Your Home Theater ULF Score - Page 118 - AVS Forum | Home Theater Discussions And Reviews
Baselworld is only a few weeks away. Getting the latest news is easy, Click Here for info on how to join the Watchuseek.com newsletter list. Follow our team for updates featuring event coverage, new product unveilings, watch industry news & more!


Forum Jump: 
 337Likes
 
Thread Tools
Old 07-12-2015, 08:05 AM
AVS Special Member
 
coolrda's Avatar
 
Join Date: Nov 2002
Location: Bakersfield, CA
Posts: 1,613
Mentioned: 27 Post(s)
Tagged: 0 Thread(s)
Quoted: 204 Post(s)
Liked: 173
Quote:
Originally Posted by dominguez1 View Post


^^^Nice work guys. This is awesome!
@coolrda , it looks like 30hz and 15hz provide the most vibration combined from the 3 axes. Does this correspond to your subjective interpretation as well?
The vibration in the Y axis, Z axis and combines Y&Z axes all feel different from a subjective standpoint. So when I first felt it with the new sub and no TA and starting to get this fine tuned for better TR @20 and 15, 20hz was this biggest gainer but the big hit was 15. Talk about laying the lumba. But was that the most tactile? Well, theres never a higher acceleration point than when you go from still to movement. Subjectively I would say yes but I know better to do that at this point. I will say for a fact that Y&Z together provide a bigger TR hit than either alone. Thats a fact.

Quote:
Originally Posted by dominguez1 View Post
One thing else I've been meaning to point out; what about vibration on the cushion on your back? How does that correspond with your seat? Is it possible to measure that vibration?
Tried that the first day of use back in March and its the same numbers, just axes change because of orientation. Don't visualized the phone's body, visualized the MEMS inside as a sphere. It doesn't matter its orientation as the tilt corrects for this. It just needs to be stable.

People, this is the REAL DEAL when when it comes to TR. Accelerometer Tech blows away are mic based. A MEMS microphone will crushed our traditional mic's used today in accuracy, bandwidth and a hundred ways. Were falling behind Fellas. The old tech doesn't work, it can't diagnose the problem, it doesn't have the skill set. We can ride in that tired old bus, Or, we can get with the program, try to use our efforts to figure out this new cutting edge tech to fix this thing. i'm tired of speculation and talk with no solution. I want this FIXED. I've read dozens of papers and tech and see and hear about daily in my job. Science fiction one day is literally become my new tool or a new product the next. Look at the shape and detail of the waveform. Count the wave per second per frequency. Blow it up and examine it. No Mic-based meter in the world gives you this details, SpecLab or otherwise. And that without using Matlab.

I get all fired up about this. How in the heck did you persevere over two years? Love of the game, I guess.

Cinema 10 Max Build

Fat Man & Little Boy Build
No Raptors were harmed in the making of these subs


Grandbabies, my chance to make the same mistakes all over again.
coolrda is online now  
Sponsored Links
Advertisement
 
Old 07-12-2015, 11:39 AM
AVS Special Member
 
3ll3d00d's Avatar
 
Join Date: Sep 2007
Location: London, UK
Posts: 1,321
Mentioned: 10 Post(s)
Tagged: 0 Thread(s)
Quoted: 683 Post(s)
Liked: 202
Quote:
Originally Posted by dominguez1 View Post
Hack, but you get the idea.
I don't really see what information the amplitude view adds as the spectro tells you the shape of the content. Anyway I've attached one that lines up the start and end v closely, has the amplitude (which you can export from audacity) and squashes the spectro down to 50Hz.

Click image for larger version

Name:	eot.jpg
Views:	17
Size:	245.7 KB
ID:	826546
dominguez1 likes this.

Last edited by 3ll3d00d; 07-12-2015 at 12:30 PM.
3ll3d00d is online now  
Old 07-12-2015, 12:22 PM - Thread Starter
AVS Special Member
 
dominguez1's Avatar
 
Join Date: Dec 2008
Posts: 2,173
Mentioned: 13 Post(s)
Tagged: 0 Thread(s)
Quoted: 329 Post(s)
Liked: 275
Quote:
Originally Posted by coolrda View Post
Subjectively I would say yes but I know better to do that at this point. I will say for a fact that Y&Z together provide a bigger TR hit than either alone. Thats a fact.
Nice. Makes sense...

How interesting would this be to hook up a transducers like the Crowson, graph the X,Y,and Z axes and compare? Could be pretty telling no matter what side of the fence you're on.



Quote:
Tried that the first day of use back in March and its the same numbers, just axes change because of orientation.
How did you fasten it to the back cushions? I've tried, but could never get it right.

Quote:
People, this is the REAL DEAL when when it comes to TR. Accelerometer Tech blows away are mic based.
That is an awesome link and explanation of what you're talking about. Everyone should read that. You hit the jackpot there, very cool. Very interesting to see a 3D view of sound. Makes total sense. Do you think MEMS could give us similar insights as the Microflown? Did the MEMS record that in free air?

EDIT: I see now it was attached to the instrument. The Microflown and MEMS are now on the wishlist.

Quote:
I get all fired up about this. How in the heck did you persevere over two years? Love of the game, I guess.
Yeah, it wasn't easy taking on Goliath...(I'm glad @Okv came around for air support every now and then ), but what drove me is that all other explanations that I was given by those much more knowledgeable then me, didn't refute the theory well enough for me to throw in the towel. I was definitely open to listening, and still am, but none provided definitive evidence of the contrary.

My hope has always been to get the ball rolling, and have others help take it to the next level....you and others have been doing a fine job lately.
coolrda likes this.

