The upgraded AV system is now installed in LR4- news item to follow.
We were faced with a number of challenges and I thought I’d talk about some lessons that came out of the exercise.
One of them is that EDID (Extended Display Identification Data) management is not as mysterious as one might think…
PC: Hello screen I want to display something. Can you tell me what resolutions and timings you support?
Screen: Sure, here’s a list.
PC: Thanks. I see that the highest resolution we can both handle is 1900 x 1200 @ 60Hz so I’ll send that.
Screen: Whatever. *displays video signal*
Easy right? Well not so easy if you are routing multiple sources of varying/unknown display capabilities through a 4×4 IO matrix onto 2 screens and 2 projectors each of which you want to display at a non-native 1280 x 1024. Get it wrong and you end up with stretched or squashed images, sluggish switching, 800 x 600 defaults or, in a worst case, no image at all.
Thankfully we can trick the system at various points into doing what we want it to do.
Firstly, the Extron matrix allows us to place a ‘phoney’ EDID file at the input to the matrix which means the information request never makes it to the projector/screens. We can lock down each input to act as a 1280 x 1024 display. This means that the matrix output and projector make just one EDID handshake to this resolution on boot and never move away from it, ensuring that the projectors aren’t wasting time re-scaling (flickery blue screen anyone?) each time an input is switched. We can then manually fix the lectern PC resolution to 1280 x 1024 to comply and enjoy quick and responsive input switching.
Secondly, in order to manage the various unknown resolutions of laptops that may be connected we place a small switcher/scaler between the laptop and the matrix. Again, we can place a phoney EDID file at the input that acts as the display’s data to force a first time laptop to use an output of 1280 x 1024. We also tell the unit to output at 1280 x 1024 so that any manually set resolutions scale correctly and never stretch or squash, instead fitting within 1280 x 1024 frame at their correct aspect ratio. Technically, we should place a higher quality scaler on each matrix output in the rack set to the projector native of 1900 x 1200 instead of relying on the projectors cheap built in scaler but the difference would be negligible.
It’s worth noting that you can still manually override the automatic EDID triggered resolution settings in Windows- most displays have a range of inputs they can handle and the Extron scaler can basically handle anything you throw at it. Interestingly, Windows registry remembers the display and the setting you used for it so if you manually set to another resolution then come back another time, the laptop will use the last resolution set on that particular display. I am yet to explore whether that can be changed- another time maybe.
The control system is another story…
This is a really stupid question (sorry!): why do you want it to be locked to 1280×1024? Is it that you want to avoid resolution changes (delay/flicker) and that’s the one most likely to support everything you throw at it? Or is there some other constraint?
It’s readability of text that defines everything.
When we design new spaces, the size of the projection screen is determined in terms of the type of content to be displayed and the distance (d) to the farthest viewer. Standard practice dictates that the height of the screen be at least d/8 for movie content (entertainment), d/6 for text led presentations (learning) and d/4 for graphical scrutiny (group working) because of the impact this has on viewing capability in a broader sense. In lecture theatres, we aim for at least d/6 although it’s not always possible within architectural constraints and existing installations.
We then have capability to change resolution within this size for further refinement. This is a balance between readable text and screen real estate. At 1024 x 768 the text is a good size but software real estate (particularly multi-window) tends to suffer. A higher resolution such as 1600 x 1200 then presents the opposite problem, that text is somewhat too small to decipher for the audience in a larger theatre. So 1280 x 1024 gives a common balance between the two.
Switching to widescreen would give us more real estate but we would require more architectural room to increase the physical width of the screen! It is far more efficient to meet the viewing distance/screen height needs with square formats than it is with widescreen- the same text readability in LR4 requires a screen width of 2.6m at 4:3 compared to 3.5m at 16:9.
In any case, there is a requirement for the continuing use of acetate OHPs which most certainly would be useless on a 16:9 screen, so 4:3 it is 🙂
So once the best resolution is decided upon, the equipment is configured to lock to it as per my original post. In a scenario where 1080p was the ideal display format, I could configure the equipment to lock to that and the behaviour would be the same. The equipment will support everything we throw at it because of the way it is configured, regardless of the output resolution.