[Jeff Hartman]

[Kip Draper] “>>Blackmagic’s gear is getting so darn shallow, someone’s going to have to start making racks that are only half a foot deep!

4” of depth works great w/ ATEM 1 M/E, but I flip the 1 M/E over so back becomes front making inputs/outputs easier to access, though there is enough clearance if mounted w/ flat, logo side facing forward. But if it’s tangle of thick HDMI cable or need to access it fast & often, backside out is best bet obviously.

I should be a bit more careful about kidding around… that was a tongue-in-cheek comment on my part. Anyone seriously considering using a very shallow rack needs to remember that the ATEM depends on convection for its cooling; unless there is a substantially unblocked body of free air behind the unit, it won’t to be cooled effectively. The other thing to bear in mind is that heavier types of coaxial cable like 1694a don’t want to bend as easily as a miniature type like 1855a. An ATEM that’s about an inch thick, plus another inch and a half of BNC plug on the cable, leaves only room to give the cable something like a 2″ bend radius to fit a 4″ deep rack — which is very likely to put excessive force on the connector.

Middle Atlantic used to have a really useful white paper on rack layout and heat management; as equipment gets denser (generating more heat in a smaller space), this is an important consideration.

— Jeff

[Jeff Hartman]

Funny… I saw both the 17″ and duo versions at a show last November, and it never dawned on me that they weren’t actually available yet. They’re both pretty impressive… one of the things that I rather liked is that the two ethernet connections let you daisy-chain a bunch of these monitors together to be fed from a single port on your network switch, reducing the volume of wiring.

Can’t believe I just wrote that last sentence: I date back to the days of tube-type Conrac monitors (that is, all tubes, not just the CRT)… there’s still something a bit over the top about plugging a monitor into a computer network!

Blackmagic’s gear is getting so darn shallow, someone’s going to have to start making racks that are only half a foot deep!

Jeff Hartman
Engineering Project Manager
Newport Television, Northeast

[Jeff Hartman]

It would be helpful to know which miniconverters you are asking about: while most add only minimal delay to the video signal (on the order of microseconds), a unit that incorporates a frame synchronizer like the up/down/cross converter will delay the signal substantially — on the order of 1/30th of a second for an interlaced format.

For a converter that doesn’t synchronize, the delay between input and output will be essentially constant. If you’re feeding many typical digital production switchers that have an input timing range of plus or minus half a line of video, the delay added by the converter is almost nil.

On the other hand, a synchronzier circuit will have a varying delay between its input and output — the exact delay at any moment depends on the relationship between input timing and the reference feeding the synchronizer.

Is your concern video system timing, or something else like minimizing latency in an IMAG system?

Jeff Hartman
Engineering Project Manager
Newport Television, Northeast

[Jeff Hartman]

[Richard Crowley] “Now if only we can convince BlackMagic that this is important information that should be given prominent place in the documentation (and even the online specs, IMHO). None of the photos are clear enough to see exactly how the four BNC and four HDMI connectors are labeled on the back panel of the hardware.”

I have reviewed a number of Blackmagic products for an online site, and poor documentation has been one of the main critiques for nearly every product. What passes for a manual would be marginally acceptable pre-sales literature for other professional video manufacturers. They make pretty good and extremely inexpensive stuff… but the buyer is making a gamble every time he purchases a new product, hoping that it will actually do what the promotional literature implies. (Or in the case of the ATEM Television Studio, hoping that it actually does what the literature explicitly claims: for instance, though it has been sold for a number of months, it wasn’t until a software release a few weeks ago that it could support standard definition formats.)

Two months ago I spoke with one of BM’s reps at a trade show about the lack of reasonable manuals; the response was that I don’t need manuals, that it was an emotional attachment that isn’t relevant today. That attitude shifted my thinking about Blackmagic products, and as a result I’m about to purchase a Ross Carbonite switcher rather than the BM switcher I was intending to buy. Blackmagic is cheap, but I don’t dare install it in a mission critical application where I might need product support.

Jeff Hartman
Engineering Project Manager
Newport Television, Northeast

[Jeff Hartman]

First off, this post is based on firmware revision 2.7.1, which changed some of this functionality. If you are running an earlier revision, you will want to first download this and update your system.

If you click the bottom “Settings” tab on the control panel, you will find pulldown options that let you choose the physical source for inputs 3 and 4. Inputs 1 and 2 are always HDMI; inputs 3 and 4 each may be either HDMI or SDI, and inputs 5 and 6 are always SDI. The significance of this detail is how the various inputs map to the multiviewer display. This screen is also where you enter the short and long names for each source; short names are used to label the buttons on the control surface, while long names appear on the multiviewer display.

