Colour perception - a couple of notes

Colour perception is a complicated business; most people have tristimulus vision - that is we perceive colour (broadly speaking) in reds, greens and blues. Your brain does the clever stuff of filling in the intermediate colours - if you're seeing a colour between red and green (yellow) the mix of stimulation of the red and green cones in your retina give you the sensation of yellow. 

Although I do a lot of colourimetry (calibrating monitors, racking cameras etc) I was unaware that some people are tetramats (they have four sets of distinct cone cells) - they are more sensitive to the colours between red and green (and hence the opposite of men who have red/green colour blindness). Interestingly they are always women and invariably the mothers of men who are red/green colourblind. Dutch scientist HL de Vries discovered this in 1948 - here is an article in Discover magazine.

Radio 4 have an interesting series of programmes about colour perception - Russian speakers are trained by their language to discern more hues of blue than the rest of us and Homer was probably colourblind!

You could also watch the podcast Hugh & I did.

EDID isn't the only thing that graphics cards look at.

Avid Media Composer insists on re-starting when it sees a change of monitor. I always assumed that it was done via Windows detecting a new EDID profile but it turns out that if you unplug and re-plug the same monitor (i.e. EXACTLY the same EDID data - even s/n) then the same happens and so something deeper is at work.

Look at pin-16; Hot Plug Detect. Basically it is held low by the graphics card but pulled high when a monitor connects - this generates an interrupt in Windows which forces the card to do an EDID refresh - request a new profile and possible re-do the HDCP handshaking if needed. The interrupt is also seen by Avid and used to force a re-start of Media Composer.

Now then, Amulet (our favorite KVM-over-IP technology) spoofs all of this; it caches the EDID profile until a new client connects and when that happens it asserts the Hot Plug Detect pin as if a monitor had been connected - essentially spoofing what happens in the real world. There are various registry tweaks to for the graphics driver to ignore pin-16 but they work variably - change driver and suddenly pin-16 is being listened to again.

When I figure out a solution I'll update this post. Meantime the customer's dream of starting off a layback or capture and then handing it off to another client is still on hold....

Dual Internet gateways and Apple TimeCapsule

I had a splendid New Years at a friends house but as ever it became a tech support visit! He has BT provided aDSL which is a long way from the exchange and consequently very poor throughput (typ. 250kBits/sec down and maybe 100 up). He'd invested in a satellite Internet service which although fast (typ. 20mBits/sec down) has terrible latency; typ. 2000mS which makes loading webpages very slow BUT means you can stream media (even the HD samples files from GoPro).

As an aside the satellite modem is just that; it has no NAT functionality or firewall; it exposes a proper, routable IP address (in the 37.x.y.z range).

So - his question to me was could he combine the two; use the aDSL for most things but force streamed media to come via the satellite modem? He has an AppleTV under his tele and so I thought it should be easily do'able. I think I've figured out a solution using only what he has, but I came across a few interesting things on the way. There "..oh, can you have a quick look" always turn into an afternoon's fun!

So, recap of the parts on the network;

• Current model BT aDSL2 HomeHub
• Current model Apple TimeCapsule (the 3TByte one)
• Netgear GigE switch
• Sitecom DVB-S satellite modem
• iMac, various wireless devices, AppleTV, couple of PCs, network printer

I was worried about exposing the satelite's IP address to the network; that has to go behind a firewall or NAT router and the TimeCapsule has NAT built in - great; I imagined I'd just have a network with two internet gateways and for the AppleTV I'd point it at the TimeCapsule and the iMac (which is used to download films, TV etc) have two network configurations which you swap over manually; it's very easy in OS-X.

My first problem was when you turn on NAT in TimeCapsule it assumes it is the only DHCP server on your network and it tends to answer DHCP requests faster than the BT aDSL router; bit of a problem as the BT was meant to be the default gateway. The answer is brutally simple, actually. Just leave DHCP enabled on the Airport, but restrict the available IP range to just a single IP. Then use a DHCP Reservation to bind that IP to a MAC address that doesn't actually exist. The Airport will silently refuse to respond to DHCP requests since it thinks that it is out of addresses that can be assigned.

