Changing The Posted Graph Parameters...

[bossobass]

Not to throw a wrench into the works, but I'd like to suggest a different frequency bandwidth on the posted graphs.

200Hz is less important and the first few octaves are more important, IMO. Even if a sub doesn't reach to single digits, being able to see if the sub is rolling off or if there's just a dip below 15Hz, as well as what the ultimate roll off is, is much more important to me than what the sub does from 100-200Hz.

As an example, here is a measurement before any tweaks with the area boxed in that would normally not be shown, if the graph stopped at 15Hz going down:



In this case, it might have appeared that the sub was rolling off sharply below 15Hz. Also, some tweaks affect VLF that would otherwise not be noticed.

Of course, it's only my opinion, and if it's too much problem, never mind, but I thought I'd ask.

Bosso

[Sonnie]

The problem with posting graphs below 10Hz is we have no idea how accurate they are, since we have no correction values below 10Hz for about 95% of our members.

Generally we post what we can adjust, thus the reason for 15Hz to 200Hz. But I'm open to the suggestion of 10Hz to 120Hz or similar if we can really justify it and others (the majority) agree.

[brucek]

There are a couple of issues here.

The vertical scale we normally use brackets the standard 75dB target level used with REW. The scale used is from 45dB to 105dB. This allows a reasonable min/max of 60dB centered on 75dB. Consistency is important for comparisons, so we like to stick to that scale.

The horizontal scale we normally use is 15Hz to 200Hz. The 15Hz is optional to 10Hz for those with an IB etc, but anything below that is not meaningful, since most people have a microphone calibration file that ends at 10Hz. Measuring below that would require a fairly accurate mic with an associated cal file.
The 200Hz was chosen to accommodate a standard value (where in association with the recommended vertical scale), an 80Hz crossover target would just fall below 45dB before 200Hz. This also accommodated adding the mains speakers to check for crossover interaction without changing scales.

We have found that the vertical (45dB -105dB) and horizontal (15Hz - 200Hz) scale to be a good overall compromise so that everyone is comparing apples to apples.







brucek

[bossobass]

I see the dilemma. The reason I shifted the horizontal axis is that it gets a bit scrunched up moving further up, since I'm adding so much to the left on a log scale.

The vertical is not a problem, I just shifted this one to show the loop back of my digital interface, which is flat to 3Hz. I also have a very good mic that's flat to 2Hz, so no cal file is needed for subwoofer measurements.

I guess there's no easy answer other than a section just for subsonic capable subwoofers. It's not just my own graph and comments that will follow, I see a lot of graphs of subs that look strong to 15Hz.

For example, the JL Audio F-113 thread. I personally have been very interested to know if there's a HP filter on that subwoofer, and if so, what order and frequency. The posted graphs could indicate that there is a rather steep roll off below 20Hz, or that there is simply a dip at 15Hz in that particular room.

I understand the idea of uniformity in posted graphs for quick comparo, but it's not much of a comparison if one sub reaches 3 full octaves below another sub, but no one knows that by looking at 2 graphs with a 15Hz cutoff.

I appreciate the responses either way.

Bosso

[Sonnie]

You can post a graph down to 2Hz, 5Hz, 10Hz, if you want... several people do... especially in other areas besides the BFD | REW Forum. 15Hz is not mandatory by any means, merely a request for this forum area, as well as the "request" for the stated axis parameters is only posted in the BFD | REW Forum.

In general, our "request" for the stated axis parameters are for those who are using the BFD and asking for help smoothing/taming their response. We are not necessarily comparing graphs in the BFD | REW Forum to see whose sub has the best extension. And as previously stated, most of us can't do anything with the response below 20Hz anyway.

I actually encourage others to post lower graphed responses myself, in the other forums such as the DIY Subwoofers and Home Audio Subwoofers forums. I like seeing extension response graphs.

[Chrisbee]

Lean on Bosso to do a calibration of a small sample of Galaxy 140s against his 2Hz mike from 2-30Hz to check for infrasonic variability and to build a proper calibration file.

Sonnie can then spread the 2Hz test limit rather effortlessly worldwide by distributing 140s with the calibration file available from the Shack.

We can call it the Bosso/Sonnie infrasonic calibration standard if bribery is absolutely necessary.

