Page 1 of 1

SOT: Latency Testing

Posted: Sun Jul 27, 2025 11:32 am
by mr_es335
Good day,
The following information may NOT be of interest to some...but for to others, such information...just might be! Thus, you are free to skip this post altogether.

However, for those that might be interested...read on...
* Note: I have provided both references and images of my current research efforts...

In the SAC User Manual [v1.4 ], 3.1.6 Tweaking Windows For Better Performance, Windows 10 (64 Bit), 12) Disable items in Task Scheduler From Auto-Starting, page 29, Bob lists a number of "tasks"...some 125 [1], to be exact..., that he "suggests" disabling.

Thus, after manually going through disabling each of these 125 entries, I decided to do some research into precisely what Windows Task Scheduler really-and-truly IS. Here is what I have discovered...

1) What is Windows Task Scheduler?
Response: "The Task Scheduler service allows you to perform automated tasks on a chosen computer. With this service, you can schedule any program to run at a convenient time for you or when a specific event occurs. The Task Scheduler monitors the time or event criteria that you choose and then executes the task when those criteria are met." [2]

Observations
1) Windows Task Scheduler is a "service".
2) Windows Task Scheduler "...allows you to perform automated tasks on a chosen computer..."
Question: What if I do NOT want to perform any such automated tasks on my chosen computer?
Answer: I do NOT want to perform any such automated tasks on my chosen computer?

2) I came across this snippet [3], wherein, postanote stated, "ScheduledTasks are just xml files which are stored here: C:\Windows\System32\Tasks"....and "By Golly!"...that is what these entries are!! Interesting!

Image

3) So, what is an "xml" file?
Response [4]: An XLM file contains one or more macros used for automating processes in Microsoft Excel, a spreadsheet application. It is a script used to perform complex or repetitive Excel functions, such as inserting data in cells or formatting charts.

More Information
Microsoft introduced XLM files in 1992 with the release of Excel 4.0. The files were commonly used by Excel users until the late-2000s, when Microsoft released Excel 2007. Excel 2007 enabled users to save macros with spreadsheets and save those spreadsheets as .XLSM or .XLTM files.
* Note: Though that above article - even though dated, notes that "...While XLM files are not as common as they once were..." I disagree, as I still "encounter" xml files on a very regular basis. For example, Overloud employs xml files, as does Native Instruments, and many others.

To continue...
Anyhow, after performing all of the necessary research, I decided to, based on the information gleaned from stackoverflow, to a) perform a disk image, and b) "nuke" the entire Tasks folder!

I employ a Behringer UMC 1820 USB2 audio interface with the latest drivers, with the buffers set to 128.

Image

I then performed a "clean" installation of SAC, and with SAC configured to 1x128. I then installed Guitar Rig 5...performing the required testing.

The Testing
I configured SAC for 12 Input channels and 1 Output channel.

I inserted 12 instances of Guitar Rig 5 with my custom, overloaded preset configuration.
* Note: One of the interesting features of Guitar Rig 5 is the ability to insert Pre and Post Tape Decks. Inserting a Pre Tape Deck permits you to "load a lick" - thus negating the need to have a performer being involved in the entire testing process. This feature is particularly handy when employing multiple instances of a plug-in...such as what I am doing here with Guitar Rig 5.

In the end, I have 12 instances of Guitar Rig 5, each with a Pre Tape Deck loaded, each with a "lick".

Then I enabled [LIVE] mode.
* To begin with, for whatever reason, when enabling [LIVE] mode the first time, I always end up with 1 slipped buffer. If I then disable and then re-enable [LIVE] mode, the buffers reset to 0.

Anyhow, after some 3 days, and 12 hours each day testing out this configuration, I have had NO issues whatsoever.

Image

This begs the question then, "What does this all mean with regards to Windows Tasks?" In my humble opinion..."Not much!"

Observations
1) SAC, as noted, loads with 1 slipped buffer, at 57% load.
2) After 20 minutes or so, and rather intermittently I must say, the buffers jump to 5-7 slipped buffers.
3) All that I CAN say thus far is that I would NEVER employ SAC in this manner.
4) My employment involves, 1 instance each of an Emcee, a Nylon Strung, a Steel Strung, and an Electric Guitar with Backing Tracks...and that is it!

In closing...
I have "other" yet tests to perform - for example, the Windows "Services"...but thus far I am very pleased with the end result that I have been experiencing.

For those that might wish to attempt such tests, please employ disk imaging at the outset...and during the entire testing process.

Any comments on the above would be greatly appreciated.

As always, I hope that the above information will be of some practicable employment?

Thank you for your time....appreciated!

