Qlab is forgetting the video output on my MacBook

[John Nolan]

I'm setting up a simple slideshow to run on, and be displayed on, a 2019 MacBook Pro, running Sonoma 14.7.1. (To be clear, it's not going to be displayed on an external device.)

I'm using Qlab 5.5

I've assigned the output to the built in display, but if I quit and reopen, I get the warning that there's no output device assigned. I re-assign the display, and QLab reports no warnings. But when I reopen the file, the warnings reappear. 

Other changes to the file persist.

I'm not encountering this issue on my Mac Studio.

What am I missing?

Thanks,

John Nolan,

Rag & Bone Puppet Theatre

[Chris Ashworth]

Hi Johns,

We’re currently working to track down this issue.  We’ll update the list once we know more, or may reach out directly if we have questions.

Sorry for the trouble in the meantime!

-C

--
Contact support anytime: sup...@figure53.com
User Group Code of Conduct: https://qlab.app/code-of-conduct/
 
Instagram: https://www.instagram.com/Figure53
Bluesky: https://bsky.app/profile/qlab.app
---
You received this message because you are subscribed to the Google Groups "QLab" group.
To unsubscribe from this group and stop receiving emails from it, send an email to qlab+uns...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/qlab/cd267195-133e-4394-a8b9-46b5d245db4an%40googlegroups.com.

[John Nolan]

Thanks Chris.

To view this discussion visit https://groups.google.com/d/msgid/qlab/CADv1a486qT%3DZMOUDgs8oD6raB%3D3HsAbzuLxqgiF3i2kXc%2BSSPg%40mail.gmail.com.

[Chris Ashworth]

Just to give a little more context for what’s happening:

Here are the current changes between 5.4 and 5.5:

- Name matching ignores suffix numbers in 5.5. In 5.4, QLab would see "LCD Projector (1)" and "LCD Projector (2)" as two different names for matching. This led to problems because macOS would sometimes swap those numbers on reboot or reconnect. 5.5 lops off the number, treating both of these as having the name "LCD Projector”.

- 5.5 keeps track of whether we've seen outputs with duplicate names. If we do see that, we save that state and disable name matching.

In both 5.4 and 5.5, this part of the algorithm is the same:

- Match on name

- Match on display ID. Display ID was the thing Apple told us to use, but they now admit can change upon reboot.

So, I think what we're seeing here is folks who have duplicate devices and display ID not behaving. In 5.4, these duplicate devices will always be reconnected using name-based matching, though sometimes they'll be reconnected backwards because macOS is unreliable. In 5.5, we don't use name-based matching for duplicate devices, and if display ID matching also fails, then we fail to reconnect.

The whole situation is frustrating, because Apple for a while pretended to give us a simple number that we could use to track displays, and then later admitted that it didn’t work, and does not provide any official API for doing it reliably. (Which is wild to me.) We’re currently working to add serial numbers into the matching algorithm, but it’s also shocking how many video outputs do not report useful serial numbers. We’ll keep working on it!

Best,

Chris

P.S.: sorry about the typo when spelling your name before, John!

To view this discussion visit https://groups.google.com/d/msgid/qlab/CAMvb2sA5%2BiTQ5cOH6vGOnTShtny5ZW3i%2BL8ATDDdzFmbr9OxxA%40mail.gmail.com.

[John Nolan]

Thanks for the details,  Chris.

So, is there a workaround possible by deleting devices and re-doing cues?

And, as my father used to say, you can call me anything but late for dinner.

To view this discussion visit https://groups.google.com/d/msgid/qlab/CADv1a49cWTKBN-TLt_vOsQ-6_MAD8ifm3GqUMaeMBsWqmvn2pw%40mail.gmail.com.

[Chris Ashworth]

Unfortunately deleting devices and re-doing cues wouldn’t help because the information QLab is getting from the operating system is simply not enough to uniquely identify video outputs, so the new video devices would have the same problems. I know others have talked about adding EDID emulators in the chain to provide actually unique display IDs, although I have not personally worked with these and don’t have strong advice about the pros / cons / pitfalls / best brands… 

(P.S.: Ha!) 

To view this discussion visit https://groups.google.com/d/msgid/qlab/CAMvb2sAV-WjLtkh9NuW%3D8G5%2BChgqXC89jhMwv5YnacvwOrfhJg%40mail.gmail.com.

[jastan...@proton.me]

A year and a half ago I had difficulty using 2 video outputs in QLab 5 on 2 M2 machines until using EDID emulators.

(2x MBPs -> 2x Plugable TB4-UDZ docks -> 2x Roland V1-HDs)

Apple, Plugable and Roland didn't feel responsible. Adding the 4 emulators between the docks and switchers solved the problem. 

This had not been a problem with Intel machines, QLab 4 and earlier OS using the same switchers and docks. Nor is it a problem on my edit machine (QLab 5 on an Intel MBP) with a CalDigit TB3 HDMI Minidock and no switcher.

Currently don't have access to the brand of EDID, but they are externally powered (using an Anker 6x USB-A power supply and adapter cables) and use HDMI jumper cables. Was afraid of using bus-powered EDID dongles and straining the HDMI connectors.

Jim

[Chad Sellers]

John,

I think we have this fixed now. Would you mind testing out this beta release to see if it fixes the issue for you:

https://qlab.app/temp/QLab-5.5.4-b1.dmg

The fixes there seem to solve the problem for others that have had display matching problems in 5.5.x, though they've all been using external displays unlike you. If you still have problems, would you mind reaching out to sup...@figure53.com so we can dive a bit deeper into your system details and your workspace?

Thanks,
Chad Sellers
Figure 53

To view this discussion visit https://groups.google.com/d/msgid/qlab/CAMvb2sAV-WjLtkh9NuW%3D8G5%2BChgqXC89jhMwv5YnacvwOrfhJg%40mail.gmail.com.

[John Nolan]

Yes, that seems to work for me. Thanks for this.

To view this discussion visit https://groups.google.com/d/msgid/qlab/CAPYr5cAaYGVHSniYxCC0C20wW4m_e6_MphZVyuCb3GF89qK-JA%40mail.gmail.com.