Anybody buying a Galaxy 140 will be assured that they can test down to 2Hz with reasonable accuracy and repeatability.

It could be a milestone in subwoofery.

[terry j]

In the few odd occasions where it is important for the discussion at hand it would be OK to post what is relevant?? I don't think anyone would be upset, I think that the only time it becomes frustrating when 'incorrect' graphs are posted is when it is obvious the person posting them has not bothered to read the help files??

In all other circumstances a 'non-standard' graph is OK if it is for a reason and useful. Surely??

[Sonnie]

You are correct terry.

I'll be happy to forward a meter to Bosso for the ultra low frequency calibration. I would even send him my ECM8000 and once I get it calibrated to 2Hz then I can check whatever then.

However, Bosso, do you have a sub capable of reproducing 2Hz well enough to perform a calibration trace?

PM me your address and I'll get both units right on out to you.

[brucek]

a 'non-standard' graph is OK if it is for a reason and useful. Surely??

Sure, don't see why not.

Here's three plots of the exact same mdat file.

Consistent and meaningful scaling is quite important for interpretation.








brucek

[bossobass]

Sonnie wrote:

You are correct terry.

I'll be happy to forward a meter to Bosso for the ultra low frequency calibration. I would even send him my ECM8000 and once I get it calibrated to 2Hz then I can check whatever then.

However, Bosso, do you have a sub capable of reproducing 2Hz well enough to perform a calibration trace?

PM me your address and I'll get both units right on out to you.

The amps are down about 6dB at 3Hz. The sub is probably dead flat to 8Hz, anechoic. The in-room response can be made flat to 3Hz (+/-)3dB.

What's a Galaxy 140?

Include detailed instructions and sure, I'll be glad to give it a whirl.

Bosso

[Sonnie]

The Galaxy CM-140 is the latest SPL meter/mic that several members are ordering now instead of the old RS meter.

Basically you would need to be able to setup a nearfield sweep (or a flat response sweep) from 2Hz-5Hz up to about 200Hz. You would take one measurement with your pro mic and then another with the CM-140 and another with my ECM8000. Then use the "trace" function in REW. I'm assuming you are familiar with REW, if not, then we can walk you through it all with no problems. The major key is keeping all mics as close as possible to the exact same position/location for each measurement... more particular the mic capsule location.

Do you have a mic amp with phantom power supply, which the ECM8000 requires?

If you are interested, PM me your mailing address and I'll ship the ECM8000 and CM-140 to you.

[terry j]

yeah understand that Bruce, and it is especially obvious in the vertical scale, tho it will stretch when we go higher in frequency too.

But if say Bosso wanted to go down to 5 hz, could we not preserve the feel of the graph by only going up to (say) 150 hz?? I'll leave the actual upper figure to someone else who understands the logarithmic scale better than I.

I would only expect to see the altered scaling quite rarely, all other posts should stay the same.

[Blaser]

I am wondering how many times we can find info below 15Hz Why care so much while most information and most noticable effects will be much above that?

B Rgds
Blaser

[Chrisbee]

blaser wrote:

I am wondering how many times we can find info below 15Hz Why care so much while most information and most noticable effects will be much above that?

B Rgds
Blaser

Because these frequencies exist in considerable quantities on many film DVDs and some music CDs. My own 16-46 SVS doesn't start to roll off until 16Hz. My IB goes on down into the murky depths. I consider these low frequencies rather more useful than ultrasonics and supertweeters even if they are not usually directly audible.

[Blaser]

Chrisbee wrote:

I consider these low frequencies rather more useful than ultrasonics and supertweeters even if they are not usually directly audible.

So is there a huge difference between having these frequencies and not.... If so, could you let me have a concrete idea.... In a blind test with and without these Ultralow freq. can you recognize them? Which movies for ex. contain useful info below 15 Hz, and is missing these freq. noticeable?

Don't misunderstand, I am just asking, no offence!

Thanks!

Blaser

[Chrisbee]

There are long threads on several forums showing Spectrum Lab waterfall graphs of extremely low frequencies or very high levels or both. Both DVD films and CDs reach sub 10Hz more frequently than you might realise.

So far very few subwoofers are known to have useful output at frequencies and levels which can make sub 10Hz sounds audible. The Eminent Technology/Thigpen rotary subwoofer driver is one of these very few.

