Q-SYS Corner

[Steve Guttag]

Having run 9.5.0 for over a month and 9.4.3 for over a week...I can confirm that they have fixed the issue where when a Dolby Atmos processor becomes the PTP GM, the CORE (110) NICs shutoff (happened about 10% of the time) has been fixed. There have been zero incidents on either version. In fact, for our designs, there haven't been any incidents of any kind on either version though the A/V world has been reporting various issues on 9.5.0 so I suspect that a 9.5.1 isn't too far off.

[John Thomas]

There is a guy from QSC arguing the point...that because they are using floating point processing, they can go above 0dB. And those of us on the counter-point is...yes BUT 0dBFS is just that FULL-SCALE. There is nothing higher than that. You can have 20dBu all you want...you cant have 20dBFS. They've been watching too many sports movies where coches tell their players to give 110%. They still don't get that user-faders shouldn't dB at all...that isn't a human way of thinking, let alone working constantly in negative numbers. Its fine and appropriate for internal gain controls...but not for a Volume control.

The concept of clipping and by extension the dBFS scale itself exists because of the limitations of analog devices/circuitry and fixed-point digital audio, but now we have a method of representing and processing audio in a way that effectively cannot clip. Q-SYS would have no problem processing a speaker level voltage plugged straight into an analog input -- if the input circuitry had that much headroom. Once a signal is inside Q-SYS and until it hits an output, its samples are represented as 32-bit floats -- values with not only accuracy identical to 24-bit fixed-point integers in both the positive and negative directions, but also a monstrous dynamic range of 1528 dB (with diminishing accuracy past the +/- 144.5dBFS marks). That's -758dBFS to +770dBFS. Let's say that imaginary unclippable input is DC-coupled. Assuming my math is correct, lightning striking that input would make the peak meter in Q-SYS read around +170dBFS, no sweat.

So yes, when 32-bit floating point audio gets louder than other signals/formats are capable of defining or reproducing, it is represented on the dBFS scale as a positive value. And obviously 0dBFS is still meaningful as the clipping point for signals/formats that do clip like analog inputs/outputs and fixed-point digital inputs/outputs/files. So this is still the most intuitive and practical solution: Use the existing scale and workflow that allows the engineer to keep a grip on gain staging where clipping does apply; and where clipping doesn't apply, levels can exceed 0dBFS because frankly that's exactly what is happening.

[Bruce Cloutier]

Ok so the above argument is confusing but I get the gist. There are advantages and likely disadvantages to representing signal as floating point.

I can imagine a floating point system as sort of a fixed point representation coupled with dynamic range/resolution. If you encode a signal at one fixed level and that approaches a point that would clip in a fixed point system you can bump the exponent and effectively double the range that can be encoded at the cost of resolution (the least significant mantissa bit then represents twice the smallest voltage that was previously encoded). You might think that where you call the 0 point (between negative and positive) might be arbitrary. I wonder what happens then (in terms of quality) in a system that varies the exponent (range/resolution) sample to sample?

In either case you can do what you like in the digital realm. I am certain there is a sophisticated science that has evolved in the industry. And, like anything else, you certainly have the freedom to muck things up as much as improve anything. But isn't the real challenge getting the digital representation back out into the real world as sound (that hopefully most ears pleasantly accept and people unanimously enjoy)? Add to that the ability to acquire audio content from the space around us and get that into the digital format in the first place? I am not even sure that these have ever been done flawlessly.

Using decibels as a scale for level control would make a lot more sense if there were a single standard that the average person could relate to. Yes there is and are standards. But the really nice thing about standards is that there are so many to choose from.

Just saying...

[John Thomas]

In terms of resolution, quantization error, dynamic range, and comparisons with fixed-point DSP, I recommend these four links, in order of least-mathy to most-mathy:

https://youtu.be/xyMUAWXjrhA?t=478

https://www.sounddevices.com/32-bit-...iles-explained

https://www.dspguide.com/ch28/4.htm

https://blog.demofox.org/2017/11/21/...int-precision/


Ultimately, no, I can't say that the ability to exceed 0dBFS without clipping is something that will instill good gain-staging, design, etc. for our type of work, and I do think it will be counterproductive in that particular aspect. You can see it first hand in music production DAW's, where beginners can get pretty far driving levels over 0dBFS without any problems... until they insert a 24-bit effect, or bypass the limiter on the master bus.