Last edited by dominguez1; 07-12-2015 at 12:26 PM.
dominguez1 is online now  
Old 07-12-2015, 12:53 PM
AVS Special Member
 
3ll3d00d's Avatar
 
Join Date: Sep 2007
Location: London, UK
Posts: 1,321
Mentioned: 10 Post(s)
Tagged: 0 Thread(s)
Quoted: 683 Post(s)
Liked: 202
one of these has the right specs -> http://www.invensense.com/products/analog/ics-40300-3/ (datasheet inc FR - http://store.invensense.com/datashee...pec_V1%200.pdf)

they sell an evaluation board for it -> http://store.invensense.com/ProductD...se-Inc/519137/

or grab a coconut, a hot glue gun and soldering iron and away you go -> http://www.instructables.com/id/Coco...ic-MEMS-techn/

Not sure how the 3d bit comes out mind you

Last edited by 3ll3d00d; 07-12-2015 at 02:00 PM.
3ll3d00d is online now  
Old 07-12-2015, 01:28 PM
AVS Special Member
 
coolrda's Avatar
 
Join Date: Nov 2002
Location: Bakersfield, CA
Posts: 1,613
Mentioned: 27 Post(s)
Tagged: 0 Thread(s)
Quoted: 204 Post(s)
Liked: 173
Quote:
Originally Posted by 3ll3d00d View Post
I don't really see what information the amplitude view adds as the spectro tells you the shape of the content. Anyway I've attached one that lines up the start and end v closely, has the amplitude (which you can export from audacity) and squashes the spectro down to 50Hz.

Attachment 826546
I agree that it doesn't work but had to see it to know. You've definitely thrown everything but the kitchen in. I like. 1920x1080 makes for a great viewing.

Cinema 10 Max Build

Fat Man & Little Boy Build
No Raptors were harmed in the making of these subs


Grandbabies, my chance to make the same mistakes all over again.
coolrda is online now  
Old 07-12-2015, 01:57 PM
AVS Special Member
 
coolrda's Avatar
 
Join Date: Nov 2002
Location: Bakersfield, CA
Posts: 1,613
Mentioned: 27 Post(s)
Tagged: 0 Thread(s)
Quoted: 204 Post(s)
Liked: 173
Quote:
Originally Posted by dominguez1 View Post
Nice. Makes sense...

How interesting would this be to hook up a transducers like the Crowson, graph the X,Y,and Z axes and compare? Could be pretty telling no matter what side of the fence you're on.
That would be my first choice of component testing. If null in TR can't be resolved in certain rooms this would be a cinch to EQ Crowsons and check TR response on all axes.



Quote:
Originally Posted by dominguez1 View Post
How did you fasten it to the back cushions? I've tried, but could never get it right.
Never fastened. Just laid against it. Don't know how good suction cup would hold to back. I'm going to test the arms for comparison though.

Quote:
Originally Posted by dominguez1 View Post
That is an awesome link and explanation of what you're talking about. Everyone should read that. You hit the jackpot there, very cool. Very interesting to see a 3D view of sound. Makes total sense. Do you think MEMS could give us similar insights as the Microflown? Did the MEMS record that in free air?

EDIT: I see now it was attached to the instrument. The Microflown and MEMS are now on the wishlist.


Yeah, it wasn't easy taking on Goliath...(I'm glad @Okv came around for air support every now and then ), but what drove me is that all other explanations that I was given by those much more knowledgeable then me, didn't refute the theory well enough for me to throw in the towel. I was definitely open to listening, and still am, but none provided definitive evidence of the contrary.

My hope has always been to get the ball rolling, and have others help take it to the next level....you and others have been doing a fine job lately.
I'm glad you persevered. You got help now. Yeah, things didn't turn out to well for Goliath, did it. Referring back to the pendulum, MEMS gives us the acceleration and Microflown the velocity. Their siblings so we can get around it if the micro flown proves elusive. I know theres a lot of work to do but I think were on the right path.

Cinema 10 Max Build

Fat Man & Little Boy Build
No Raptors were harmed in the making of these subs


Grandbabies, my chance to make the same mistakes all over again.
coolrda is online now  
Old 07-12-2015, 02:01 PM
AVS Special Member
 
coolrda's Avatar
 
Join Date: Nov 2002
Location: Bakersfield, CA
Posts: 1,613
Mentioned: 27 Post(s)
Tagged: 0 Thread(s)
Quoted: 204 Post(s)
Liked: 173
I do wanna tried it free air just to see what happens.

Cinema 10 Max Build

Fat Man & Little Boy Build
No Raptors were harmed in the making of these subs


Grandbabies, my chance to make the same mistakes all over again.
coolrda is online now  
Old 07-12-2015, 03:00 PM
AVS Special Member
 
coolrda's Avatar
 
Join Date: Nov 2002
Location: Bakersfield, CA
Posts: 1,613
Mentioned: 27 Post(s)
Tagged: 0 Thread(s)
Quoted: 204 Post(s)
Liked: 173
Quote:
Originally Posted by 3ll3d00d View Post
one of these has the right specs -> http://www.invensense.com/products/analog/ics-40300-3/ (datasheet inc FR - http://store.invensense.com/datashee...pec_V1%200.pdf)

they sell an evaluation board for it -> http://store.invensense.com/ProductD...se-Inc/519137/

or grab a coconut, a hot glue gun and soldering iron and away you go -> http://www.instructables.com/id/Coco...ic-MEMS-techn/


Not sure how the 3d bit comes out mind you
Very interesting.