 

 The next problem is that the BT router's ARP table is tied to it's DHCP table; if the router gets requests from an IP address that it didn't serve then it refuses to route them to the outside world. You can switch the iMac's networking from DHCP assigned to manual and enter the same network details the router provided and no traffic will flow. 

I guess both of these issues are because the gadgets are aimed at non-technical people who (if they ever get into the remote interface) don't want to be worried by DHCP, NAT etc.

So, the final solution was the allow the TimeCapsule to share the satellite modem's connection via wireless and the BT to share the aDSL via wireless and wired. The iMac and the AppleTV choose which connection they need by the wireless network they attach to.

Integrating a Blackmagic Universal VideoHub pt. 2

As mentioned in a previous post I'm integrating one of the big 288x288 VideoHub routers - 572 3G High Def video connections (mix of coax and single-mode optical) and 288 RS422 remote ports. The thing is not very deep (maybe 100mm) and only 18u high and therein lies the problem; how do you cable it neatly and in a state where re-configuring/maintaining it is possible? Each of the 96 interface cards has either eight BNCs or four duplex-LC optical connectors and there is a proprietary 4-way RS422 port in the centre of each card. 

In the left-hand image you can see we've taken the optical feeds up the right-hand side of the bay so they can come over the top of the patch panels where all of the facilities single-mode tielines (all run in loose-tube cable!) terminate - top of the right-hand picture. Many of the rooms are nearly 100m away from the CAR and so for reliable 3G performance SMPTE 297M is a must. Those pre-made single-mode patch cords are protected in Copex.

All the coax feeds got down the bay to CTPs in bays either side for all the incoming/outgoing jackfields. Nylon sock allows us to tame the coax and keep it neat.

The RS422 was the real challenge. As mentioned before they have a 1m pre-made cable that breaks out each card to 4 x 9-pin(D) connectors which we chopped off! These we ran into little 0.5u cat6 patch panels mounted on the intermediate rackstrip. In the LH picture they are the bundles running up the middle and in the right hand pic you can see the incoming feeds from the suites/VTRs/Avids etc.

This thing has been running reliably for a month now and I am staggered that BlackMagic can build and ship a 288x288 3G/RS422 matrix with fibre for sub £100k. You could make a real pigs ear of cabling one of these very easily which would limit it's utility. I think we've got it about as neat/maintainable as is possible.

3G SDi parts becoming commodity?

My pals at Lindy (who I buy lots of computer connectivity parts from) sent me a bunch of engineering samples of 3G-capable SDi parts; DAs, fibre transceivers etc. from one of their OEM manufacturers in the Far East. For very modestly priced pieces they were excellent. The eye-pattern (above) is the output of their re-clocking DA. Whilst talking to them about the various measurements I made (which you'd only understand if you're a broadcast engineer) I recalled Tektronix's excellent SDi physical layer webinar and pointed the guys at it. 

 http://www.tek.com/webinar/hd-and-3g-sdi-physical-layer-webinar

TCP & Networking, part 2; the protocols

I continue my conversation with Hugh going over some of the lower-level protocols that are used in IP networks. Find it on iTunes, vanilla RSS, YouTube or the show notes website.

The Engineer's Bench podcast - TCP & Networking 101

Gone are the days when every cable carried a synchronous video stream. Contemporary engineering staff have to be aware of packetized networks and how they impact the modern facility. This part 1 (of a two-parter) covers the fundamentals of the protocols and practises that drive all internet-derived networks. Find it on iTunes, vanilla RSS, YouTube or the show notes website.