Really, I just wanted to nitpick Steve because it's rare that I get a chance to intelligently disagree with him on a technical level, and it's not the first time I've heard the argument. I do recommend reading up on 32-bit floating point DSP though, because it's interesting in its own right as well as sort of a gateway to learning about the problems and workarounds that DSP programmers had to fight with before it was a thing.

Also that QSC guy should have just sent you this:

dynamic_range_chart-768x945.png

[Marco Giustini]

Incredibly interesting and useful thanks. I found the second link more understandable and I could also download the sample files and see in person the positive dBFS thing on my PC!

Some systems I installed suffered of severe clipping at the amplifier input (not output and with reference levels of course) and I remember being puzzled by those positive dbFS numbers. I suppose something went wrong with the conversion from the floating point signal to the DAC?

QSC as usual stuck their head in the sand.

[Steve Guttag]

John, I have no problem with floating point DSP having the wide range but where QSC is showing the +dBFS is on the output stage...at the D-A conversion. Furthermore, check out the bottom end (no signal)...-120dB...their bottom end on all input/output stages. If they were representing a floating point signal, at that point, should it not be a lot lower?

Screen Shot 2022-10-17 at 9.17.03 AM.png

We should be out of the floating point double-precision range by then...we're at the amplifier itself and feeding the speakers. What is above FFFFFFFF....?

One of the nice things about working in FP double-precision is that it is virtually impossible to "clip" the signal, internally, while one is manipulating it with various filters, EQs, expanders...whatever. One just has to get it back into the human range for the output stage. One really should only have to worry about getting their gain right at the inputs and outputs where the signal is touching the analog world.

[Marco Giustini]

Exactly, Steve. I was totally unaware of that. I was actually pretty baffled when I loaded the 32bit FP file this morning in my DAW and started adding stupid gain without seeing any clipping - to then see the signal being massively clip once converted to 24bit before the level was adjusted.

Good point about the location of the meters and their end scale.

[Steve Guttag]

QDS 9.6.0 has just be released. For those using 9.5.0, this is probably a good upgrade as it resolves bugs (that should have been fixed in a 9.5.1). It also adds the three new cores. The CORE 110f V2 (the version 2 omits the GPIO as well as the front panel display). These were the parts of the CORE110 that were keeping it from production so, the V2 allows a 110 offering. The CORE 610, a Dell based rackmount (1U) computer with Q-SYS installed replaces the CORE 510i (the CORE 510c having been discontinued along with "cinema"). The CORE 610 does not have card slots but, then again, QSC has canned the whole card based I/O (and the I/O frame). There are now QIO devices to replicate many of the I/O needs though just with a single Q-LAN, so no redundancy.

The last CORE is the vCORE (virtual CORE). This is licensed software (so there will be a reoccurring fee) to run on properly set up hardware. This allows one to implement a control environment if the sound part is already covered. I'm not sure it has much application in cinema but it might for some.

Note, QDS 9.6 may be a jumping off place for those that have older installations. The CORE 250i and CORE 500i are deprecated.



Version 9.6.0 includes these new features, updates, and resolved issues.

Platform

vCore Virtualized Processor
Core 110f v2
Core 610
Deprecated: Support for Core 250i, 500i, and I/O Frame 8S
NL, NM, and QIO Series device limit on Core 110 Series processors

Amplifiers

Disable PowerSave control

Loudspeakers

AD-C6T-HP and AD-C6T-HC
AD-DWL Series Video

NC-110 camera imaging improvements / new default values

Resolved Known Issues

• Platform: You can once again install Q-SYS Designer Software (QDS) on an ARM-based platform, such as an M1 Mac running Parallels, without an installation error. Note: Installation of QDS on ARM-based platforms is not officially supported.
• Platform: An issue that could cause audio noise from a Core 510i in IO Mode during redundant Core failover has been resolved.
• Amplifiers: An issue that could cause 4-channel CX-Qn amplifiers to become unresponsive during bootup with firmware v9.5 has been resolved.
• Loudspeakers: The AD-S28T sub, inadvertently removed in v9.5, has been added back to the Inventory.
• Video: CEC status and display power toggling now work as expected.
• Video: Signing out of Windows no longer results in an NV-32-H video format error and black video on a connected display.
• Video: An issue that could cause Mediacast streams to not resume after a camera reboot has been resolved.
• Video: NV-32-H peripheral logs (qsyslog) once again show complete information.
• Video: An issue affecting Mediacast stream performance with simultaneous NV-32-H USB Video Bridge 360p30 YUY2 format video has been resolved.
• Video: Source audio connected to Generic HDMI Display component Breakaway Channel pins is now audible regardless of whether HDCP is enabled or disabled.
• Video: An issue that could cause NV-32-H USB Audio Bridge distortion during heavy Encoder video utilization has been resolved.
• Video: An issue that could cause degraded I/O-USB Bridge performance with MJPEG video has been resolved.

