Let’s take a look under the hood of two broadcast VT formats: both are called HDCAM, both do some form of HD, but apart from the operational differences, the funky coloured flap and two extra letters, what is the real difference between them? Let’s find out.

HDCAM (non-SR)

HDCAM was introduced by Sony ages ago in 1997. No other tape based HD broadcast format was around at the time, and it quickly became clear that one was needed for HDTV to succeed. The HDW-M2000 machines are based on the DVW-2000 series of VTRs and record with a datarate of 144 Mbps (that’s a 50% increase compared to Digital Betacam’s datarate of 96 Mbps). HDCAM features 4 audio channels and can handle all common framerates between 24 and 30Fps.

But here’s the snag: a full HD frame is 1920×1080 pixels. To keep the datarate low enough to put this on tape, HDCAM – like its consumer cousin HDV – records frames at 1440×1080 pixels, where each pixel is horizontally squashed and expanded 1.33 times upon playback. The same idea is employed in 16:9 SD signals, where each “square” pixel is stretched 1.78 times to fit the wide screen aspect ratio.

In addition, HDCAM uses the same DCT compression that Digital Betacam uses very successfully. But unlike its predecessor, HDCAM compresses the YUV signal in 3:1:1 rather than 4:2:2. That’s not so good – but again the aim is to keep the datarate low.

Tape stock is essentially the same as Digital Betacam with the same particle density, and the only difference is a bright orange flap and some more holes at the bottom.

In a nutshell: HDCAM doesn’t actually record every detail that the HD-SDI signal potentially has to offer – so a completely new recording format needed to be created. Sticking with the same cassette shells made famous with the introduction of large Betacam SP tapes in 1986, Sony gave us…

HDCAM SR

HDCAM SR (as in “Superior Resolution”) was introduced 6 years after HDCAM and hit the market in 2003. SR records with a choice of 440 or 880 Mbps, which is over three times as much data as HDCAM.

The format prominently features 12 individual audio tracks and has a built-in Dolby E decoder (i.e. feed it Dolby E as an audio pair, and see the result as decoded single channels if you choose). The SRW-5000 VT machine looks and behaves completely different than its predecessors, which gives you the first impression of a very different beast.

SR’s main advantage is the ability to record the full HD signal “as god made it”: Square 1920×1080 pixels in full 4:2:2, 4:4:4 or even 8:8:8 (it depends on what you feed it; 8:8:8 recording at 880Mbps is achieved by twice the tape speed). Instead of DCT, SR uses the more efficient MPEG4/H264 compression for video (audio is uncompressed).

“Why would I need to record that much data, Jay? Isn’t that a bit overkill?” – Well it depends on your application. If you’re shooting on SR and you’re using it as a replacement for film, then you want to capture everything the HD image sensor has to offer. HDCAM SR can so this for you. SR tapes also come with a pre-attached Tele-File sticker, which contains a 1K memory chip to store meta data about your tape.

12 audio channels give multi-lingual broadcasters a distinct advantage, not to mention less deterioration in post production. At the time of writing it looks like HDCAM is now assumed to be a good TX delivery format, while SR is mainly used at the pre and post production stage.

Even though SR’s tape comes in Betacam cassettes, the particle density is twice as high as in HDCAM and Digital Betacam tape stock. Sony’s official stock brochure states: “We developed particles some 50% smaller than our previous best to capture wavelengths as small as 0.29mm — a human hair can be 300 times wider! The result is incredible performance: two times the output and 2.7 times the recording density of HDCAM tape”.

However, according to usage reports this higher density comes at a cost: minor shocks can cause severe impact on your precious recordings – which is why a new tape box has been introduced with the stock. Featuring a honeycomb pattern on the inside, sometimes even bubble wrap and tighter locks on the outside. Sony do not recommend for any stock to be re-used. That’s probably not feasible in reality, so beware of dropouts upon re-usage of your stock.

There you have it. Share your own experiences with HDCAM below if you feel like it.

Sources and Further Reading

Wikipedia’s page about the History of Betacam (covers HDCAM and SR)
http://en.wikipedia.org/wiki/Digital_Betacam

Forum Post at Creative Cow
http://forums.creativecow.net/thread/240/3

SR Stock explained in detail:
http://www.tapeonline.com/faqs/hdcam
http://www.tapestockonline.com/hdcamtapefaqs.html

Official Sony SRW5000 page (HDCAM SR VTR)
http://pro.sony.com/bbsc/ssr/cat-videorecorders/cat-rechdcamsr/product-SRW5000/

Official Sony HDW-M2000 page (HDCAM VTR)
http://pro.sony.com/bbsccms/ext/cinealta/post/hdwm2000.shtml

Official Sony HDCAM SR stock brochure:
http://pro.sony.com/bbsccms/assets/files/mkt/recmedia/brochures/prodbroch_hdcamsr.pdf

This is how nPSK works...

