Hotshoe accessory connector pinout spec (Niche question)

[Mattyw]

Advertisement (gone after registration)

I am inquiring if there has been any detailed research or analysis conducted on the pinouts of the proprietary hot-shoe connector used in Leica M10/M11 cameras, specifically regarding its data output. Has there been any reverse engineering undertaken to decode the signals, formats, and general pinout configuration? Is the data output comparable to HDMI or Component video, and what other data types are transmitted via this connector? Are there any available schematics that detail this?

Additionally, I am exploring the possibility of developing an adapter for studio or specific use use cases that allows for external live monitoring (via HDMI/USB-C) without relying on the Leica app, which has been noted for its latency, compression, and occasional instability. Would there be interest in within the community? Correct me if this already exists .  I know this isn't the intended use for the M cameras,  and should just buy another more aligned to the situation but i think it would be a nice ad-on for people with poor vision and client monitoring when tethering isn't possible. 

[Advertisement]

The Leica Q3 is available for pre-order at B&H Photo Video and Adorama

[Photoworks]

I don't think you can drive the external screen from the M cameras, data rates and power would not be enough. 

there is also a pause between shots that is quite long before Video Out is reactivated.
I would consider another camera.

USB-C can deliver Live view in Capture One.

[Mattyw]

@Photoworks thanks,  not interested in the power rails,  i wouldn't expect an external screen to power much more than the visoflex, but more interested in the data that is driving the Visoflex, and the feasibility of of less tethered external views.  I didn't realise Capture One had a "live view".  Im mostly interesting in the underlying proprietary pinouts more from a exploration POV, and understand whats possible.  

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

 

[Vec]

The easiest (and I mean this really loosely) is probably try to open up the Visoflex to try to reverse engineer it from that side. My guess is there are a few channels of LVDS that are typically used to drive LCDs using something like MIPI as the higher-level signaling.

Most of these tend to be either proprietary or hidden behind expensive NDAs. (Or, I suppose, both)

A bonus from this is you'll get a Visoflex connector that you can use as that's also a proprietary Leica connector I believe.

Presumably, there is dedicated hardware on the Visoflex side to decode all this. Reverse engineering would entail some high-speed FPGAs while doing the reverse engineering (and likely productionalizing as this is presumably low volume. Anything with LVDS signaling will require some high-end test equipment as well -- likely well into the GHz range. Even dealing with channel integrity on circuit boards can be a bit of a challenge at these speeds.

A quick back-of-the-envelope calculation gives me a minimum raw bitrate of around 1.8Gbps (3.7 megapixel/frame * 8 bits/pixel * 60 frames/second), but of course, there will be plenty of other sources of overhead that would likely get that well over 2Gbps. This is also assuming the lowest bit depths being pushed over the interface, as well as only sending one color channel per pixel.

Is it impossible: No. Nearly impossible: Probably.

It would likely be cheaper to fly to Wetzlar, hang out in bars near Leica HQ, get someone who knows something drunk, and just social engineer something closer to the answer.

[Photoworks]

since you are going to take everything apart, can you come up with a sync cord socket that uses the center pin to add a PocketWizard in the studio?

Someone made on in the past for the M10, but I can use it for the M11

[Mattyw]

@Vec Thanks for detailed and excelling answer,  and yeah - it does seem like quite a challenge.  I will tear down both versions, and see what can be unpicked and monitored - would be a little overkill to create dedicated protocols but i defiantly wouldn't put it past Leica to put the effort into the protection of the design with the risk of third-party products breaking things and their hardware .   

Failing that,  like you say its probably easier to find alternative methods  

https://www.yukosteel.com/2024/05/disassembly-leica-visoflex-typ-020.html

[Photoworks]

Advertisement (gone after registration)

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

 

[Mattyw]

14 minutes ago, Photoworks said:

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

 

Interesting - flash sync...  

Edited May 27, 2024 by Mattyw

[Vec]

2 hours ago, Mattyw said:

I will tear down both versions, and see what can be unpicked and monitored - would be a little overkill to create dedicated protocols but i defiantly wouldn't put it past Leica to put the effort into the protection of the design with the risk of third-party products breaking things and their hardware .

I have both the V1 and V2 Visoflexes (though the V1 is with my friend across the country). The capabilities of my scope top out at ~ 250-500MHz, the front-end of my scope is unlocked to 500MHz, but I don't have proper differential probes to go much beyond than 250MHz. Most of this stuff goes much faster than that so I'm more than a bit hesitant to crack into my V2 to try to offer any assistance. 

Maybe the easiest (still, a relative term) approach would be to get a PCB fabbed that has the same overall specs and maybe somehow make an interposer board that has all of the connections exposed so you can probe a live connected flash. The challenge would be to get the connector the real Visoflex would plug into. I suppose another challenge, assuming we're talking differential lines, would be to figure out which pairs would need to be run in parallel to get the correct impedance for things to function correctly. A teardown of the camera would make that a lot easier... but that's another hard challenge in itself.

Extending the flash sync, like that posted above, well... that would be a lot easier!

My assumption is that there's some debug mode where the image data would be exposed on the USB-C interface. It would make the manufacturing, alignment, and calibration a whole lot easier. But that's a whole other problem.

[Vec]

I was doing some digging and it looks like the likely display panel for the Visoflex is a Sony ECX337A. They don't publish the datasheet for it, but they do state that it's LVDS and sub-LVDS signalling going into that module. I don't have a clue if the camera is driving that directly, or if the Visoflex has a frame buffer inside it -- either is plausible depending on tradeoffs.

One thing to keep in mind is that "dots" is talking about sub-pixels. So assuming that's the panel used by Leica, it's only 1280x900xRGB resolution, which seems pretty low all things considered. It may be a show-stopping limitation regardless of anything else.

Best info I can find: https://www.framos.com/wp-content/uploads/media/pdf/00/56/c0/Flyer_FRAMOS-M-OLED_EN_WEBOFlTu3yn1JOoV.pdf

Edited May 27, 2024 by Vec

[Mattyw]

Thanks every for the feedback,  regarding the resolution, i was hoping that there was some higher resolution that makes forward compatible to newer EVF's.  Only reason for this is the earlier versions for the device were lower resolution - so hoping there might be some higher-datarate that is being downsampled to 1280x900,  I do agree,  if its not higher than that natively its a bit of no-goer - it'll just look rough even if we manage to reverse engineer it.

Regarding your suggestion PCB fab, this is exactly what i was thinking, something that sits between the two devices so we can probe and check everything - that's prob the only non destructive way we can manage this.  @Vec you sound like you know your stuff,  a lot more than me anyway - been speaking with a few engineers who are experience with video reverse engineering, but obviously non of em have a M10/M11 lying about so it's going to be challenging.   Also thinking about the GPS in the earliermodel and proximity sensor (that turns the main screen off when you're close to it) this suggests there is bi-directional stuff going on which complicates things further.

Thanks for the links @Vec