REW macOS DMG (includes private Java 8 JRE)

REW Beta Release REW macOS DMG (includes private Java 8 JRE) V5.20.3

You Must Login or Register to Download

hestepare

Registered
Joined
Apr 24, 2020
Messages
12
Great updates, John. A minor suggestion: as I'm an avid reader of the help file I notice that scroll behaviour is different from regular MacOS apps. It's much more sensitive and a little finicky to use.
 

John Mulcahy

REW Author
Thread Starter
Joined
Apr 3, 2017
Messages
5,542
John Mulcahy updated REW macOS DMG (includes private Java 8 JRE) with a new update entry:

REW V5.20.1 macOS DMG

This is the official release of V5.20, the version number is 5.20.1 to distinguish it from the release candidates. It will only be available here for the next week or so, then I will also add it to the REW website and enable the version detection to report it as available on REW startup. It has taken a tad longer than I expected to release 5.20 (more than 2 years longer!) but on the plus side it does have a lot of new features. :)

There are a few changes from V5.20 RC15:

  • There is...
Read the rest of this update entry...
 

sam_adams

Member
Joined
May 23, 2017
Messages
174
Congratulations on the release, John. It has indeed been a long slog. Perhaps a well deserved holiday is in order.
 

Theo Aronson

Registered
Joined
Sep 21, 2021
Messages
5
Hi John - new to the club here, but can't figure out for the life of my how to UNINSTALL v5.19 which I accidentally installed on 10.15.7. Can you advise?
 

Theo Aronson

Registered
Joined
Sep 21, 2021
Messages
5
You can just drag it to the trash.
Thanks for the quick reply, John. There aren't any other preference files or legacy java files I need to remove as well? I saw the following .plist files were modified as well in my Library but does the REW node remove itself upon deletion? Can I delete all four of these files as well before I install 5.20.1?

com.apple.java.util.prefs.plist
com.install4j.installations.plist
com.install4j.4549-9647-2313-4375.uninstaller.plist
roomeqwizard.launcher.plist

To clarify, just trying to give computer a scrub before I install any other files to avoid having unused files sitting in OS Library.

Cheers,
Theo
 

John Mulcahy

REW Author
Thread Starter
Joined
Apr 3, 2017
Messages
5,542
Both versions use the same Preferences node. If you have any other Java application that saves preferences those nodes will also be in the java.utils.prefs.plist file.
 
Last edited:

Theo Aronson

Registered
Joined
Sep 21, 2021
Messages
5
Both versions use the same Preferences node. If you have any other Java application that saves preferences those ndoes will also be in the java.utils.prefs.plist file.
Copy that. Thanks again for quick reply (and for creating this software) -- about to dig in to 5.20.1!
 

John Mulcahy

REW Author
Thread Starter
Joined
Apr 3, 2017
Messages
5,542
John Mulcahy updated REW macOS DMG (includes private Java 8 JRE) with a new update entry:

REW V5.20.2 macOS DMG

V5.20.2: A small update to add some checks to detect invalid impedance calibration results and fix a cosmetic bug on the Measure dialog.

  • Check for mono input data in impedance measurements
  • Check scale of measurement input vs ref input in impedance short circuit calibration to detect invalid results
  • Bug fix: Rinput and Rleads labels and controls could appear on the Measure dialog when measuring SPL
Read the rest of this update entry...
 

KaTsCorner

Registered
Joined
Feb 24, 2021
Messages
2
My AV System  
Computer Audio
Digital Performer
I just noted the reference in the release announcement of 5.20.1, where it says:

* There is now a "dB Average" button in the All SPL graph controls that averages the dB values of the selected traces (e.g. +5 dB and - 5 dB average to 0 dB).

I'm wondering if that is a misprint, or if you are averaging something I'm not grasping. When I do the math (via logs and powers of 10), +5dB and -5dB do not average to 0dB, they average to +2.4dB. I'm just seeking a little clarification, thanks! :-)
 

KaTsCorner

Registered
Joined
Feb 24, 2021
Messages
2
My AV System  
Computer Audio
Digital Performer
Ahh. So you are intentionally taking an arithmetic average from among a series of exponents. Good to know. Since that process effectively yields the Nth root of the product of N numbers, (please correct me if I am wrong) the solution it seems to seek is, "Given N distinct sources at various SPLs, find the SPL for a hypothetical new source which would then by itself present a level equal to the average of the SPLs of the other distinct sources, none of which being measured in combination with any other." I can see that having value in, for example, trimming individual loudspeaker output levels to achieve equitable power sharing either across channels or across individual array elements within a channel. Most times my focus is on the combined effect in the room of all elements taken together, so I have thus far used formulae which combine rather than individualize component contributions. I am intrigued however to apply this approach in an effort to smooth element-to-element interaction before addressing the full system. Thank you!
 

John Mulcahy

REW Author
Thread Starter
Joined
Apr 3, 2017
Messages
5,542
Well, I can't recall having ever asked myself that question :) but yes, it produces a geometric average.
 

John Mulcahy

REW Author
Thread Starter
Joined
Apr 3, 2017
Messages
5,542
John Mulcahy updated REW macOS DMG (includes private Java 8 JRE) with a new update entry:

REW V5.20.3 macOS DMG

Changes in V5.20.3:
  • Remember and restore the state of the ASIO secondary output selection
  • Allow smoothing to be applied to RTA data (previously could only be applied to Spectrum data)
  • Added a Resistance trace to the Impedance & Phase graph to show the real part of the impedance
  • Bug fix: If multi-input capture was selected soundcard calibration would stop on the Measure dialog
  • Bug fix: The watermark text field in the View settings was blank after restart
  • Bug...
Read the rest of this update entry...
 