Cinema 10 Max Build

Fat Man & Little Boy Build
No Raptors were harmed in the making of these subs


Grandbabies, my chance to make the same mistakes all over again.
coolrda is online now  
Old 07-12-2015, 03:28 PM
AVS Special Member
 
3ll3d00d's Avatar
 
Join Date: Sep 2007
Location: London, UK
Posts: 1,321
Mentioned: 10 Post(s)
Tagged: 0 Thread(s)
Quoted: 683 Post(s)
Liked: 202
Quote:
Originally Posted by coolrda View Post
Very interesting.
I read the article again and realised I had misread it completely

Now I have read it again, I am pretty sure this is something that could be implemented on a phone. It gives you a digital feed at the sensor rate and vibsensor can export the raw data stream. One could then fft this in exactly the same way as speclab/rew et al do (because ultimately a mic is a voltage source that gets translated via an ADC into a stream of numbers that represents the waveform in digital form). The main difference is the low sample rate, 100Hz is obviously tiny/slow compared to audio data rates. I imagine this will be a significant limiting factor in the subsequent processing.

I suspect this is the sort of thing arduino, or similar, should be good for. You are then free to choose your hardware and arduino gives you the infrastructure to build a data collector on top of it. If you can get the data then it could be moved onto a full pc for further processing.

Actually I wonder what would happen if we converted vibsensor data into a wav and fed it through speclab. Perhaps I will try that to see if anything sane comes out. I guess normalising to some notion of 0dBFS (eg assume 1g is 0dBFS), convert the vibsensor data to a 100Hz 3 channel wav (or 3 mono wavs) then feed into speclab for visualisation.
coolrda likes this.

Last edited by 3ll3d00d; 07-12-2015 at 03:33 PM.
3ll3d00d is online now  
Old 07-12-2015, 03:58 PM
AVS Special Member
 
3ll3d00d's Avatar
 
Join Date: Sep 2007
Location: London, UK
Posts: 1,321
Mentioned: 10 Post(s)
Tagged: 0 Thread(s)
Quoted: 683 Post(s)
Liked: 202
Actually that explains why the vibsensor PSD chart goes to 50Hz, this is the Nyquist frequency for the data as the sample rate is 100Hz. Perhaps we can use speclab to tease out a view on the vibsensor data that is more than just amplitude after all.
3ll3d00d is online now  
Old 07-12-2015, 11:20 PM
AVS Special Member
 
coolrda's Avatar
 
Join Date: Nov 2002
Location: Bakersfield, CA
Posts: 1,613
Mentioned: 27 Post(s)
Tagged: 0 Thread(s)
Quoted: 204 Post(s)
Liked: 173
We only have two samples but the same issue with both. @desertdome and I have different rooms with different subs system but have the same null in Z axis TR@15hz. At the point where we have the highest or a peak in SPL, we have the lowest TR. It happens to be at the same frequency for both our rooms. This can be explained simply with Correlation. In this case positive and negative correlation. What is that? Very simply if you have two variables and when one variable is moved to the positive, and the other does as well, thats a positive variable. Negative variable is the opposite, when one variable moves to the positive, another moves to the negative. The other important thing is to find the causation of the particular problem or issue were having. We have to identify the cause if we are to come up with a solution. Let’s review to this point.

This is my response with and without Time Alignment.



Time Aligning multiple subs allows you to achieve the maximum spl at the MLP. This show my frequency response from 5-50hz. Notice the peak around 40hz. Theres a corresponding gain in TR that matches the FR over the band the subs in. We have a positive correlation. You turn up the volume, it gets louder and TR increases. All pretty simple.
But………. low frequency tactile response seemed much weaker. Playback of ULF heavy media, War or the Worlds, Incredible Hulk, Edge of Tomorrow all felt weak with time alignment. zeroing TA made a huge improvement to the tactile feel of the clips. It also flattened the peak. Trying to identify the problem with Omnimic proved frustrating and fruitless. This wasn’t working. I needed to find the right tool. The rest is history.

This is the tactile response.





Above 24hz, we have a positive correlation but below that, theres a negative correlation and the opposite happens. So now that we know the effects, we need to identify the cause. Correlation doesn’t always lead to causation or it could lead to more than one. in this case theres more than one. There is indeed a primary and a secondary causation. In In my room you clearly see a larger trough than in DD’s room. On the EoT test clip I have a 20hz and 15hz null and in DD’s room just the 15hz null, where his SPL was highest and TR was the lowest. So DesertDome’s and my room have something in common and something that isn’t. Eliminating the TA delay remedies the 20hz null. Thats demonstrated very clearly and easily. Lets call this the secondary causation. With it eliminated, TR improved dramatically at 25 and 20hz. Now both rooms mirror each other with a significant Z axis null at 15hz. DesertDome’s room is different as well as his type of subs and placement. But our rooms share a common denominator. What is that? I believe the Z axis is the answer, and very simply the Z axis is the rooms height. This may be the one and only thing ours rooms have in common, the height. With basement theaters ranging from 7-8ft or more and single or two story HT ranging from 8-10ft on average, it makes sense that our rooms TR would match fairly close in the Z axis. I believe if we measure more rooms, we may see a pattern there in the Z axis. If that proves to be the case, then it would stand to reason that room height is the single biggest deterrent to TR@15hz or near that frequency depending on height. If thats the case, then the primary causation of TR dip is the boundary. For the record my room height is 9ft. At this point its too early to talk solution(even though eliminating boundaries appears early on to be the answer, or using and equalizing axes specific tactile transducers), even causation because we’re still trying to figure a test protocol. Actually we are in the experimentation phase. BTW, as a by product of this research, it appears we have stumbled on a way to identify the frequencies that we need to excite by way of TT’s. Now that we have a way to meter and set TT’s, they can be expertly set and blended seemlessly with a sub’s and room’s natural response and eq’d accordingly. If this is the only solution, so be it. It would be a noteworthy accomplishment.

