Home Theater Forum and Systems banner

41 - 60 of 82 Posts

·
REW Author
Joined
·
6,711 Posts
I see the measurement sweep signal is modified a little so I intend to update my external sweep file accordingly.
The measurement signal hasn't actually changed, just the way that start and end are shown in the signal generator controls to make them consistent with the measurement window settings.

I looked to see if the measurement sweep export feature now includes a new option to save a stereo version of the sweep with the new acoustic timing sweep on the reference channel. I don't see that option yet. I am thinking that it will make it possible to run external sweeps that provides the capability of the new acoustic timing option.
The meas sweep export needs an additional dialog to choose whether or not to include a timing signal and where, I haven't done that yet. It would also lend itself to a mode where REW can be set to start capturing and wait to see the timing signal before capturing the measurement, removing the complication of trying to time things by hand. On my todo list :)
 

·
Registered
Joined
·
195 Posts
Very nice indeed, John! Now I need to go back and research all of the instances in which my measurement accuracy was called into question because of the lack of a timing reference. Serious homework to do!

And I feel an update to my beginners REW guide will be coming soon... :smile:
 

·
Registered
Joined
·
24 Posts
Thank you very much. With this addition andyc56 (the creator of Multi-sub optimizer) can make his software work with USB mics which makes it much easier.
Hi all. I'm the guy who wrote the Multi-Sub Optimizer (MSO) software.

I mentioned this in another thread, but it's worth repeating here. I used to have big, bold warnings about using MSO with USB mics, which I've since updated to reflect the current REW status with regard to the new acoustic timing reference feature. I had those warnings there because if one attempted to use measurements taken with a USB mic without an acoustic timing reference, MSO would appear to work fine, but give wrong results. This is an insidious type of "silent failure" error that needs to be clearly warned against in the documentation, especially considering that USB mics now dominate the measurement scene. There are actually no modifications to MSO required for it to give correct results when used with measurements taken with a USB mic, as long as one uses the acoustic timing reference when performing the measurement.

Kudos to John, not only for the quality and breadth of features of REW, but also for his amazing commitment to it.
 

·
Premium Member
Joined
·
13 Posts
JohnM, many thanks for the acoustic timing reference addition for USB microphones. I had the chance to do some measurements today - in preparation for some MSO calculations (thanks andyc56) - and the new system is so much easier to use, previously I had to crawl behind my AV cabinet to disconnect speaker cables and connect up some pre-amp outputs to a Behringer ADC.
The quick high frequency sweep gave identical timing accuracy to my previous method - so I knew I was good to go!!!!

Now, how about a quick access panel for selecting the HDMI output channel, with inc/dec buttons???????????

Regards, Mike.
 

·
REW Author
Joined
·
6,711 Posts
A bit more progress on the acoustic reference support, V5.15 beta 4 (available from the installers directory) has added the ability to include the timing reference signal when saving a measurement sweep as WAV and there now is a "Wait for timing reference" check box on the measurement dialog. When that box is checked REW will delay capturing the main measurement sweep until it has seen the timing reference signal, which allows coping with unlimited amounts of latency and makes it easy to measure by replaying the measurement sweep WAV if the REW output cannot be used directly.

I've also made some small changes to the waterfall and spectrogram plots.

The changes compared to beta 3 are:
  • Added option to add timing reference signal when saving measurement sweep as WAV
  • Made the signal generator export filenames more descriptive
  • Show a warning if using a timing reference with ASIO but haven't selected a timing ref output
  • Removed the fade to graph colour at the bottom of the spectrogram to better define the lower parts of the plot
  • Restored Violet as the bottom colour for the Rainbow spectrogram colour scheme
  • Added a cursor frequency line over the surface of the waterfall plot
  • Added a measurement option to wait for the timing reference signal to be detected when using an acoustic reference
  • Bug fix: Spectrogram contours were not redrawn when plot range was changed
  • Bug fix: Frequency axis grid lines were not drawn on floor of waterfall when using linear frequency axis
 

·
Premium Member
Joined
·
1,562 Posts
JohnM,
Thanks!
I spent the day figuring out how to get this setup correctly. I am using the 'acoustic timing' feature and the 'hold for trigger' function when using an external signal from my DLNA file server. After some confusion and dumb mistakes I finally got it all sorted out. I now have accurate IR timing with foobar fed REW signal files and can thus confirm that close phase tracking is properly retained when the ReRhase FIR phase correction filter activated. Very nice!
 