References
[1] Windows 10 LTSC lists 173 entries in total, whilst Windows 10 Pro lists some 210 entries.
* On the other hand, Windows 7 Pro has 55 tasks!!
[2] https://learn.microsoft.com/en-us/windo ... -scheduler
[3] https://stackoverflow.com/questions/56926116/how-to-delete-folder-from-task-scheduler-with-powershell
* See Item 8

* Jul 8, 2019 at 6:24
[4] https://fileinfo.com/extension/xlm

Re: SOT: Latency Testing: UPDATE!!

Posted: Sun Aug 03, 2025 4:15 pm
by mr_es335
Good Day,

As a "Heads Up!"...after some 6 days now [07/27/25], the complete removal of the
"C:\Windows\System32\Tasks" folder has had absolute NO side-effects as a result of that complete removal!

I am able to run SAC at a lower latency and with fewer "buffer overruns" as well!

If interested, I will keep you all updated!

Re: SOT: Latency Testing: UPDATE!!

Posted: Thu Aug 07, 2025 10:49 am
by Ruben VS
mr_es335 wrote:Good Day,

As a "Heads Up!"...after some 6 days now [07/27/25], the complete removal of the
"C:\Windows\System32\Tasks" folder has had absolute NO side-effects as a result of that complete removal!

I am able to run SAC at a lower latency and with fewer "buffer overruns" as well!

If interested, I will keep you all updated!
I find this interesting. I'm going to try it next week to see if it improves my system. Thanks.

Re: SOT: Latency Testing: UPDATE!!

Posted: Wed Aug 20, 2025 7:42 am
by Ruben VS
mr_es335 wrote:Good Day,

As a "Heads Up!"...after some 6 days now [07/27/25], the complete removal of the
"C:\Windows\System32\Tasks" folder has had absolute NO side-effects as a result of that complete removal!

I am able to run SAC at a lower latency and with fewer "buffer overruns" as well!

If interested, I will keep you all updated!

I've been trying to delete that pesky Tasks folder in system32 for a week now, but I haven't been able to. Could you tell me how you did it?

Re: SOT: Latency Testing: UPDATE!!

Posted: Wed Aug 20, 2025 9:05 am
by mr_es335
You have to "Take Ownership" of the folder first.

See this...[Click_Me]

Re: SOT: Latency Testing

Posted: Mon Oct 20, 2025 1:04 am
by operationwhat
Hello Dell,

great info...
does this also helps in win7?

Re: SOT: Latency Testing

Posted: Fri Oct 24, 2025 10:49 am
by Warren
Wow Dell
Thanks for this
Working GREAT on Win 11
Now I am (Well my system, not me) is stable at a buffer of 32 lowest setting, when before I had to be at 128 buffers.
Two ea. HDSP9652 interfaces.
Intel I7 64 gig ram
I also use Process Lasso so I had to configure the new empty "Tasks" folder so it would at least load that but nothing else.
I just made a copy of the old Tasks folder renamed it Tasks old and copied only what I needed to be running to the new Tasks folder.
Loving this new setup. :) :cool:

Thank you Sir.

Warren

Re: SOT: Latency Testing

Posted: Fri Oct 31, 2025 7:19 pm
by mr_es335
I do not see why not?!?

PS: By-the-way...all are VERY welcome!!

So, WIN100, Huh! Not for this camper!!!

Re: SOT: Latency Testing

Posted: Sat May 30, 2026 5:20 pm
by Donnie Frank
Hello! Great information here. I will just add a couple comments.

Note: My information and experience is with legacy SAC (Win XP), so this information may be dated. But my gut is that the basic SAC engine hasn't changed much over the years.

Information: When I did latency testing, buffer testing and CPU testing, it didn't matter if I passed audio or not. The results were exactly the same.

My theory: This makes sense to me, as you're still passing the same information regardless of audio levels or variety. With no audio, the VOLUME of data doesn't change. There are never null values. There are only zero values. Only the values of said data changes. e.g. Passing and processing a bunch of zeros at a given bitrate doesn't differ from passing a gaggle of varying values.

I would invite anybody to confirm or refute this. I have no ego in this.

Note: This "no audio theory" may not apply to plugin testing, but I don't see why not. I assume plugins processes data much the same as SAC does.

D

Re: SOT: Latency Testing

Posted: Sat May 30, 2026 8:55 pm
by cgrafx
When the audio engine is live it is passing audio even if the levels are all zero and there is no audio present on the inputs. All the audio engine machinery is still functional. Think of it as a conveyer belt moving boxes that contain audio samples, The belts and boxes are still moving even if the boxes are empty.

Re: SOT: Latency Testing

Posted: Sun May 31, 2026 3:30 pm
by Donnie Frank
cgrafx wrote: Sat May 30, 2026 8:55 pm When the audio engine is live it is passing audio even if the levels are all zero and there is no audio present on the inputs. All the audio engine machinery is still functional. Think of it as a conveyer belt moving boxes that contain audio samples, The belts and boxes are still moving even if the boxes are empty.
Well said.

D