Cinema 10 Max Build

Fat Man & Little Boy Build
No Raptors were harmed in the making of these subs


Grandbabies, my chance to make the same mistakes all over again.
coolrda is online now  
Old 07-12-2015, 11:34 PM
AVS Special Member
 
DotJun's Avatar
 
Join Date: Dec 2012
Posts: 1,327
Mentioned: 5 Post(s)
Tagged: 0 Thread(s)
Quoted: 365 Post(s)
Liked: 93
Quote:
Originally Posted by desertdome View Post
tsmuxer or eac3to don't take bass management into consideration. You just get the raw channels. And neither method is easier for that either.



It literally takes me 5 seconds to get any clip from any movie with or without bass management.
  1. Enter playback range in file tag
  2. Click Convert Format
  3. Check Convert Video to Audio
  4. Click Convert
  5. Done







I don't always even convert. I can enter the playback range and then play back live while viewing simultaneously in JRiver, Spectrum Lab, REW's SPL Logger, REW's SPL Meter, and REW's RTA.




Sorry @desertdome , I must have missed something in your previous post. What do you mean that neither app takes bass management into consideration?
DotJun is online now  
Old 07-12-2015, 11:55 PM
AVS Special Member
 
coolrda's Avatar
 
Join Date: Nov 2002
Location: Bakersfield, CA
Posts: 1,613
Mentioned: 27 Post(s)
Tagged: 0 Thread(s)
Quoted: 204 Post(s)
Liked: 173
Quote:
Originally Posted by 3ll3d00d View Post
I read the article again and realised I had misread it completely

Now I have read it again, I am pretty sure this is something that could be implemented on a phone. It gives you a digital feed at the sensor rate and vibsensor can export the raw data stream. One could then fft this in exactly the same way as speclab/rew et al do (because ultimately a mic is a voltage source that gets translated via an ADC into a stream of numbers that represents the waveform in digital form). The main difference is the low sample rate, 100Hz is obviously tiny/slow compared to audio data rates. I imagine this will be a significant limiting factor in the subsequent processing.

I suspect this is the sort of thing arduino, or similar, should be good for. You are then free to choose your hardware and arduino gives you the infrastructure to build a data collector on top of it. If you can get the data then it could be moved onto a full pc for further processing.

Actually I wonder what would happen if we converted vibsensor data into a wav and fed it through speclab. Perhaps I will try that to see if anything sane comes out. I guess normalising to some notion of 0dBFS (eg assume 1g is 0dBFS), convert the vibsensor data to a 100Hz 3 channel wav (or 3 mono wavs) then feed into speclab for visualisation.
OOTB thinking right there. Why can't we do that? I swap file, pics, vids, updates, backups, you name it through the usb. Or better yet Bluttoth or wifi. Its only a matter of time. Their do it in the med industry. A micbased and MEMS based speclab type readout would be crazy.

Cinema 10 Max Build

Fat Man & Little Boy Build
No Raptors were harmed in the making of these subs


Grandbabies, my chance to make the same mistakes all over again.
coolrda is online now  
Old 07-13-2015, 12:15 AM
AVS Special Member
 
coolrda's Avatar
 
Join Date: Nov 2002
Location: Bakersfield, CA
Posts: 1,613
Mentioned: 27 Post(s)
Tagged: 0 Thread(s)
Quoted: 204 Post(s)
Liked: 173
Quote:
Originally Posted by dominguez1 View Post
Nice. Makes sense...

How interesting would this be to hook up a transducers like the Crowson, graph the X,Y,and Z axes and compare? Could be pretty telling no matter what side of the fence you're on.




How did you fasten it to the back cushions? I've tried, but could never get it right.


That is an awesome link and explanation of what you're talking about. Everyone should read that. You hit the jackpot there, very cool. Very interesting to see a 3D view of sound. Makes total sense. Do you think MEMS could give us similar insights as the Microflown? Did the MEMS record that in free air?

EDIT: I see now it was attached to the instrument. The Microflown and MEMS are now on the wishlist.


Yeah, it wasn't easy taking on Goliath...(I'm glad @Okv came around for air support every now and then ), but what drove me is that all other explanations that I was given by those much more knowledgeable then me, didn't refute the theory well enough for me to throw in the towel. I was definitely open to listening, and still am, but none provided definitive evidence of the contrary.

My hope has always been to get the ball rolling, and have others help take it to the next level....you and others have been doing a fine job lately.
Found this on WIKI so maybe you can do some creative Particle velocity testing. We'll have to try it.

Accelerometers are used to measure the motion and vibration of a structure that is exposed to dynamic loads.[22] Dynamic loads originate from a variety of sources including:

Human activities – walking, running, dancing or skipping
Working machines – inside a building or in the surrounding area
Construction work – driving piles, demolition, drilling and excavating
Moving loads on bridges
Vehicle collisions
Impact loads – falling debris
Concussion loads – internal and external explosions
Collapse of structural elements
Wind loads and wind gusts
Air blast pressure
Loss of support because of ground failure
Earthquakes and aftershocks
Under structural applications, measuring and recording how a structure dynamically responds to these inputs is critical for assessing the safety and viability of a structure. This type of monitoring is called Health Monitoring, which usually involves other types of instruments, such as displacement sensors -Potentiometers, LVDTs, etc.- deformation sensors -Strain Gauges, Extensometers-, load sensors -Load Cells, Piezo-Electric Sensors- among others.