Chuck Zwicky

Registered
Joined
Sep 8, 2017
Messages
61
Experiencing odd timing issues with 5.20.8

OSX 10.13.6

When performing a measurement, the progress bar will often get halfway through before the sweep oscillator begins, and measurement is truncated before the sweep is finished, leading to erroneous results.

I've tested several interfaces and had the same issue.

EDIT: Experimenting with Buffer settings now, seems that it is working only when using a 32k buffer and multiple sweeps
 
Last edited:

John Mulcahy

REW Author
Thread Starter
Joined
Apr 3, 2017
Messages
5,542
Sounds like there is a high latency somewhere in the replay path, best solution for that is to use the acoustic timing reference so REW waits to hear the timing signal before starting its capture.
 

Chuck Zwicky

Registered
Joined
Sep 8, 2017
Messages
61
Ah, yes, excellent idea, will try and report back. The latency isn't more than a few ms. Why would it work only with particular buffer settings and only when using multiple sweeps?

EDIT:

I tried again using the acoustic timing reference, but what happened was it did a sweep then got stuck here (at 9%):
52511
 
Last edited:

John Mulcahy

REW Author
Thread Starter
Joined
Apr 3, 2017
Messages
5,542
The latency isn't more than a few ms. Why would it work only with particular buffer settings and only when using multiple sweeps?
Buffers need to be filled. It doesn't work with multiple sweeps, you will be getting parts of different sweeps incorrectly merged.

EDIT:
I tried again using the acoustic timing reference, but what happened was it did a sweep then got stuck here (at 9%):
Per the message, it is waiting to hear the timing reference, which is a short high frequency sweep. Is output channel 1 (the ref output) connected to a speaker that has a tweeter?
 

Chuck Zwicky

Registered
Joined
Sep 8, 2017
Messages
61
All connections are electrical.

Loopback is channel 1, Ref is channel 2.

I am measuring a filter in a hybrid circuit which involves an A/D--->DSP--->D/A---Analog Filter--->A/D--->DSP--->D/A the overall latency is 6ms.

Further experiments here show that clean sweeps are possible only when using no timing reference.

This is what a sweep looks like when using the loopback timing reference



52512




No timing reference:


52514
 
Last edited:

John Mulcahy

REW Author
Thread Starter
Joined
Apr 3, 2017
Messages
5,542
"Ref output" on the Measure dialog is where REW sends the acoustic timing reference signal, which on the image you posted was channel 1.

52513


If that goes through a filter that doesn't pass 5 kHz - 20 kHz (the range of the timing ref signal) and/or it doesn't come back on the measurement input (which is channel 2 on the dialog image) then REW won't detect it. If you use a loopback timing reference instead the ref output and ref input specified are used, but excessive latency would still be an issue.

It is difficult to reconcile "overall latency is 6 ms" with "the progress bar will often get halfway through before the sweep oscillator begins". When not using the timing reference REW starts the sweep (sends audio data to the OS) and starts capture (requests audio data from the OS) at the same time. The output has 100 ms of dither before the sweep starts, and there is 500 ms of silence after the sweep ends. The capture continues for a further 1 second after that, so losing the end of the sweep means latency is more than 1.5 seconds.
 

Chuck Zwicky

Registered
Joined
Sep 8, 2017
Messages
61
I swapped the input and reference channels for the last 2 measurements above - both physically and in the preferences. No difference.

I'll examine further.

It is difficult to reconcile "overall latency is 6 ms" with "the progress bar will often get halfway through before the sweep oscillator begins". When not using the timing reference REW starts the sweep (sends audio data to the OS) and starts capture (requests audio data from the OS) at the same time. The output has 100 ms of dither before the sweep starts, and there is 500 ms of silence after the sweep ends. The capture continues for a further 1 second after that, so losing the end of the sweep means latency is more than 1.5 seconds.
Yes, this seemed to depend on the REW buffer setting, at settings higher than 64k the progress bar can get further than 50% before the first sweep happens, and any subsequent sweep just gets cut off. The only accurate measurements were made with no timing reference, or only when using a 32k buffer and multiple sweeps.

EDIT: Just changed the REF and Input settings again, managed a clean sweep, then this:

52515
 

Chuck Zwicky

Registered
Joined
Sep 8, 2017
Messages
61
Using these settings:

52516




52517


Runing an initial sweep gave a horribly fuzzy response.

BUT... using that sweep to determine change the offset:
52518


Resulted in a very clean sweep..

Before and after.. (offset for clarity).
52519



QUESTION: I have been using REW for years and never been confounded like this, is 6ms too much latency? the direct sweeps (bypassing the DUT) were all perfectly clean. What was I doing wrong here? Why were some buffer settings working and some not? I am not confused about the timing reference routing, etc, I've made hundreds of useful measurements with REW..
 
Last edited:

John Mulcahy

REW Author
Thread Starter
Joined
Apr 3, 2017
Messages
5,542
EDIT: Just changed the REF and Input settings again, managed a clean sweep, then this:

View attachment 52515
Please provide the details of the error, or attach a zipped copy of the REW log files folder.

QUESTION: I have been using REW for years and never been confounded like this, is 6ms too much latency? the direct sweeps (bypassing the DUT) were all perfectly clean. What was I doing wrong here? Why were some buffer settings working and some not? I am not confused about the timing reference routing, etc, I've made hundreds of useful measurements with REW..
No, 6 ms is not too much latency, less than 1500 ms is fine. Please attach the measurement mdat files.
 

John Mulcahy

REW Author
Thread Starter
Joined
Apr 3, 2017
Messages
5,542
Can you attach the mdat files for the responses that were cleaner after changing the timing offset, as posted in your "Before and after (offset for clarity)" image? Why aren't you running the current REW version?
 
Top Bottom