Dropouts in Stepped Sine Measurements - 5.20RC3 and 5.20RC4 on macOS

[sam_adams]

@John Mulcahy I have been testing the new release candidates and have been having an issue with signal dropouts in the stepped sine measurements. I have attached images to look through to illustrate the problem. These were tested on a MacBook Pro running 10.11.6 and an iMac running 10.13.6. I have been using both of these Macs with REW for quite some time with excellent results on both. All of these measurements were made using the builtin audio on the Macs with a simple pair of loopback cables. I haven't yet had a chance to test the latest RC on my Mac Minis. The i7 mini that I was using for REW is now doing duty as my main Mac since my MBP 17 inch finally died as a result of the GPU issue.

An interesting contrast to the current RC issue is I have also tested the build that you posted in this thread where you enabled a flag in the build to try and fix the issue that the posting member was having. That particular build does not have the dropout issue as can be seen from the attached images marked as 520AT. The dropouts in the RCs also occur when using the signal generator to generate a stimulus signal. The odd thing is that if I use an external application like Audacity to generate a stimulus signal, the dropouts do not occur.

As a small bright spot, the issue that I brought to your attention in this post regarding the disappearing UI elements on the i7 Mini with the Intel HD Graphics 3000 adapter, seem to have been resolved with the newer JRE that the latest builds are based on. Small steps.

 

[John Mulcahy]

Here is a build of RC4 without the flag to try (the audio test version removed the flag rather than adding it): http://roomeqwizard.com/REW_macos_5_20_RC4_noflag.dmg

Other things can also cause problems during a test of course, such as heavy network activity (backups, for example), background updates, other applications making high resource demands etc. All runs OK on my creaking late 2012 i5 Mac mini running 10.15.7, so processor capability shouldn't be a factor. I did have to max out the memory in the mini (16 GB) to get it to run more recent macOS builds decently though.

 

[sam_adams]

John, the test version works correctly. No dropouts—see attached. Thank you for addressing this so quickly.

As a point of reference from an REW performance perspective, the MacBook Pro I have been using is a mid-2009 Core 2 Duo MBP, 2.53 GHz, 8GB RAM,. When testing with REW, it's the only application running and I've never seen the processor running over 15%. It seems that REW really doesn't stress the hardware very much.

One thing about the RTA window though, if you look at the all the attached pics, you'll see that the title bar indicates LINE IN L, but all of these tests were done with the Right channel as the selected input channel.

Also, is there a way for you to change the level setting in the stepped sign settings dialog to allow for two decimal places to the right of the decimal point?

 

[John Mulcahy]

I have fixed the input channel in the title bar for the next build. I have also changed the stepped sine level controls to show 2 decimal places, but that means the values will change in steps of 0.01 dB when the arrow buttons on the spinners are used.

 

[sam_adams]

Thank you, John, for all the hard work. Looking forward to hopefully the last RC and the release of 5.20.