Cinema 10 Max Build

Fat Man & Little Boy Build
No Raptors were harmed in the making of these subs


Grandbabies, my chance to make the same mistakes all over again.
coolrda is online now  
Old 07-13-2015, 04:11 AM
AVS Special Member
 
3ll3d00d's Avatar
 
Join Date: Sep 2007
Location: London, UK
Posts: 1,321
Mentioned: 10 Post(s)
Tagged: 0 Thread(s)
Quoted: 683 Post(s)
Liked: 202
Quote:
Originally Posted by coolrda View Post
OOTB thinking right there. Why can't we do that? I swap file, pics, vids, updates, backups, you name it through the usb. Or better yet Bluttoth or wifi. Its only a matter of time. Their do it in the med industry. A micbased and MEMS based speclab type readout would be crazy.
I've checked the sox docsand sox can handle this sort of conversion easily. It can read a .dat file which is a textual representation of the waveform (i.e. the vib data) and also has a spectrogram option so it should be possible to apply some basic formatting to the vib data and pipe that through the spectrogram and/or convert it to a wav to feed into speclab.
3ll3d00d is online now  
Old 07-13-2015, 04:45 AM
AVS Special Member
 
3ll3d00d's Avatar
 
Join Date: Sep 2007
Location: London, UK
Posts: 1,321
Mentioned: 10 Post(s)
Tagged: 0 Thread(s)
Quoted: 683 Post(s)
Liked: 202
Quote:
Originally Posted by coolrda View Post
Time Aligning multiple subs allows you to achieve the maximum spl at the MLP. This show my frequency response from 5-50hz. Notice the peak around 40hz. Theres a corresponding gain in TR that matches the FR over the band the subs in. We have a positive correlation. You turn up the volume, it gets louder and TR increases. All pretty simple.
I might be missing something but why you would ever use the 7ms configuration when it yields a 9dB spike in the response between 30-50Hz?
3ll3d00d is online now  
Old 07-13-2015, 08:23 AM
AVS Special Member
 
desertdome's Avatar
 
Join Date: Aug 2004
Location: Elkhorn, NE
Posts: 2,393
Mentioned: 60 Post(s)
Tagged: 0 Thread(s)
Quoted: 502 Post(s)
Liked: 513
Quote:
Originally Posted by DotJun View Post
@desertdome , an easier method to extract the audio file would be to use tsmuxer or eac3to.
Quote:
Originally Posted by DotJun View Post
Sorry @desertdome , I must have missed something in your previous post. What do you mean that neither app takes bass management into consideration?
I assumed your first statement was in reference to my post about using JRiver to extract audio for the purpose of bass analysis by Spectrum Lab. The charts at data-bass are not just of the LFE channel. They are all of the LFE channel plus redirected bass from other channels with level reduction done a certain to account for a worse case scenario, i.e. Bass Management. In order to combine the redirected bass with other programs you must first extract with one program and then use another to combine them. Then you have to use the proper mix down math to make sure you are accurately replicating the bass management done by Blu-ray players, receivers, etc. Finally, with eac3to at least, you need a separate decoder installed to decode DTS-HD. I'm not sure about tsmuxer.

With JRiver all steps are done at once and it literally takes 5 seconds per clip to have a wav or flac file that can be analyzed in Spectrum Lab with correct bass management.

Many media programs, like mpc-hc, rely on LAV Filters for decoding, demuxing, or remuxing the audio. The sole developer of LAV Filters, Hendrick Leppkes, is one of JRiver's main programmers and lets his filters be used by the other programs. Some credit LAV Filters, but he has seen others just copy his code. JRiver adds the nice GUI and flexible DSP that makes it so simple.
desertdome is online now  
Old 07-13-2015, 08:25 AM
AVS Special Member
 
desertdome's Avatar
 
Join Date: Aug 2004
Location: Elkhorn, NE
Posts: 2,393
Mentioned: 60 Post(s)
Tagged: 0 Thread(s)
Quoted: 502 Post(s)
Liked: 513
Quote:
Originally Posted by coolrda View Post
The EoT in Digital waveform from the source in REW.
Quick correction - that is the waveform shown in Audacity, not REW.
desertdome is online now  
Old 07-13-2015, 08:34 AM
AVS Special Member
 
desertdome's Avatar
 
Join Date: Aug 2004
Location: Elkhorn, NE
Posts: 2,393
Mentioned: 60 Post(s)
Tagged: 0 Thread(s)
Quoted: 502 Post(s)
Liked: 513
Quote:
Originally Posted by coolrda View Post
I see, nice rig. If there was a weird phasing mixed in at certain freqs, that would show up, correct? Ok so you have a manifold single point entry into the room? No front to rear array? At first I thought it was a front/rear time alignment and it is at 20hz w/7ms delay. But I still have trough at 15hz even with front subs muted. So I think, whether the wave is cancelled from a second array or its bouncing off a boundary onto itself, that appears to be what's happening.
I never answered this. Yes, you should be able to see weird phasing.

I have two manifolds with four 15" drivers each. Each manifold is slightly outside and behind my mains so I can use a stereo bass configuration, too.

