The LZX Patchable Video Standard

Category: Unknown · Tags: — · Posts: 10

#1 — creatorlars · 2019-01-08

[WORK IN PROGRESS, FEEDBACK IN THREAD WELCOMED]

LZX Industries Patchable Video Standard V1.0

The Patchable Video Standard proposes an electrical and interface specification for wide bandwidth analog computing instruments intended for creative and expressive applications. It is optimized for, but not limited to: (1) generating and processing analog RGB video graphics in SD resolutions, (2) affordable devices which are accessible to working artists and not just big studios, and (3) maximum patchability via a universal DC voltage range and high impedance connections.

Presented below are the guidelines we follow at LZX Industries when designing our products to communicate with each other. We openly invite and warmly welcome others interested in developing instruments which also speak the Patchable Video Standard.

Voltage Ranges

• An input voltage of +/-1V corresponds to +/-100% of a parameter’s range.
• Some inputs respond only to positive voltages. This is OK. For example, the red channel input on a video encoder module responds to 0% to 100% red, but negative values are ignored. However, when subtracting one signal from another using the Passage module, negative output voltages are retained.
• Signal generators output unipolar voltages, with 0V to +1V corresponding to the generator’s minimum and maximum values. For example, the red channel output of Visual Cortex input decoder section will output 0V when there is no red in the input image, and 1V if there is a maximum amount of red in the image. As another example, Prismatic Ray’s outputs oscillate between 0V and 1V across their frequency period.
• Inputs must be tolerant of voltages up to +/-12V.
• Output voltages must not exceed +/-12V.

Impedance & Bandwidth

• Inputs are terminated to 0V through a 100K ohm resistor. Inputs are usually immediately buffered with a video bandwidth op amp to avoid lowering input impedance.
• Outputs are usually driven through a 470-499 ohm resistor. Output drivers must be capable of driving up to four separate inputs without signal loss greater than 5% or a corresponding loss in bandwidth. Usually outputs are individually buffered with a video bandwidth op amp part.
• Signal paths must be capable of processing signal bandwidths from DC to 10MHz. This is to ensure SD video signals are passed without degradation in edge sharpness. In practice, this means op amp ICs specially designed for video bandwidths are used throughout the entire signal path. Some modules may be designed to intentionally alter the image quality. For example, a low pass filter designed to blur parts of the image, or an intentionally lo-fi approach to edge keying. This is OK so long as user expectations are set accordingly.

Electromechanical

• Inputs and outputs use 3.5mm TS phono jacks.
• Connections are made with 3.5mm TS phono cables less than two meters in length.
• Outputs may be split using passive mults, stacking cables and splitting hubs.

External Interfaces

• External connection to video cameras and displays, MIDI interfaces, ILDA laser projectors, etc. shall follow established electrical standards for the device in question. Conversion interfaces between Patchable Video and other signal formats should perform all processing and conditioning required to account for valid output under all input conditions. For example, the composite video outputs on Visual Cortex follow all expectations for consumer video devices: connection via a yellow RCA jack, 75 ohms output impedance, driven at double gain, sync signals and color subcarrier embedded, etc. With the Cyclops module, the output is via a DB-24 port used by International Laser Display Association for laser projector interfaces. In both of these cases, there are signal clippers, conditioners, and any other processors required to ensure valid output under all patchable video input signal conditions.

Video Synchronization

• Devices interacting with a video signal’s timing must be capable of synchronization to a master timing reference.
• Video synchronization methods follow broadcast video expectations for genlock using black burst or composite sync signals passed over RCA or BNC connectors. For example, on our current EuroRack modules there is often RCA sync in / passthru jacks located on the rear of the module for daisy chained sync connections.
• A synchronization input must be capable of accepting any valid video source as the master reference input. For example, I could synchronize a Prismatic Ray module to the Sync Out from the rear of Visual Cortex, or to the composite video signal from my Sony Camcorder.
• As an RGB graphics system by design, synchronization circuits need not worry about chroma subcarrier lock in composite video, which is handled as a function of output video encoder modules. Composite video inputs are typically time base corrected and decoded to RGB before being passed out as Patchable Video Standard color channels.

Note: The 14-Pin Video Sync Distribution Cable format, used by discontinued Visionary series EuroRack modules and Cadet/Castle DIY series modules, is deprecated for the purpose of this standards document.

#2 — transistorcat · 2019-01-08

Might want to add something explicit to the “external connections” part about the output module being responsible for valid non-harmful signal levels and conditions for all values in (i.e. the full ±12V range)

e.g. output modules not just being 12V tolerant, but also ensuring valid output levels for all inputs in the 12V range, and iirc. cyclops ensuring values not harmful to the laser.

Apart from that, nice work.

#3 — 337is · 2019-01-08

creatorlars wrote:

Some inputs respond only to positive voltages.

I’m curious to understand more about this. In your example of the red input on VC, what informed the decision to have it ignore the negative portion of a signal?

#4 — creatorlars · 2019-01-08

Re: Negative voltages.

Most voltage controlled parameters are unipolar by nature. In the case of red video for example, there’s nothing less red than black (and black is 0V.) Same with a VCA, there is no more off than off (and off is 0V.)