Multicast addresses in IP

I thought I knew TCP & UDP/IP but I was reminded this week about the 224.0.0.0/4 multicast subnet. If you're ever in a position where you need to identify a device's IP address (even on a different subnet, but the same LAN segment) you can PING 224.0.0.1 and everything on the segment will respond to the PING (firewall settings permitting).

 

So, if I set my machine's IP address to 192.168.1.220 on a 10.100.100.x network and then PING the multicast address;

You can see that all the machines on the 10.100.100.0/8 network respond.

This comes in very useful with Amulet DXiP cards which you configure over a web interface. Our demo kit came back from a customer who had forgotten what they had hard-set the cards' IP addresses to and this technique was a life-saver.

Thanks for reminding me of this Graham and Don Poves!

Wikipedia article on multicast IP

Some more training notes

Video Measurement - Principals (4 days)

Course Aim The aim of the course is to provide grounding in video fundamentals, compression and picture quality analysis for staff who work with video processing, detection and vision systems.

Course Structure 

• Day 1 - TV Fundamentals; Scanning and Sampling, Colour Systems, Analogue Composite Coding - notes
• Day 2 - TV Fundamentals; Digital Component Coding, Conversion of Film to Television, High Definition - notes
• Day 3 - Compression; DCT principles, Intra-Frame vs Inter-Frame Encoding, blocks and macroblocks etc. Video Tape Recording; Magnetic recording principles, rotary recording, Simple VT maintenance - morning notes, afternoon notes
• Day 4 - Television Measurements; Analogue, Digital, Picture Quality Analysis; Analogue picture impairments, digital picture impairments, compression, TV Displays; CRT, LCD and Plasma displays. Problems with LCD and Plasma, Projection systems - morning notes, afternoon notes

Cladding vs Core Alignment fusion splicers

For seven years we've been using Tritec Fase II fusion splicers; we have two kits composing the splicers themselves (camera or microscope inspection), clevers and finishing ovens for the splice-protectors. If you're in the business of bespoke fibre then fusion splicing really is the only way to go as you get superior performance in termination and you can base infrastructure entirely on loose-tube multi-core cable (tight-buffered cable is really only suitable for use within cabinets or at most between cabinets - it isn't man enough for running in voids and risers).

Anyway, the Tritecs have been solid workhorses clocking up many thousands of terminations between them and we've been entirely happy. In 2004 and then again in 2007 when we upgraded them they were the best sub £10k machines.

The only thing you could criticise them for is that they are "cladding alignment splicers" - they rely on the diameter of the cladding being correct (125 microns, fact fans!) and the core being correctly positioned within the fibre. The guys at Tritec tell me this is a non-problem as contemporary fibre optic stock is always spot on - ten years ago, not so much, but for now it's a problem that's solved and it is fair to say we have seen very little badly made fibre since we started; and that's many hundreds of kilometers of fibre cable!

 

The other side of the coin is the newer style of fibre machines called "core alignment splicers" where two cameras set at 90 degrees examine the cleaved ends of the fibres  and the software dynamically aligned the ends and dried the splicing arc. The machine is then able to illuminate the join and make an estimation of the loss across the join. Fujikoura were the pioneers in this field in the mid-noughties with machines like the FSM-series costing more than £30k - clearly you'd have to do a lot of fibre work to pay off one of those! Guys who use both types a lot reckon that core-alignment machines allow you to work about twice as fast as the machine produces a much greater consistency of good splices and doesn't rely on the operator to maintain precise splice measurements. Ask any wireman about the 'fibre-blindness' that sets in around 15:00 each day; you just need to walk away and forget about squinting down a microscope for an hour or so; these are very small measurements after all!

So, we've just splashed out on an INNO IFS-10 which gets really good crits when compared to the current FSM-60 machine from Fujikura. We're going to take it on a big job we're just starting and so I'll get a good feel for it.

Here is an excellent "torture test" video of the machine in action.