Once you get the physical sources selected and named, clicking the top menu “File” / “Preferences” opens a window that lets you set which physical source is mapped to each button. The bottom line is that you may assign physical sources to buttons in any order — even assigning a source to more than one button. You may also select “none” as a source, which greys out and disables the button. Since the keyboard shortcuts (the numerals 1-8 for the preview bus, and SHIFT 1-8 for the program bus) correlate directly to the buttons on the screen, how you set up the preferences also determines how the physical inputs map to the keyboard.

Finally, if you’re using the GPI and Tally interface, its tally relays 1-6 always map to physical switcher inputs 1-6, and are unaffected by the button mapping — this lets you customize your user interface at will without worrying about messing up the tally.

Regards,

Jeff Hartman
Engineering Project Manager
Newport Television, Northeast

[Jeff Hartman]

Just a bit of background to help you interpret what’s going on…

Unlike most units of measurement, decibels are a relative measure, always compared to something else. It’s not like measuring X volts or saying something is N inches long; rather, it’s like saying a signal is three times more than… something. So a reading in decibels is not just the bare number; you also need to know what that “something” is — the reference.

In the world of analog equipment, one use of decibels is to describe the voltage of an audio signal; the reference used is the amount of voltage that you have when the signal delivers one milliwatt of power into a 600 ohm load. That measurement is usually shown as dBm (the “m” signifying “milliwatt”). Most professional analog audio gear nowadays uses +4dBm as the nominal operating level; back when I started in television, +8dBm was the more common reference level. Generally speaking, most analog equipment is capable of handling peak voltages around +24dBm, which means there’s about 20dB of headroom above the nominal operating level before clipping begins. The exact point of clipping varies from device to device, so a systems designer needs to be mindful to avoid serious mismatches that would lead to early distortion.

Now if you look at a piece of analog equipment with a VU meter — a console or tape recorder will do — you’ll see a scale in decibels with zero about two thirds of the way to the right. This is measuring level with respect to the nominal operating level. So if you have a mixer that has a +4dBm reference level as its standard, you will see +4dBm when the meter reads 0; drop the level so the meter reads -10dB, and the output voltage will drop to -6dBm. Raise the level to +6dB, and the output becomes +10dBm. Where analog VU metering falls short is in representing peaks; very few analog meters go anywhere near the typical clipping point of +20dB VU. But in the analog world, it’s pretty unusual to spend any real time there; moreover, a lot of analog devices can handle these overloads fairly gracefully.

Digital equipment, on the other hand, does not suffer overload well at all. There’s no room for “just a little bit more”; once the numbers hit the maximum value, that’s as far as things will go. And voila: instant clipping. This is such a serious issue that a great many digital devices measure audio in terms of that brick wall limit: and the measurement is dBfs, decibels with respect to Full Scale. Since the reference is the maximum possible value, that makes all dBfs measurements a negative number (which is the first thing that tends to hang people up). As a general case, most digital gear makes the assumption that the normal operating level is 20dB below the maximum, so for these devices -20dBfs will show as 0 VU, and will produce +4dBm on an analog output.

So… how does this play out in the real world? If you have a typical system with a mix of analog and digital equipment and throw tone on your mixer so that the VU meter shows 0dB, you should expect to measure a voltage of +4dBm on an analog output; feed it into a digital recorder, and you should see a reading of -20dBfs. This is entirely normal, and is exactly what you should be shooting for.

I could go on about the differences between average versus peak metering, but that’s more intuitive and not directly pertinent to the issue you’re having.

Regards,

Jeff

Jeff Hartman
Engineering Project Manager
Newport Television, Northeast

[Jeff Hartman]

Excellent. Thanks!

— Jeff

Jeff Hartman
Engineering Project Manager
Newport Television, Northeast

[Jeff Hartman]

Glad you were able to confirm our findings.

While you’re looking at this particular card, how about also adding some control over the reference: in particular, the ability to explicitly choose frame 1, frame 2, or the card’s direct BNC input? And second, to be able to control the output timing? As a practical matter, a synchronizer is much more useful if you can actually set it to be in a particular phase relationship to other sources.

Thanks,

Jeff

Jeff Hartman
Engineering Project Manager
Newport Television, Northeast

[Jeff Hartman]

Ouch! As it is, we have a rack and a half filled with FS1s (we’re a regional centralcasting hub, so we’re both sending and receiving every format short of PAL. It’s funny you say this, though: I’ve been wondering if they will wind up swallowing Evertz.

— Jeff

[Jeff Hartman]

BM just released software 2.7.1 last week, which added support for standard def, plus a number of other enhancements. I see they’re up to 2.7.2 today.

Jeff Hartman
Engineering Project Manager
Newport Television, Northeast