Maybe I need to rephrase it, because a negative voltage isn’t ignored if it’s summed with a positive offset knob (which is almost always the case in module design.) This is merely a reference to the CV range the destination parameter in question responds to.

For example, the Cadet II RGB encoder has just Red, Green and Blue inputs. Anything below 0V (black level) and above 1V (white level) is clipped off.

On the other hand, on most modules there is a CV pre-processing circuit, which involves a DC bias/offset (the main parameter knob control) which is summed with the CV input voltage (often processed by an “attenuverter” circuit, which may have negative voltages passing through.) The result is then a voltage that controls the value. This “pre processing” circuit is standard across most Expedition modules for any parameter (exceptions obviously being Bridge and Arch.)

In other words, parameters always respond to 0V to 1V range as min/max values. And generators always output 0V to 1V signals. But anything within +/-12V may result as these signals are mixed and processed throughout the patch (lots of headroom!)

Why 1V amplitude? Well the response of op-amps designed for high bandwidth signals is typically optimized for video signals, which also have a 1V amplitude. If we were to have 5V signals, we’d have one fifth the bandwidth given the same circuitry, and to compensate that means more expensive parts, higher current draw, and so on.

We also like 1V due to its ease of understanding in terms of algebraic functions and translation from floating point math. It is easy to think of 1V as 100% (full scale) value. Consider the following relationships for a 4-quadrant multiplier (bipolar VCA):

1.0V * 1.0V = 1.0V

0.5V * 0.5V = 0.25V

-0.25V * -1.0V = 0.25V

0.75 * -1.0V = -0.75V

Or in image compositing, where subtracting white from white equals black:

1.0V - 1.0V = 0V

#5 — creatorlars · 2019-01-08

Re: External connections.

The intention is that in the case of, for example, an interface module for video output, the inputs (patchable video format) would follow all the patchable video standard format rules, and the outputs would follow the existing standards for broadcast video – meaning that black/white level clipping, sync insertion, proper amplitudes and driving, etc – is implied. I could clarify that a bit better, and should probably add a section elaborating on how video signals and other interfaces are handled (i.e., with color channels and syncs as separate signals) and some reference documents for broadcast video specs, etc. I’m just trying to state the core guidelines as concisely as possible. A designer guide going into detail on handling things like video encoding and decoding would be a nice follow up.

Also, I’m not trying to define necessarily a spec for EuroRack modules here, or limit the specification to EuroRack power supplies. The +/-12V tolerance patchable video IOs exist so that you can cross-patch between audio instrumentation easily. For example, Vidiot is entirely a +/-5V power supply, but inputs are tolerant of +/-12V signals due to the presence of schottky clipping diodes on all inputs (post attenuator.)

There may be some cases for devices that exceed +/-12V on the external connections. For example if you designed a high current servo driver interface or something (in which case you should follow guidelines and protections for high current servo drivers.)

#6 — joem · 2019-01-09

I like this documentation being all in one place and clearly written like this. Nothing jumps out at me as glaringly wrong or omitted.

I think your Re: Negative voltages and Re: External connections responses would make excellent footnotes, especially since they’re a bit less rigidly formal than the spec. I love informal (or not-as-formal) footnotes to formal texts. Bring on more footnotes too!

#7 — transistorcat · 2019-01-09

Generally agreed.

The point was that as an engineer i read the “for all inputs, even non-valid (but inside abs-max)” as being strongly implicit, but wearing the user hat i don’t feel certain that every implementor would read it that way.

One of the most important ideas in modular to me is that with a proper setup nothing you can do to it is “dangerous”, especially to the system outside the instrument.

If a user feels absolutely sure that the worst thing that can happen is a non-interesting result, they will experiment in an entirely different way than if not.

I very much feel the need for brevity in the standard, but i feel like that specific point merits the extra half-sentence required

#8 — reverselandfill · 2019-01-09

creatorlars wrote:

A synchronization input must be capable of accepting any valid video source as the master reference input. > *> For example, I could synchronize a Prismatic Ray module to the Sync Out from the rear of Visual Cortex, or to the composite video signal from my Sony Camcorder.> *

To me this can be read in 2 ways:

you can also sync the prismatic ray to a camcorder

or

you can sync the VC to a camcorder

I don’t know if both statements are correct.

#9 — creatorlars · 2019-01-09

Re: External interfaces. OK, I get what you mean better now. I added something to that effect. This is of course one of the major tenets of the format itself; that any outgoing conversion interface conditions any input signals to be “safe” and within expectations of standards for correct, non-glitchy operation of the recipient device. (Goes back to design goal #3, maximum patchability for creative freedom.)

#10 — creatorlars · 2019-01-09

you can also sync the prismatic ray to a camcorder>>>> or>>>> you can sync the VC to a camcorder

Both statements are correct, but we’re talking about electrical standards/requirements here – not system usage. While any genlock input/sync input module should be able to sync to any valid NTSC/PAL video source, that does not imply this is the intended use case (to patch a camera directly to a module’s rear for synchronization purposes.)