Very large multi-driver IB subwoofers with sufficient cone area can produce prodigious levels of infrasonics. Though not quite in the same league as the Thigpen fan subwoofer in present sizes. It takes at least 8 x 18" drivers to bend concrete floors and move the furniture about, I believe. At least one 16 x 18" IB is known to exist. It can move the sofa carrying its excited occupants bodily across the room.

Even if such low frequencies are not directly audible they are known to affect mood.
Very low frequencies can cause euphoria, fear or discomfort and have been used for this purpose both consciously and subconsciously for hundreds if not thousands of years. Very large musical instruments, cathedral organs, drums, gongs, bells, waterfalls, surf, earthquakes and storms are all common sources of infrasonics.

Deep, loud bass seems quite addictive judging from the millions of subwoofers sold worldwide. Not to mention the endless online discussions by the happy owners of their latest bass high. (or should that be "low"?) The SVS grin was once a common subject on many forums.

Loud, deep bass from high performance subwoofers can have much the same effects as fair rides. Excitement, fear and elation are common. Add the desire to go on repeating the exercise until the excitement is finally exhausted and you get the picture. People are apparently willing to risk damage to their homes and their reputations with their neighbours in their search for bass nirvana.

Sensitivity to VLF is variable between human subjects and may be affected by the stress of testing for response confirmation. In blind tests some subjects detected remarkably low frequencies but missed others higher in frequency. It is still largely unknown territory since so few commercial subwoofers can provide such low frequencies at sufficiently high levels to make them commonplace.

[Blaser]

Well!! Good luck! I think I have never experienced such thing... I am still in research of a good 16-20 Hz area only

Thanks for your long, informative post.

Blaser

[DrWho]

Ya know....the data doesn't change when the scale on the graph changes. Not to be rude, but I really gotta question the ability of a person to interpret data if they feel a certain scaling is required. I totally agree that a standardized scaling makes it easy to cast a glance and make some judgement calls, but this whole obsession with the frequency response seems a bit crazy considering all of the other cool features built into REW. , just change the window for the FFT and you'll get +-6dB no prob on the calculated frequency response...

Ok, enough ranting...

One last comment....I would argue that the frequency response needs to include higher octaves to better show the transition from mains to subs. Comparison to the 4kHz region is much more informative than the 200-600Hz region...

[Sonnie]

I really gotta question the ability of a person to interpret data if they feel a certain scaling is required.

I wouldn't... ... but let's clear this up so we have no misunderstanding. The scale is not "required", it is "requested" and again it is more for when someone is requesting help in the BFD | REW Forum for smoothing their sub response. It's not that any of us cannot interpret various size scales, it's more for the sanity of those that look at these things day in and day out and respecting what they do and what might help them give the fastest response.

[Chrisbee]

blaser wrote:

Well!! Good luck! I think I have never experienced such thing... I am still in research of a good 16-20 Hz area only

Thanks for your long, informative post.

Blaser

If you thought my rambling above was long you should read about the TRW subwoofer in this interesting article.

Everything you ever wanted to know about the reproduction of infrasonics but were afraid to ask.

http://www.iar-80.com/page142.html

[brucek]

Not to be rude, but I really gotta question the ability of a person to interpret data if they feel a certain scaling is required.

That is certainly one of the more bizarre statements I've read on this forum.

Oh well, you can't please everyone. My thirty two years working in electronics have taught me to appreciate how important scaling is in graph interpretation, especially for the inexperienced. This forum tries to cater to a fairly broad cross-section of people.

Many posters have very little experience with interpretation of the results that REW produces. This is evidenced by the constant requests for advice on just that subject. By requesting a scaling standard for posting subwoofer graphs, I think it makes it easier for the novice to learn and better interpret the output of REW. The scaling we chose is completely appropriate for a dataset that represents the usual vertical swing and horizontal extension of subwoofers. It's a scale that couches most situations we encounter. The fact that someone could choose a ridiculous scale to exaggerate a reading and subsequently perform a correct interpretation, doesn't seem like something that should be required to garner your respect.

Everyone who posts here can't be an expert in graphs such as yourself. Hopefully you can use your expertise to help others, even though they might be novices.

brucek

[bossobass]

blaser wrote:

So is there a huge difference between having these frequencies and not.... If so, could you let me have a concrete idea.... In a blind test with and without these Ultralow freq. can you recognize them? Which movies for ex. contain useful info below 15 Hz, and is missing these freq. noticeable?