·
Registered
Joined
·
21 Posts
Discussion Starter #48
Had some success and some failure with this so far. One one of my setups it worked flawlessly once I figured it out. On a second setup I get the test signal also coming from the channel I have setup as the timing reference. Obviously a full range signal from the tweeter :( Might be I've messed something up in the setup but pretty sure I mirrored the working setup on the other PC. I used the same gear.
 

·
Registered
Joined
·
195 Posts
John, just to be clear, when selecting the channel to use for timing reference, there is no requirement that the speaker on that channel be closer to the MLP than the speakers being measured, is there? For example, when measuring my left, right, and center channels, I am using a side surround as the timing reference. That speaker is ~.5 ft further away than the mains from the MLP.
 

·
Registered
Joined
·
261 Posts
A limitation of this initial implementation is that the reference channel is only used for the timing signal, it can't be measured while using it as the timing reference (e.g. if the left channel output is being used for the timing ref, the measurement sweep will only be on the right channel output).
Would you expect a meaningful increase in precision over just measuring that channel without a reference? Put another way, what are the variable latency sources that this feature would eliminate (in the case of measuring the reference channel)?
 

·
REW Author
Joined
·
6,711 Posts
Using the timing reference doesn't make any difference to the accuracy of the measurement, the method copes with static latencies. In any case, I have removed the restriction on measuring the ref channel in beta 6, which is now in the usual place.

Changes in beta 6:
  • Added a left/right output channel selector in the Soundcard preferences when using Java drivers, meas sweep only appears on selected channel unless using loopback, when it appears on both
  • When using acoustic timing reference can now also measure the timing ref channel
  • Added an equaliser setting for the Anti-Mode 2.0 Dual Core
 

·
REW Author
Joined
·
6,711 Posts
A few more changes for beta 7:

  • Rearranged the input and output selectors on the Soundcard preferences panel
  • Added a "Both" option to the output channel selector when using Java drivers
  • Added output selectors on the measurement dialog
  • Added display of timing reference headroom below the timing ref level control on the measurement dialog
  • Added the timing reference mode and timing reference output name to the measurement Info panel
  • Added a display of the CEA-2010 peak level on the RTA graph
 

·
Premium Member
Joined
·
13 Posts
Although I have not had a chance to try out Beta7 (with output selection on the Measurement dialog) I felt that a 'like' simply did not go far enough.
Many, many thanks John for all the steady improvements to REW - absolutely unbeatable.

Best regards, Mike.
 

·
Premium Member
Joined
·
13 Posts
I will generously assume that other people have already realised an issue with using the new acoustic timing reference - but just in case you have not, I'll try to explain what I have realised.

A proper electrical timing reference will generally allow you to measure the flight time from speaker to microphone. Moving the mic relative to the speaker will show a change in timing due to the distance change - all good, and expected. Moving the mic along an arc, while keeping the speaker distance constant - will show no change in timing, again all good, as expected. Now try the same thing with an acoustic timing reference (using a surround speaker as your reference for example) - if the mic remains stationary the relative timing of various speakers will be accurate - but if you move the mic relative to the reference speaker, all the timings will change - even if you maintained a constant distance to the measurement speakers.

Therefore, if taking single point measurements - acoustic timing works just as well as electrical. But if you plan on taking multiple position samples (that require accurate timing) - don't use acoustic, stay with electrical. It is possible to correct for the reference movement - but it's a bit of a pain. I will grudgingly return to electrical loopback for my next session.

I'm not being critical of John's work - it works as intended, I've used it - just trying to let others know of an inherent issue. That's if others can understand my ramblings :smile:

Regards, Mike.
 

·
Registered
Joined
·
24 Posts
Therefore, if taking single point measurements - acoustic timing works just as well as electrical. But if you plan on taking multiple position samples (that require accurate timing) - don't use acoustic, stay with electrical. It is possible to correct for the reference movement - but it's a bit of a pain. I will grudgingly return to electrical loopback for my next session.
Hi Mike,

Are you AV_mike over at AVS? I was asking because you mentioned MSO in an earlier post. Anyway, I thought about this very thing a few weeks ago. At first, I thought it might invalidate MSO results. But then I realized that the acoustic timing reference allows the summation of complex acoustic pressure amplitudes of multiple subs and main speakers taken individually at a given listening position to match what the measurement would be if it were performed with simultaneous excitation. Since each summation of an MSO measurement group normally corresponds to only a single measurement position, the integrity of each summation (the MSO measurement group graph trace) is preserved. If the measurements in a given MSO measurement group were taken at different listening positions, there will be a problem though, and the summation won't be valid. MSO is flexible and allows measurement groups to be defined in this way, but such a summation has no useful physical interpretation that I am aware of.