[Ian Freer]

Barco Series 4 Digital Cinema Projector v1.0 plugin has been released

Summary:

• This Q-SYS-developed plugin allows control and monitoring of Series 4 Barco Digital Cinema Projectors, with automated projector functionality including macros, power, lamp, douser, and lens adjustment along with key status indicators, from a Q-SYS network touch screen controller, iOS UCI viewer, or web browser. It also allows for monitoring of the projector via Q-SYS Reflect Enterprise Manager.

Find it in the Asset Manager

[Steve Guttag]

Yes, it's true but I wouldn't be too excited about this one. I'm about to submit some comments to Q-SYS. I've only tested (briefly) with an SP4K-25C and, thus far, I've found the following:

• Notifications are missing (the ones that would let one, essentially, create the missing "taillight" for the errors, notices, warnings...etc.
• It does not indicate which Macro/Preset are selected
• While it provides for 100 macros and 100 presets, it provides no way to assign presets so one is still constrained to the silly 9 preset limit.
• When the projector goes into standby ("Eco Mode"), the status changes to the red disconnected/missing rather than showing Eco Mode.

This one needs a 1.1 version.

[Steve Guttag]

Doing a bit more research on why the S4 plugin is so poor...it seems to stem from the API itself lacking key features present in legacy projector communication. There is no way, in the current "REST" API to query the projector for the active Macro, for instance.

[Steve Guttag]

New software announcement. QDS 9.4.4 LTS. This one is part of the LTS series so there are not many new "features" and it is mostly about bug fixes. I'd say notable features for cinemas include the ability to disable the "Power Save" feature on amplifiers (CX-Q, DPA-Q). I have not found the power save feature to be an issue and it can cover up some of the noise issues of the Q amplifiers (they are hissy while at idle) but some will prefer to have things steady-state.



What's New in 9.4.4 (LTS)

Version 9.4.4 is a Long Term Support (LTS) release of Q-SYS Designer Software. It includes these updates and resolved issues.

Note: See the Downgrade Notices section for important information concerning downgrading to version 9.4.4 (LTS) from version 9.5.0 and later. Platform

NL, NM, and QIO Series device limit on Core 110 Series processors

To maximize processor efficiency, there is now a 32 peripheral limit for any combination of NL, NM, and QIO Series peripheral devices in a design using a Core 110 Series processor.

Note: NM Series Microphones (NM-T1) are not supported in Q-SYS 9.4.4 (LTS) but will be supported in an upcoming release. Amplifiers & Loudspeakers

Disable PowerSave control
AD-C6T-HP and AD-C6T-HC
AD-DWL Series Video

NC Series camera imaging improvements / new default values Resolved Known Issues

• Platform: An issue that could cause audio noise from a Core 510i in IO Mode during redundant Core failover has been resolved.
• Platform: The Collaboration Bundle license is no longer inadvertently shown as being required with certain designs using External USB Audio and AEC.
• Platform: An issue that could cause NL-SB42 and QIO audio peripherals to temporarily not pass audio at bootup has been resolved.
• Platform: Control Link connections between redundant Cores are now maintained if a design is pushed or a Core is rebooted while the backup Core is offline.
• Platform: An issue that could cause PA Router Core-to-Core paging connections to drop momentarily has been resolved.
• Platform: An issue that could cause an "internal fault, load failed" message when pushing a design to a Core 8 Flex immediately following a Core reset has been resolved.
• Loudspeakers: The AD-S28T sub has been added back to the Inventory.
• Control: Blocks no longer disappear after editing and saving them in Block Controller or upgrading a design.
• Control: MTR now works properly if a network proxy service is configured (and the proxy service has been configured to allow proxy bypass).
• Control: HID Keyboard component functionality has been restored when using Q-SYS Core processors and peripherals with Google Meet hardware.
• Control: Control names for QIO peripheral input and output components now properly start at 1 instead of 0, in parity with other component control names.
• Video: The default imaging settings for NC Series cameras have been optimized for conference room applications with average color temperature (4000K).
• Video: The Red and Blue gain controls for White Balance > Manual mode for PTZ-IP Series cameras are now functional.
• Management: It is once again possible to bulk-select files in Core Manager > Files. Use Shift + Click to select a range of files or Ctrl + Click to select multiple individual files.