Amulet - they fixed OS-X's Temporal Dithering issue

As mentioned in a previous post temporal dithering on OS-X has proved troublesome with Amulet (and in fact it bedevils all KVM-over-IP systems). The guys at Amulet have written a Kernel Extension that stops the card turning on temporal dithering. James, their engineer, explained to me that it's a different technique between nVidia and Radeon, but they've got it licked;

 

 

kexstat is a utility to show which Kernel Extensions are loaded. Here I've grep'ed the output to exclude all the Apple ones. You can see the Amulet one at 0x2000

The proof of the pudding is that now all the MacPro clients on this particular Amulet system look splendid; even full-screen replay of 1080P material.

New podcast - RS232, fifty years and still going!

Hugh and I go into the details of RS232C and how it is still used in broadcast engineering for configuration and test. Find it on iTunes, vanilla RSS, YouTube or the show notes website.

Getting cheap/unsupported laser printers working on OS-X

Got a Dell, Samsung or Xerox printer where the manufacturer has dropped support AND it was too cheap to have a full Post Script implementation? Whereas Windows seems to support a very broad range of budget printers natively OS-X is more picky when it comes to PCL-models.

So - this saved us at the workshop; http://guigo.us/mac/splix/

Miranda, Grass Valley and good-old GPI tallies

I never realised that lots of manufacturers use ACOS protocol to send tally and UMD-names between vision mixers, multi-viewers and routers. I assumed that big switchers still used GPI closures for tallies but when I got to a customer's site to make a GVG Kayak mixer talk to a Miranda Kaleido multi-viewer I couldn't find the RS422 port on the back of the Miranda; turns out it's an option!

So - back to the good old dry-GPI (i.e. relay closure) tallies.








After a couple of false starts here is the cable used to connect the GVG to the multiviewer.

The next thing is to associate the appropriate GPI ins with the video inputs on the multi-viewer; this is X-Edit, Miranda's network config tool. Once you've dragged the GPIs to the source lines you have to go into the layout tool and associate the UMD red-light-up (it the RHS of the UMD definition) with the GPI event - you can set as many or as few of the virtual displays to activate.

Windows HD .WTV files; what can you do with them?!

I record off-air TV using Windows Media Centre. Those channels that broadcast in HD do so over the DVB-T2 standard, and in the UK at the moment that means H.264 in the transport stream. Media Centre wraps them as .WTV file which are almost unusable (other than playing back). You can't even re-wrap them to the earlier .dvr-ms format that other PVRs use (and is more convertible)

But, it turns out that the new version of Handbrake can handle them and convert them to vanilla H.264 with lots of pre-sets for iPad and other platform. Then you can use Quicktime to trim them (remove ad-breaks etc).

Some interesting domestic electronic fixes

Fridge/Freezer earthing conundrum

I've had my Miele fridge for maybe six years and it's one of those frost-free models; I assume an extractor fan maintains negative pressure inside so that moist air that has entered when you open it gets drawn out. About three years ago (after I'd re-arranged some of the mains in the kitchen) it started icing up every few months like an old-style freezer. Of course it was outside it's warrantee so we lived with it. Around eighteen months ago whilst cleaning behind it I noticed the mains cord had been tugged and the earth was disconnected. Of course I re-ended the cable but since then the freezer has stopped icing up.
I've searched online for a schematic for this model to see if I can figure out what is going on. I'm reluctant to disconnect the earth to prove it, but I'm left wondering (and no other engineer I've mentioned it to can give me an answer) why the anti-frost mechanism didn't work without the safety earth? I stuck the freezer on my trusty Martindale 2100 PAT tester and it was fine; no excessive residual earth current and all isolation good; even at 8A.

When is a micro-switch not a switch?

My Valiant combination boiler has the usual configuration of a butterfly valve that routes hot water to either the taps or the heating pump. There is a micro-switch on the valve such that when you open the hot water tap the boiler is forced on (it's unlikely it was on heating the radiators exactly when you needed to wash your hands). The boiler is fifteen years old and I've replaced that micro-switch two or three times before. Earlier this year it seemed to have gone again yet removing the connections and testing for continuity when the valve opened showed the switch (seemed!) to be working. In the end I admitted defeat and got a heating guy in. He concluded the same as me and he 'phone the Valiant tech support line.