Vibration data at the manifold is obviously of just one manifold. However, when I measured from the listening position I had both manifolds running. I need to retake a measurement with just one manifold and then see if there is any cancellation from both manifolds and what improvement can be made.
desertdome is online now  
Old 07-13-2015, 08:58 AM
AVS Special Member
 
coolrda's Avatar
 
Join Date: Nov 2002
Location: Bakersfield, CA
Posts: 1,613
Mentioned: 27 Post(s)
Tagged: 0 Thread(s)
Quoted: 204 Post(s)
Liked: 173
Quote:
Originally Posted by desertdome View Post
I never answered this. Yes, you should be able to see weird phasing.

I have two manifolds with four 15" drivers each. Each manifold is slightly outside and behind my mains so I can use a stereo bass configuration, too.

Vibration data at the manifold is obviously of just one manifold. However, when I measured from the listening position I had both manifolds running. I need to retake a measurement with just one manifold and then see if there is any cancellation from both manifolds and what improvement can be made.
I'll be watching.

Cinema 10 Max Build

Fat Man & Little Boy Build
No Raptors were harmed in the making of these subs


Grandbabies, my chance to make the same mistakes all over again.
coolrda is online now  
Old 07-13-2015, 08:59 AM
AVS Special Member
 
coolrda's Avatar
 
Join Date: Nov 2002
Location: Bakersfield, CA
Posts: 1,613
Mentioned: 27 Post(s)
Tagged: 0 Thread(s)
Quoted: 204 Post(s)
Liked: 173
Quote:
Originally Posted by 3ll3d00d View Post
I might be missing something but why you would ever use the 7ms configuration when it yields a 9dB spike in the response between 30-50Hz?
Good question. Ok let's go through the progressions of a person's(AVSer) purchase and use of subwoofer/s over time, typically years. For example, I built my first sub in 1985. Over the 30 years, I bought and built subs starting this cycle with DIY, then commercial bought, then DIY, then Comm, and finally back to DIY about the last 4 years. Pretty much the same cycle with mains. Anyway so you buy or build your first sub. You get used to it, then you want better integration so you move it around the room , do the sub crawl, whatever. Time passes, your more experienced, you buy a measurement system or another sub or both. Now your see the FR of your sub/room together and want to fix it. So the next step is adding an eq, by way of an outboard unit or onboard an amp. Next step is to TA and EQ. We were always taught to TA then eq. Only makes sense right. It's what we do with our multichannel playback systems either manually or automated. Depending on what we do with our placement, we will have varying degrees of response, frequency and tactile. With perfect TA we may see a ideal 6db gain at the MLP. Now TR may or may not do likewise. To put it in a term we can understand let's use DB. Maybe our TR only improved 3 or 1db. It better than one sub but not quite what your expecting. You grumbled and move on. Wait I'll moved them around the room, add subs, take away subs, add acoustical treatments and on and on and on. Our bodies are telling us something because their great sensors in the 0-20hz range. Our senses are telling us what's going on but we can't see it, can't meter it so oh well. So we again start messing with sub placement. We find a spot we think is better. Then we TA and EQ. We say that's a little better or worse, kinda puzzled. Remember the primary and secondary causation? Can't do nothing about the primarily causation and when we're done the first thing we do is compound the TR null reintroducing the secondary causation by re TAing. How hilarious is that. I've done that many many times myself. I'm picturing the thousands of YouTube videos of cats chasing their tales. Hilarious. Now too be sure, there's tons of variables. Two front corner symmetrically placed sub will naturally have less TR loss than more complex systems which is one reason why I stuck with that and avoided the near field over the years. And there's dozens more. I can easily see the home theater builder of the future giving his client a velocity/acceleration/frequency response doc to show the whole picture of what they've accomplished.

Now the simple answer, the 30-50 is a combined speaker response/room modal response. You TA then EQ it down. You extract the max SPL, then smooth. It's the proper way. We do it. Room correction software does it. It's the way we're taught to do it. FR measurement hardware and software is absolutely needed. It's like an X-Ray. It's the right tool in many cases and the first step. But when you need to see further, when you need to see the whole picture in much greater detail, we need an MRI. Accelerate/Velo MEMS could vary well be that.

Cinema 10 Max Build

Fat Man & Little Boy Build
No Raptors were harmed in the making of these subs


Grandbabies, my chance to make the same mistakes all over again.

Last edited by coolrda; 07-13-2015 at 09:24 AM.
coolrda is online now  
Old 07-13-2015, 09:52 AM
AVS Special Member
 
coolrda's Avatar
 
Join Date: Nov 2002
Location: Bakersfield, CA
Posts: 1,613
Mentioned: 27 Post(s)
Tagged: 0 Thread(s)
Quoted: 204 Post(s)
Liked: 173
@desertdome , or whomever, would you be interested in putting together a suite of test tones and sweeps? I'm really sick of EoT. We need 0db from full scale, maybe -3db and even -20db. I tired of switching from stereo to surround to mono. I have the luxury of turning off amps I don't what sound from. Many do not. I want LFE channel only but bass managed added as well. I want it to lock the pre/pro or receiver into 5.1 or 7.1 so there's no variation of signal amplitude because of BM. We're gonna needs sweeps and tones and transients, each at different durations. Take into account a the sampling rate of 100hz and FR of 1-50hz only. I thought just down to 5hz but might as well go to 1hz. PCM or bitstream, I don't care. Dts-hd and Dolby truehd if possible. And with appropriate leadin or whatever they call it to make it burnable for playback in br format. I guess we can do the later. I just wanna pick the particular track, make the Run and go onto the next. Set it up how you see fit. That'd be cool.