Downgrade Notices


If you are downgrading to version 9.4.4 (LTS) from version 9.5.0 and later, we recommend performing these actions as part of the downgrade process:

• If your design includes NC-110 cameras, perform a Factory Reset following the downgrade. This resets the cameras' Max ePTZ Zoom limit back to 8x zoom. For Factory Reset instructions, see the NC Series User Manual at qsys.com.
• If your design includes an NV-32-H (Core Capable) processor in Core Mode, perform a Factory Reset following the downgrade. This enables the System Log to be usable by Q-SYS Support, if needed and authorized. For Factory Reset instructions, see the NV-32-H (Core Capable) User Manual at qsys.com.
• If your design includes USB Sound Card components, consider upgrading to version 9.6.0 before downgrading to version 9.4.4 (LTS). This preserves these components within the schematic without the need to re-add them from the Inventory.

[Elia Orselli]

I've found some spare time to try the IMS3000 audio processor in conjunction with a core 110f via AES67. After some work (and some hint from Steve...) I've found a working solution in a simple 7.1 environment.
At first I was getting some lost packet or duplicate packet errors in the Atmos receiver module in Qsys design. I believed that it was related to my switch settings, as I'm also testing a cheap TP-link TL-SG3210, but after some try I've seen that it was related to IMS settings, as they are slightly different from the QSC guide dated 2019:
immagine.pngimmagine.png
QSC guide indicated to enable "unique RTP destination UDP ports", but the checkbox does not exist in IMS3000 with 3.5.13 sw. Instead I had to enable "show/hide advanced mode" and change the RTP Destination UPD port manually. That is explained at the end of this Dolby training video: https://vimeo.com/435856534 (about 23.30 mins). As I'm in a 7.1 setup I did it only for 1-8 channels, but it has to be done for all channels needed in an atmos setup.

[Steve Guttag]

Another thing to consider is that the IMS3000 does not play with IGMP and will not respond to a query (or respond properly). So, if IGMP is enabled on your network, the IGMP will shut multicast off to the IMS3000...and that will mess up an AES67 stream from the IMS3000. While Dolby has improved the IMS3000s AES67 implementation (it is much better than it used to be but it still isn't quite there.

If you are using IGMP (and it is a good idea to avoid amps, touchscreens...etc. from going "missing"). most switches, TP-Link included, provide for forcing a port on for Multicast. This should be done on the port for the IMS3000 to the normal query/snooping don't slam it closed. In the case of TP-Link you declare the IMS3000 (and I would do the DMA amplifiers, if they are in use too) as "Static Router Ports." That will keep them open to multicast traffic.

[Steve Guttag]

Hear ye, hear ye,

QDS 9.7.0 has just been released.

Of note, potentially for cinema, a CORE 8 Flex can now be used in Peripheral Mode so, if you are in a dual-network installation and need some extra analog I/O, this could be an [expensive] option. The CORE NANO can also be used in peripheral mode but, unless you need a USB bridge, this is probably of no or little use for cinema.

Potential "gotchas" in include that how 9.7.0 handles streams changes a bit if you use dual-networks where up to twice as many streams could be consumed as with prior versions. This is not an absolute as I tested it on one of my CORE 510c systems and yes, the stream count nearly doubled (though WAY below maximum) but on a CORE 110c the stream count only went up by 2 and in both cases the DSP load and bandwidth consumption went down. However on yet another CORE 110f, where I was at about 84% DSP processing, moving to 9.7.0 sent it over 103% (non-functional). So, check your design before committing to 9.7.0 (and beyond).

My recommendation remains, for cinema in particular, stick with the LTS branch of QDS and the current version of that is 9.4.4. It will be the most stable as it does not get features as fast as the latest and greatest. Think of those people as YOUR beta testers! Let them iron the bugs out. 9.4.4 LTS has been stable for me on all installations I have it on running CORE 110, CORE 510, Dolby ATMOS and multi-screen with 5.1/7.1.

Here is the list of goodies for 9.7.0


Version 9.7.0 includes these new features, updates, and resolved issues.

