It looks like XDCam, DNxHD or ProRes will be acceptable in many venues.
Just exported a static test chart:
ProRes: 203,393 KB
DNxHD: 65,712 KB
XDCam 1,257 KB
All are 4:2:2
ProRes: 158 Mb/s
DNxHD : 50 Mb/s
XDCam 50 Mb/s
I’m assuming your test exports were only for basic file size comparisons.
For more detailed tests, the clips should contain bars and tones, but also full
screen single colors like red, green, blue.
Alternating between them at 2 sec intervals is good.
Although saturated colors (color bars and single colors) are important, it’s often difficult to see any subtle color shifts using them, thus there should also be video with motion, sports is always good.
Fast pan shots are good for testing interlacing and shutter.
Fine detail patterns like a cross hatch is also nice to have.
Although many potential problems and degradation cannot be picked up by eye alone, it’s certainly a good start to double check your exports.
Don’t forget to include audio, tone followed by some full range music and back to tone.
Do a check for any clipping and also ensure that the audio levels are within spec.
Keep in mind that LUFS are the preferred method, don’t think any broadcasters use RMS anymore.
What is really nice to have, but out of reach of most casual users, is a way to play out
your test exports via SDI and feed that into a good calibrated reference monitor and waveform and vector analyzer.
TVlogic, Plura (amongst others) make reasonably priced reference monitors with
built-in waveform and vector display as well.
Reconditioned ones can also be picked up second hand.
Just make sure that they can be calibrated and what the cost will be as I suggest this
every six months at a minimum.
You may want to double check that, it looks odd for 1080 interlaced.
There is a DNxHD at 36Mb/s (1920X1080 @ 25p)
43Mb/s (1280X720 @25p) then aprox. 50Mb/s for 1280X720 @29.97p.
Right, and I was looking at the export bit rate.
I’ll leave it to others to make the grand test file you describe with crosshatches and music.
We NEVER judge color rendition with the naked eye. We use a calibrated test pattern and an eyedropper program.
Be careful with that approach as it may well be giving you the result where the built-in color profile of the screen is included.
It is much safer to use an external calibrated reference monitor and hardware vector scope.
It’s not 1080 interlaced; it’s 720p 59.94.
Aha, OK that makes sense.
I assumed because this discussion took a turn towards SC and options to interlace footage that those bit rates were all for interlaced footage.
It’s not. I’ve tested the he** out of it and the colors I put in are the ones the eyedropper puts out.
I know a little something about metrology. It’s far more accurate than eyeballing it. Shotcut has very good color accuracy now thanks to me being the squeaky wheel.
Unless you want to buy me a $$$ hardware vectorscope 
Well priced used WBR5000 (and earlier models) are available, spread the word amongst
colleagues, check equipment brokers, contact some local TV stations.
This is exactly what I did about a year ago and another station had just upgraded a bunch of equipment.
They were only too happy to get rid of their older stuff cluttering the shelves.
Picked up a WFM5000, a 24" Sony reference monitor and a pair of Yamaha MSP5 studio speakers for 1200 bucks.
They had plenty more kit, but I had no more available $$$. 
An eyedropper program works just as well for color checking on a PC. An eyedropper reads the color information ahead of any monitor calibration adjustments. If you know the magic equations you can also calculate Y, Cr and Cb as well as saturation.
Perhaps it’s my lack of knowledge about video cards, GPU’s and the like, but I don’t
feel comfortable relying on my computer to tell me what the colors are.
There is just too much that can get in the way (as I see it), things like OpenGL, firmware, drivers, screen profiles and who knows what else.
I’m much prefer trusting my SDI card connected to my Sony and Tektronix.
Again, I’ve done enough testing to feel confident in my testing rather than relying on hunch and superstition.
What testing have you done to make sure your measurement techniques aren’t introducing error? Have you verified that your measurements aren’t being affected by such things as monitor calibration?
I don’t really get your question.
Measuring the way I do is absolutely standard in a broadcast environment.
Both my reference monitor and Tektronix are calibrated by the same people that and come and do the station’s monitors and vector scopes every quarter.
Costs me 200 bucks every time so I only do mine twice yearly.
If your vectorscope is on the same video output as your monitor, then your measurements will be affected by your monitor calibration settings.
Also, what kind of test signal are you measuring on your vectorscope? Are they SMPTE RP 219 bars?
My technique measures ahead of the monitor calibration software so it is not affected.
Here are some spec-compliant 75% color bars. The RGB values should be16 or 180, i.e. green should be 16-180-16.
Sorry to spoil the fun, but discussion of color accuracy (interesting though it may be) is a little off-topic for this thread. I would like to get back to figuring out what codecs are actually supported by modern playout servers.
I’m seeing H.264 Long GOP and AVC Intra 100 in here. Other networks have indicated support for MPEG-2 Long GOP and some even accept ProRes (which was news to me too). What is happening with those files? Are they playable directly by an Omneon, or is the station being generous and transcoding? Is there a manual for these playout servers that says exactly what codecs they support? What are the most popular playout servers these days? For networks that allow the MPEG-2 Long GOP option, can ffmpeg (not ffmbc) create a ready-to-play file using the -c:v mpeg2video codec?
You are missing the point, the vector scope is not on the same output.
There are two HDMI outputs used for the editing screens (not calibrated) and an extra SDI
card which has it’s own separate output which feeds the SDI reference monitor and the Vector scope.
The SDI card has it’s own color profile and settings which are completely isolated from the standard HDMI outputs.
It depends.
How Omneon works is, you buy the hardware then a license for each codec you would like to playout.
We have licenses for XDCAM and DNxHD.
So either the stations in question have a Prores license or are transcoding elsewhere as a favour.
That is a very tough one to answer as it often comes down to non technical things like, price, support, historical use and how often the suppliers and the station’s buyer play golf together. 
A good guess would definitely be Omneon, Grass Valley, Cinedeck, EVS (if slo-mo is important, mostly for sports) and on the lower end of the scale (think community TV stations), 3Play is pretty popular.
Austin: It seems you are looking for a one-size-fits all solution – do I have that right?
I’m under the impression that you’d have to deal individually with the stations/networks you’ll be delivering to because each has its own delivery requirements. Paul, do you concur? It would probably pay to send your client(s) a test file to see if it passes muster.