Cinema 10 Max Build

Fat Man & Little Boy Build
No Raptors were harmed in the making of these subs


Grandbabies, my chance to make the same mistakes all over again.
coolrda is online now  
Old 07-13-2015, 10:23 AM
AVS Special Member
 
3ll3d00d's Avatar
 
Join Date: Sep 2007
Location: London, UK
Posts: 1,321
Mentioned: 10 Post(s)
Tagged: 0 Thread(s)
Quoted: 683 Post(s)
Liked: 202
Quote:
Originally Posted by coolrda View Post
Now the simple answer, the 30-50 is a combined speaker response/room modal response. You TA then EQ it down. You extract the max SPL, then smooth. It's the proper way. We do it. Room correction software does it. It's the way we're taught to do it. FR measurement hardware and software is absolutely needed. It's like an X-Ray. It's the right tool in many cases and the first step. But when you need to see further, when you need to see the whole picture in much greater detail, we need an MRI. Accelerate/Velo MEMS could vary well be that.
Just to be clear, the 7ms one produces a peak that needs to be eq'ed down but the 0ms is naturally flat(ter). Is that right?
3ll3d00d is online now  
Old 07-13-2015, 11:05 AM
AVS Special Member
 
coolrda's Avatar
 
Join Date: Nov 2002
Location: Bakersfield, CA
Posts: 1,613
Mentioned: 27 Post(s)
Tagged: 0 Thread(s)
Quoted: 204 Post(s)
Liked: 173
Quote:
Originally Posted by 3ll3d00d View Post
Just to be clear, the 7ms one produces a peak that needs to be eq'ed down but the 0ms is naturally flat(ter). Is that right?
Yep.

Cinema 10 Max Build

Fat Man & Little Boy Build
No Raptors were harmed in the making of these subs


Grandbabies, my chance to make the same mistakes all over again.
coolrda is online now  
Old 07-13-2015, 11:40 AM
AVS Special Member
 
coolrda's Avatar
 
Join Date: Nov 2002
Location: Bakersfield, CA
Posts: 1,613
Mentioned: 27 Post(s)
Tagged: 0 Thread(s)
Quoted: 204 Post(s)
Liked: 173
Quote:
Originally Posted by 3ll3d00d View Post
Just to be clear, the 7ms one produces a peak that needs to be eq'ed down but the 0ms is naturally flat(ter). Is that right?

This is an old compression graph but it illustrates the point. As I go from 1 to 2 to 4 to 8 subs I get the same freq sweep overlay but with 6db increases up to an 18db total increase. It doesn't look bad from this graph. The one above is chopped off at 50hz and uses a 1db scale for max resolution.

Last edited by coolrda; 07-13-2015 at 11:44 AM.
coolrda is online now  
Old 07-13-2015, 11:46 AM
AVS Special Member
 
3ll3d00d's Avatar
 
Join Date: Sep 2007
Location: London, UK
Posts: 1,321
Mentioned: 10 Post(s)
Tagged: 0 Thread(s)
Quoted: 683 Post(s)
Liked: 202
Quote:
Originally Posted by coolrda View Post
Yep.
here's the spectrogram of your y axis vibration data, tactility seems to be dominated by the 30+Hz region (corresponds to your FR change)

Click image for larger version

Name:	spectrogram.png
Views:	19
Size:	102.9 KB
ID:	828122
3ll3d00d is online now  
Old 07-13-2015, 02:50 PM
AVS Special Member
 
coolrda's Avatar
 
Join Date: Nov 2002
Location: Bakersfield, CA
Posts: 1,613
Mentioned: 27 Post(s)
Tagged: 0 Thread(s)
Quoted: 204 Post(s)
Liked: 173
Quote:
Originally Posted by 3ll3d00d View Post
here's the spectrogram of your y axis vibration data, tactility seems to be dominated by the 30+Hz region (corresponds to your FR change)

Attachment 828122
How are you doing that? Is there a chain of conversion and handoffs or are you going straight from the data file to that? I guess I thought wrong when I thought matlab was needed to do the spectrogram. Thats amazing. That's exactly what I was hoping for. So can we put the digital spectro top windowed, freq mid windowed and combined all axes tactile bottom windowed for direct comparison? I don't think I necessarily want all three axes shown in spectro but a combined window. What is the vibrational amplitude scale calibrated too? Just for illustration or use the highest peaks of each above the room gain. My choices today would be the three windowed axes view above the spectro and the three windowed spectro of source/frequency/tactile.

Cinema 10 Max Build

Fat Man & Little Boy Build
No Raptors were harmed in the making of these subs


Grandbabies, my chance to make the same mistakes all over again.
coolrda is online now  
Old 07-13-2015, 03:01 PM
AVS Special Member
 
3ll3d00d's Avatar
 
Join Date: Sep 2007
Location: London, UK
Posts: 1,321
Mentioned: 10 Post(s)
Tagged: 0 Thread(s)
Quoted: 683 Post(s)
Liked: 202
Quote:
Originally Posted by coolrda View Post
How are you doing that? Is there a chain of conversion and handoffs or are you going straight from the data file to that? I guess I thought wrong when I thought matlab was needed to do the spectrogram. Thats amazing. That's exactly what I was hoping for. So can we put the digital spectro top windowed, freq mid windowed and combined all axes tactile bottom windowed for direct comparison? I don't think I necessarily want all three axes shown in spectro but a combined window. What is the vibrational amplitude scale calibrated too? Just for illustration or use the highest peaks of each above the room gain. My choices today would be the three windowed axes view above the spectro and the three windowed spectro of source/frequency/tactile.
that particular example just quickly hacked out as follows