Platform

Windows 11 Pro support
NM-T1 microphone support
Core 6000 CXR support
Core Nano and Core 8 Flex: Peripheral Mode support
Network Redundancy: Stream usage
Discontinued: PTZ-IP Series cameras

Audio

Media Stream Receiver: HTTPS support
Generic Speaker: New limiter parameters
Removed: "Maximum DSP" property

Video

NC Series: PTZ camera imaging improvements / new default values
Cameras: High resolution graphic support for UCIs, Mediacast to HDMI, and USB Video Bridging
NV Series: High resolution default graphics
NV Series: Support for 21:9 aspect ratio on HDMI inputs and outputs
TSC-G3 Series: Q-SYS AV Bridging
I/O-USB Bridge: Improved USB Video Bridging quality
Auto Privacy
Video Freeze


Control

Call Sync enhancements
Asset Manager: Lua Modules
Asset Manager: Plugin "Certified" and "Ready" badges
Lua Scripting: Crypto.Cypher support for AES_256
Lua Scripting: TcpSocketServer connection IP addresses


Management

Enterprise Manager: Presets enhancements
Enterprise Manager: Design Backup (Beta)
Core Manager: Cameras > Firmware Update page removed


Resolved Known Issues

• Audio: Grandmaster mismatch errors no longer occur when connecting QIO inputs.
• Audio: An issue that could cause "Packet Missing Streaming Errors" from QIO audio devices on networks with heavy traffic has been resolved.
• Platform: The Collaboration Bundle license is no longer inadvertently shown as being required with certain designs using External USB Audio and AEC.
• Platform: An issue that could cause PA Router Core-to-Core paging connections to drop momentarily has been resolved.
• Platform: An issue that could cause UCIs to not be discoverable in the Q-SYS Control iOS app when a Core's LAN A and LAN B interfaces are both connected to networks has been resolved in Q-SYS 9.7 with app version 3.6 and later.
• Platform: An issue that could cause Control Compute CPU usage to go to 100% with four or more Media Stream Receiver components in the design has been resolved.
• Loudspeakers: The unusable Status control pin for NL Series Loudspeaker Output components has been removed.
• Loudspeakers: Limiter settings have been corrected for some older Q-SYS loudspeaker models in 70V/100V mode.
• Video: CEC On/Off commands and the Display Power Status LED now function properly in conjunction with heavy Shift AV processing on NV-32-H (Core Capable) devices.
• Video: It is no longer possible to inadvertently control a PTZ-12x72 or PTZ-20x60 camera from two different designs.
• Video: The Status component for NV-32-H Encoders now properly indicates the Mediacast input Source Name.
• Video: An issue that could cause USB Video Bridges to show as disconnected and stop bridging video after a period of time has been resolved.
• Video: The default imaging settings for NC Series cameras have been optimized for conference room applications with average color temperature (4000K).
• Video: An issue that could cause USB bridged audio and video from an NV-32-H Encoder to a Mac to eventually stop has been resolved.
• Control: Blocks no longer disappear after editing and saving them in Block Controller or upgrading a design.
• Control: It is now possible to update the Script Access property for multiple components simultaneously without error.
• Control: MTR now works properly if a network proxy service is configured (and the proxy service has been configured to allow proxy bypass).
• Control: HID Keyboard component functionality has been restored when using Q-SYS Core processors and peripherals with Google Meet hardware.
• Control: Control names for QIO peripheral input and output components now properly start at 1 instead of 0, in parity with other component control names.
• Control: SNMP requests in scripts no longer cause increasing memory usage and eventual Core unresponsiveness.
• Control: TSC-G2 Series touch screens now properly auto-negotiate to 100 Mbps regardless of the connected network switch model, resulting in a stable connection.
• Control: Settings for port speed, data bits, and parity are now sent to QIO-S4 devices from Block Controller and Command Buttons components when a design starts running or after a device reboot.
• Control: Frequent camera switching in Mediacast Router no longer results in UCI Viewer or Q-SYS Control for MTR app instability when connected to a Core via LAN B or AUX LAN.
• Streaming I/O: In a Core-redundant configuration, System Link AV and audio streams now resume as expected after forcing a Primary Core back to Active status.
• Management: It is once again possible to bulk-select files in Core Manager > Files. Use Shift + Click to select a range of files or Ctrl + Click to select multiple individual files.
• Management: Core Manager now displays an error when trying to configure an invalid subnet mask (Netmask) for a LAN adapter.
• Management: The America/Cancun time zone selection in Core Manager no longer inadvertently adjusts for DST.