibration & setup, actually two, search for the one with "tactile" in the title.

 

You already have pressure measurement, but of course you already knew that.

 

Velocity measurements are useful to see what you ended up with, if you have sub units both up front and back, it is also possible - to some degree - to change things with calibration.

Share this post


Link to post
Share on other sites

Such as the character of the bass - a big system sounds better, even at low volume. How is that - certainly more than capacity comes to play here, because the difference is present at low volume, when a small 8" sealed sub should sound similar - according to simple theory. It does not. 

 

The difference is transient response and sound field properties - how the bass couples to the air. A large radiating surface couples better, and creates a sound field with more power and tactile feel, for the same spl. We have also experienced that a ported box has more impact than a sealed.

 

Sorry to take it off topic with the sound velocity meter but I really didn't as I've noticed what you described above as well. Why does a 12" midbass have more impact than a 6"/8" midbass when played back at the exact same sound pressure level? I've asked multiple times, at least on AVS and I haven't ever received a satisfactory answer. I was hoping measuring the velocity would help point me in the right direction so I can understand this phenomena better. If you have ideas on the physics, I'd love to hear it.

 

And I'd love a pointer to a thread (couldn't find the one you referred to) on a velocity measurement kit to build or device to buy. Purchase depends on cost as well as if that will help me understand this size difference better. Thanks.

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!


Register a new account

Sign in

Already have an account? Sign in here.


Sign In Now
Sign in to follow this  
Followers 0