Code:
$ awk -F"," '/^[0-9]/ {print (NR-1)*0.01" "($3*(1/0.312886))}' eot_vib.csv > eot_vib_y.dat
$ sox -V eot_vib_y.dat -n spectrogram -x 1500 -z 60  -X 50
the 1st line picks out the 3rd column, scales it so the largest absolute amplitude (which happens to be 0.312886) is 1 (i.e. treated as 0dBFS) and then prints it in sox dat format which is like this

Code:
; Sample Rate 100
; Channels 1
0.01 0.000852505
0.02 -0.00135834
0.03 0.00218499
0.04 0.00212094
0.05 -0.00827867
0.06 -0.0123336
0.07 -0.00965217
0.08 -0.0154972
0.09 -0.00962884
0.1 -0.0143329
0.11 -0.00259119
0.12 -0.00844311
i.e. a space delimited time/channel n amplitude description of the waveform

the 2nd line then tells sox to generate a spectrogram that is 1500px wide, uses gives 50px per second and has a 60dB depth on the z axis
3ll3d00d is online now  
Old 07-13-2015, 03:46 PM
AVS Special Member
 
3ll3d00d's Avatar
 
Join Date: Sep 2007
Location: London, UK
Posts: 1,321
Mentioned: 10 Post(s)
Tagged: 0 Thread(s)
Quoted: 683 Post(s)
Liked: 202
@coolrda here's all three channels, x is at the top then y then z

Click image for larger version

Name:	eot_vib.png
Views:	12
Size:	322.9 KB
ID:	828730

and the content being played

Click image for larger version

Name:	eot.jpg
Views:	8
Size:	221.0 KB
ID:	828738

I can't easily put them on one chart because it is being generated by a different package (and also the chart would be getting rather large... it is doable of course, just requires effort... Simplest way would be to convert the dat to a wav, play that in speclab and export the fft to a txt file for gnuplot to plot).

FWIW the script to hack this out is https://github.com/3ll3d00d/ulf-util...er/generate.sh

I think the key points this shows are;

- x is markedly weaker than y or z but is obviously reacting to the same content
- vibrations get weaker as frequency descends & the 10Hz tone is really doing nothing much
- the tactile feel of the 15Hz fundamental is dominated by the 45Hz 3rd harmonic overtone (which is in the original content), especially on the y axis

EDIT: replaced the spectro as I realised I copied the y data to the z channel and hence concluded y and z were really v similar after all, facepalm all round

Last edited by 3ll3d00d; 07-13-2015 at 04:26 PM.
3ll3d00d is online now  
Old 07-13-2015, 05:15 PM
AVS Special Member
 
coolrda's Avatar
 
Join Date: Nov 2002
Location: Bakersfield, CA
Posts: 1,613
Mentioned: 27 Post(s)
Tagged: 0 Thread(s)
Quoted: 204 Post(s)
Liked: 173
Quote:
Originally Posted by 3ll3d00d View Post
that particular example just quickly hacked out as follows

Code:
$ awk -F"," '/^[0-9]/ {print (NR-1)*0.01" "($3*(1/0.312886))}' eot_vib.csv > eot_vib_y.dat
$ sox -V eot_vib_y.dat -n spectrogram -x 1500 -z 60  -X 50
the 1st line picks out the 3rd column, scales it so the largest absolute amplitude (which happens to be 0.312886) is 1 (i.e. treated as 0dBFS) and then prints it in sox dat format which is like this

Code:
; Sample Rate 100
; Channels 1
0.01 0.000852505
0.02 -0.00135834
0.03 0.00218499
0.04 0.00212094
0.05 -0.00827867
0.06 -0.0123336
0.07 -0.00965217
0.08 -0.0154972
0.09 -0.00962884
0.1 -0.0143329
0.11 -0.00259119
0.12 -0.00844311
i.e. a space delimited time/channel n amplitude description of the waveform

the 2nd line then tells sox to generate a spectrogram that is 1500px wide, uses gives 50px per second and has a 60dB depth on the z axis
I like the peak amplitude being set based on that method but only if its sample in the positive correlation territory. In my case any spot above 24hz, preferably between 40-50hz. As long as our max amplitude can be based on a sample there, and locked in for those series of test runs, I think it's reliable, until we figure out something better. That formula doesn't work when based on negative correlation territory samples as that gets skewed then the tactile domain data is worthless when compared to the freq domain. I want test files with single hertz narrow band transient pulse from 1-20, linear sweeps from 5-10, 10-20, 15-5. I want take some fast sine sweeps from the Omnimic test CD and compared to EoT. When I do I'll send you both for comparison. I'd like to see how VS handles the sweeps.

Cinema 10 Max Build

Fat Man & Little Boy Build
No Raptors were harmed in the making of these subs


Grandbabies, my chance to make the same mistakes all over again.

Last edited by coolrda; 07-13-2015 at 11:24 PM.
coolrda is online now  
Sponsored Links
Advertisement
 
 

Tags
ae ib15 subwoofer , Bic Pl 200 Acoustech Platinum Series Subwoofer , Bowers Wilkins Asw 610 , Danley Sound Labs Dts 10 , Elemental Designs A7s 450 , Epik Empire , Fv15hp Subwoofer , Hsu Vtf 15h Subwoofer , Klipsch Sw 311 , Seaton Sound Submersive H P , Svs Pb12 Nsd Black Vinyl 12 Inch Powered Subwoofer , Svs Sb13 Ultra Piano Gloss 13 Inch 1000 Watt Powered Subwoofer , Velodyne Eq Max 15 15 Subwoofer
Thread Tools


Forum Jump: 

Posting Rules  
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off