Don't misunderstand, I am just asking, no offence!

Thanks!

Blaser

Hi Blaser,

I've been headed toward a somewhat specific set of blind listening tests to see if I can derive a pattern either way regarding with and without VLF.

The mistake most people make when entering the discussion, IMO, comes from all the usual talk of non-linear human hearing and house curve sorts of information. This leads many to think that the very low frequencies have to be at extremely high SPL to be perceived.

I use the illustration of the low rumble of a very distant thunderstorm as an example. It gets your attention regardless of how high the ambient noise is whitout having to be louder than the surrounding everyday sounds. At the same time, it's usually umistakably thunder, even though you can't hear any of the upper frequencies (which is because of the inverse square law that prohibits higher frequency sound from traveling very far before becoming inaudible and the fact that very low frequencies defy the inverse square law and can travel multitudes farther without losing intensity).

Add to those facts the thought that all unfiltered music transients contain VLF, regardless of the fundamental frequency, you get an interesting scenario where VLF reproduction actually can make the presentation more realistic just by including the entire spectrum of the original event.

Here are two graphs of a music transients that show the surprisingly (to me, anyway ) widely spread spectrum of the pluck of a single note of an instrument that is generally thought to be well above subsonic range:




So, the question becomes whether or not there is a perceived difference at normal playback levels of the same piece of music, with and without VLF. Of course there is a perceived difference when playing the DVD, WOTW, the lightning scene, where some of the strikes are mostly VLF and encoded at rather high levels.

Bosso

[SteveCallas]

Bosso wrote:

I use the illustration of the low rumble of a very distant thunderstorm as an example. It gets your attention regardless of how high the ambient noise is whitout having to be louder than the surrounding everyday sounds

And at what magnitude do you think the single digit frequencies created by a thunder strike are at when they are shaking your house? I don't think we are talking about 70db levels Let' say it's Xdb. Now what if at that exact moment, your subwoofer is playing a note 5hz higher at X+20db. That rumbling will easily mask the lower frequency rumbling.

In regards to the topic though, why would 100-200hz be less important than 1-10hz? 100-200hz is the region that greatly influences bass "sound quality" and we are much more percpective to it. <10hz is rumbling.

[Blaser]

Chrisbee wrote:

If you thought my rambling above was long you should read about the TRW subwoofer in this interesting article.

Everything you ever wanted to know about the reproduction of infrasonics but were afraid to ask.

http://www.iar-80.com/page142.html

I am first willing to experience a good hearing experience which could include by the way some infrasonics.... But Infrasonis as such are not my target. Despite I am no expert in infrasonics, I have some basic knowledge, and would never be afraid or shy of asking Nobody can know everything!

Ultra-infrasonics are much more expensive to achieve than even good low shaking sounds down to 15 Hz, and although they are present on DVDs and CDs, they need to be "exagerately" amplified to be experienced (correct me if I'm wrong), hence the high cost!

But again you certainly know much more than I do, I have never experienced an IB, but reads quite impressive things about.

But It is also a matter of taste: Many peoples like shaking bass in movies. One day I have placed one of my subs. behind the couch, calibrated it with REW, and then put JPIII. The couch was shaking a lot, but to my ears the sound was not equivalent to shaking, and I felt the sub effects and sound fake and week compared to the intensity of shaking, and indeed I could not enjoy the movie! I am sure that many peoples would like that , but I didn't and sent the sub back to the front wall. Don't mis-understand me, I love shaking bass but when they resut from realistic high SPL (which gives IMO a fuller experience)!!

I think my post has derailed from this thread's track, and apologize for that

B Rgds
Blaser

[Blaser]

SteveCallas wrote:

And at what magnitude do you think the single digit frequencies created by a thunder strike are at when they are shaking your house? I don't think we are talking about 70db levels Let' say it's Xdb. Now what if at that exact moment, your subwoofer is playing a note 5hz higher at X+20db. That rumbling will easily mask the lower frequency rumbling.

In regards to the topic though, why would 100-200hz be less important than 1-10hz? 100-200hz is the region that greatly influences bass "sound quality" and we are much more percpective to it. <10hz is rumbling.

Agree with you!