"Replace the micro-switch"

"No, no - it's not the switch, it buzzes out correctly when the valve closes"

"Never mind that, replace the switch"

Because it's a common model of boiler and the micro-switch goes often he had the kit in his van. Sure enough, he replaced the switch and the boiler starting working correctly. I was now beginning to doubt a lifetime of electronics knowledge! I got him the call the helpline back and the chap at the other end explained that it's not just a DC voltage that gets switched but the i2c data link! Yes, I was amazed; I can only assume it's part of a wider control system (maybe common across much more complicated boilers); but if there is any electrical noise or impedance on the switch data doesn't get back to the control board and the boiler CPU doesn't fire up the gas.

XBox 360 and failed DVD drive.

What would imagine would be the part most likely to fail in any kind of consumer device? The mechanical part that has a 50p laser diode AKA the DVD drive. Given that it has two connection - SATA and power it should be a user-serviceable part and given you can buy them from eBay and spare parts suppliers home tinkerers like me should be able to fix their XBoxes. But no, the XBox OS is keyed to the firmware serial number of the drive that is installed at the factory and if it sees another drive it not only refuses to read the disk, but it informs the XBox Live! (why the exclamation?) mothership and your account is banned for being a game-stealing pirate.
In the end you have to get the same model of DVD drive (there have been four revisions over the lifetime of the '360) and swap the board between drives so that your new mechanics have an old firmware. It worked for me. 

How to do RS422 properly on the big Blackmagic Universal Videohub?

A job we're working on at the moment needs a big 288x288 HD/SDi and RS422 router. Budget means that the Black Magic is the only one that can be considered and it's hard to argue with sub £100k - Probel, Quartz-Everts etc would all cost four times that amount and so we're trying to figure out how to cable this one nicely.

The issue is going to be with the RS422; video will dress nicely on the cable tray down each side of the cabinet but the remotes are presented on a proprietry connector in the centre of each card. The only way of connecting to this is via a pre-made breakout cable that in only a metre long. So, after a lot of thinking and measuring we've come up with a solution.

 So far I've not been able to locate this connector - anyone have any thoughts?!

 Chopped breakout cable to see what the colour code of the cores is.

Nice 1/2u high cat7 jackfield which we'll use (need a dozen of them) to run down the intermediate rackstrip at the rear of the matrix to allow proper RS422 breakout.

The colour code we've settled on for this router when it's being cabled on cat6.

Avocent HMX-series KVM-over-IP, a few notes

Although I'm a fan of Amulet Hotkey for sending 2 x DVI, USB & audio i/o over ethernet networks (and even slow connections; works over a domestic internet connection) we did fit an Avocent HMX system on a recent job. 

Unlike the Amulet where you can do remote admin of both the transmitters and the receivers (referred to as Zero Clients) over a web interface with Avocent you have to use a serial connection (see my post about RS232 for making a null-modem cable!) with a slightly unusual comms setup.

The clever thing is that if you want to connect to a transmitter (the gadget that sits behind the server or workstation) you have to do it via a receiver that is logged onto that transmitter over the KVM-network. I suppose this allows you to administrate a very distant workstation's sender from where you are.

So - RS232 port (male, therefor a DTE, therefor a null-modem cable required) and you can see what you might not be getting any video from the remote computer.

What I discovered is that unless the EDID (or DDC if you're doing it over SVGA) profile matches the resolution of the sending machine you get no video! What is the point of that - the only reason to pass an EDID or DDC profile is to allow the graphics card to tell the system what resolution is being expect.

Anyhow - fixed that by logging onto the machine locally.

By the way - is there a decent PuTTY-like app for OS-X. Good job I keep Windows on my MacBook Pro for situations like this.