So it's all good as far as normal usage of MSO is concerned. Normal usage would be the typical scenario for which all the measurements in each MSO measurement group were all taken at the same listening position. Taking phase differences between measurement group summations won't be valid, but I can't see any reason to do that. Also, MSO uses only the magnitude, and not the phase, of the complex summation of the components of a measurement group in its calculations.
 

·
Premium Member
Joined
·
13 Posts
Hi Andyc56,

This is my alter-ego...............................

I must admit, I thought MSO needed the phase to perform the complex summation - as phase and delay share similarities, and would affect how two or more sources interact within the room?

I have been using MSO to combine dissimilar subs - never even got close trying to do it manually - many, many thanks Andy.
I have also been using MSO to get the best sub-sat splice in two listening positions - this is where microphone movement is required. I noticed that my center channel impulse had shifted by over four feet - and yet the center speaker is equi-distant from each listening position. The change was because the mic moved relative to the surround speaker (timing ref source). My initial MSO analysis gave some pretty poor results - until I manually corrected for the impulse timing movement.

While you're here Andy, any changes to MSO on the horizon? I have tried the Target Curve option - but it does not operate as I expected (I'll try to make some notes). A permanently visible RMS error display, and iteration counter would be useful.

Regards Mike (Mike2001, AV_mike)
 

·
Registered
Joined
·
24 Posts
I must admit, I thought MSO needed the phase to perform the complex summation - as phase and delay share similarities, and would affect how two or more sources interact within the room?
Yes, MSO does need the phase to perform the complex summation. However, it does not require that the absolute delay of each measurement be correct. It only requires, for a given complex summation, that the measurements preserve the delays relative to one another of each speaker or sub measured.

This is not unlike how an AVR works internally. An AVR uses zero delay for the most distant speaker to the MLP, then delays closer speakers based on their relative distance to the MLP from the most distant one. So the AVR only uses the relative delays for this calculation. It's a similar idea with MSO.

I have also been using MSO to get the best sub-sat splice in two listening positions - this is where microphone movement is required. I noticed that my center channel impulse had shifted by over four feet - and yet the center speaker is equi-distant from each listening position. The change was because the mic moved relative to the surround speaker (timing ref source).
If all else is done correctly, this should be okay, as the sub distances should all shift by four feet also, thus preserving the delays relative to one another of each speaker and sub. The latest REW beta allows using a channel under test as the acoustic timing reference, so if this shift in center channel distance bothers you, you can use the center as the timing reference.

My initial MSO analysis gave some pretty poor results - until I manually corrected for the impulse timing movement.
I would absolutely recommend against any second-guessing of REW data by manual time shifting in this way. As long as the measurements were performed correctly, and there's no bug in this new REW feature, REW should take care of all of this internally. In this case, "taking care of it" should manifest itself in a four-foot shift in the impulse responses of the subs at the new position, preserving the relative delays of the subs to one another and to the center channel.

While you're here Andy, any changes to MSO on the horizon? I have tried the Target Curve option - but it does not operate as I expected (I'll try to make some notes). A permanently visible RMS error display, and iteration counter would be useful.
For the time being, I'm not working on it, as I'm dealing with home repairs due to water damage from an unfortunate plumbing incident. I'll add those requests to the list, and I have some other planned improvements too. But they may be a while. I can help you with the target curve in the main MSO thread.
 

·
REW Author
Joined
·
6,711 Posts
Not strictly related to acoustic timing reference, but I have extended the output selection to the signal generator and made a few other changes in beta 8, available in the usual place.

  • Added output selection to the signal generator
  • Added bit length selection to signal generator WAV export (16, 24 or 32-bit WAV files can be generated)
  • Rearranged the signal generator controls
  • Added next and previous 1/3rd octave frequency buttons to the CEA burst generator
  • Added the output name to the measurement Info panel
 

·
REW Author
Joined
·
6,711 Posts
I have uploaded beta 9 to the installers directory. This version is earlier than intendede due to an error in the way beta 7 and beta 8 were built. Files made in beta 7 will not open in beta 8 or later and files made in beta 8 will not open in beta 9 or later. The workaround is to open files made in beta 7 or beta 8 in beta 6 or earlier, then save them again. The saved files will open in any later version. There are a couple of small changes in beta 9, long file names should not cause wrapping in the measurement panel and the CEA peak value turns red if any harmonic limits are exceeded. I have removed beta 7 and beta 8.
 
41 - 60 of 82 Posts
Top