Fifteen years of HD displays in one morning!

I had to visit a new client to calibrate four different display devices and it struck me that they represented the gamut(!) of HD displays since the mid-90s. They didn't have an OLED display for me, but a Sony BVM-D24, a VuTrix Pro-24, an AutoCue G-series and a JVC D-ILA 2k projector.

The CIE 1931 chromaticity chart shows the extents of human vision with the RGB primaries mapped onto a pair of X-Y primaries that represent the colour content only and says nothing about the illumination of a colour. In the centre of the triangle is an "equal mix" of red, green and blue which is the very definition of white light. The various standards for the white-point used in different territories and applications is shown as a curved line. In TV pretty much the whole world now regards 6500k (actually, 6504k - had to be updated in the seventies when physicists realised they'd got Planck's constant wrong!). When I test myself I can only just discern the difference between 6500 and 6600k colour temperature.

Anyhow, metamerism aside (and that's another blog post!) you can't use the same probe to calibrate different display technologies. My venerable old CRT probe is no good for LCDs and vice versa (and that's before you consider plasma and OLED displays). That's why the £200 Huey-type probes you can buy on Tottenham Court Road are worse than useless.

• Sony BVM-D24 - Old faithful, the HD display we've all been using forever. I'm starting to notice that the '24s I see now are in the autumn of their tube's lives. They are still the gold-standard for colour and unless the EHT is wrong or the tube is really knackered it's a nice and easy to job to get a D24 correct in the blacks and the whites and generally they track splendidly. As ever they look best when the white point is set below 100Cd/m2
• VuTrix Pro-24 - These could be could be considered first generation broadcast LCDs as they are rear-illuminated with fluorescent strips. This particular one had to be really wound out from it's factory defaults to get it even near standard. They also have no sensible way of setting the colour in dark areas of the picture ("bias" on a Sony monitor). These have poor off-axis black performance as well.
• AutoCue G-series - This was a new monitor, the G-series are their best ones with LED backlit LCD panel and a really nice look. Easy to get it looking correct from a colour point of view and it's factory preset was pretty darn close to illuminant-D. They must have some sort of fancy dichroic glass on the front because the off-axis performance is pretty good.
• JVC D-ILA projector - because I don't have a spectral-radiometer (a wide band colour analyser), only photometers (colour probes that rely on having the same metameristic failure as the the device being analysed) I have to eye-match projectors to a know calibrated LCD or CRT. This is the bit of the job I like the best. Since I spent a decent amount of the nineties racking studios cameras I'm used to colour matching. I don't have a good colour memory (I'm always surprised how milky D6500 white is compared to what I've remembered).

Once I've got the probe happy with the light coming from the front of each display I do an eye-match as a final tweak. BBC test-card F is perfect for this as it has lots of grey-scale and real pictures. It's always nice to give Carole an outing; she's the person who's been on TV the most in the whole world, ever!

You can see here the monitor on the floor is a bit redder in the whites than the projector; since I'd got the monitor correct I eye-matched the projector as best I could; but it's always a struggle because different display technologies just look different because they make colour in different way.

What I spotted at IBC 2012

I had a couple of days in Amsterdam at IBC (the TV industry's European trade show); I talked to lots of people and had a splendid night out with Bryant Broadcast (my main cable supplier) at the Braziliaans grill restaurant.

Anyway; the only thing that really grabbed my attention was the updated NewTek Tricaster - the model 8000. The really clever additions are;

• Automation; mixer tasks can be grouped into macros
• Control of outboard routers (essentially expanding the number of HD/SDi inputs)
• Eight re-entrant MEs
• A motion tracker (yes, in a studio vision switcher!)
• Hot-spots that the talent can activate by their actions - think the weatherman being able to reach out and "pull" in his next map.
• Up to four status monitors
• A bunch of social media exports ("send that clip to Facebook now!)

Watch the video (Kiki features, as ever!)