We made some television history the other night: in a test for the NFL International Distribution, Adrian from TV2Go and us folks at IMG Mediahouse have successfully brought a 32APSK encoded signal across the Atlantic, with a datarate of 30Mbps using a 9MHz carrier!

Since the start of the current NFL season in 2009, part of my job is to bring in feeds that originate in the US for an uplink stateside to NSS7. Our clients then downlink these feeds in Europe. In North America, DirectTV offers the NFL (and nearly 200 other channels) to domestic customers via their fleet of ten satellites. Using MPEG2 and H264 encoding, QPSK and 8PSK modulation are commonplace in satellite transmissions.

Ray was looking after this year's NFL season (I was behind the camera of my mobile phone)

However satellite bandwidth is expensive, and we’re dealing with demands of ever increasing datarates. Compression alone will only get us so far, so the only other way to increase throughput would be to increase phase modulation on the TX path. Lucky for us, the DVB-S2 spec has another trick or two up its sleeve.

Using 32APSK modulation and Newtec equipment, Adrian, Ray and I successfully brought over a stable 30Mbps connection using only a 9MHz carrier. With our 5.6m dish, carrier to noise was 18dB – which is not too shabby. Newtec kindly provided an NTC 2263 receiver for this, which provides an ASI output – so we had to wire it into a Tandberg 1290 IRD to decode the signal into SDI.

The results are gooood! In layman’s terms: 32APSK increases the amount of data that can be sent in comparison to 8PSK by about 30%: with 8PSK modulation on a 9MHz carrier, we’d be getting a maximum data rate of 20Mbps. In 32APSK we’re getting 30Mbps!

8PSK Gray Coding - imagine this x3 for 32APSK

As I understand it, 8PSK sends all bits in one phase revolution, or orbit. 16APSK uses twice as much using two orbits, and 32APSK uses THREE orbits. How exactly this works is beyond me – but it’s sure as hell exciting to see the results!

On the transmission front, Adrian was using an Adtec Encoder for this project. Although MPEG2 is commonly used in satellite transmissions, we were utilising MPEG4/H264 compression which is more efficient than MPEG2 – so that’s TWO new technologies in one fair swoop!

The Newtec seems to be the only demodulator that can currently cope with 32APSK downlinks. Even though the Tandbergs have 16APSK and 32APSK as an option in the Licensed Features list, I understand these are hardware upgrades, not software upgrades.

And speaking of Tandberg 1290 receivers, we encountered a slight problem: every once in a while (say every 10mins or so), the 1290 would glitch for a couple of seconds, as if it’s not coping with error correction (at least that’s what it looked like). Apparently this phenomenon happens because of i-frames sent at irregular intervals from the encoder. I understand that due to different on-board chipsets, the Tandberg motherboards 3.x are NOT causing these glitches, whereas the version 2.x do (at the time of writing, we currently have the latter ones here at IMG).

“But what’s the application, Jay? Why go through all that trouble and pay for a new infrastructure?” I hear you ask. Simple really: saving money. If you can use the same satellite space for more data, then you’ll be able to use a 6 MHz carrier instead of 9 MHz or provide more services in the same space. Adrian suggests carriers as narrow as 3MHz for decent broadcast quality – superb for news gathering.

We’re hoping to utilise 32APSK with MPEG4 for the forthcoming NFL season 2010/2011 starting in September. Let’s keep our fingers crossed!

Newtec NTC 2263 demodulator wedged between two Tandberg 1290 IRD's

My own notes, invaluable in times of severe panic

Sometimes a feed comes in, gets an English voiceover and is put straight to air on ESPN UK. But on Sunday, we pulled various aces from our sleeves to bring the Serie A “Milan Derby” to an expecting UK audience.

The feed comes in from Eutelsat W3 in HD, the studio for pre match presentation is SD, the commentators are on site in Milan and come in via ISDN. Question is: How do we synchronise comms with pictures, and how do we create HD graphics during the match?

We use pretty much EVERY piece of kit IMG Mediahouse has to offer, that’s how.

In this article, I’ll explain some of the challenges my dayjob in MCR has to offer. My colleagues and I feel like rockstars when a feat like this can successfully be pulled off, and I’m happy to say we did it once again. We’ve never done this particular route before, so it deserves a mention here.

First of all, the audio commentary will reach us instantly via 128 kbps ISDN circuit. Even though we have our man Paul Ryan on site, we have no influence on the picture circuit. It’s provided by Infront Media, who do a similar thing to the Italian Serie A football that we (IMG) do for the English Premier League: international distribution. This involves getting a feed from the stadium (either via fibre or satellite) to a dedicated facility, add graphics and potentially commentary and send it out for further satellite uplinks around the world.

more notes...

Here’s the snag: By the time we see the signal, it could potentially be bounced off two satellites, each path taking roughly one second. Take into consideration compression and decompressions along the way, the pictures could arrive roughly 2-4 seconds later than the commentary reaches us – how long exactly requires a sync test. At this point, we’re not sure if the host broadcasters will let us step in front of the camera.

Even so, one delay unit alone won’t be able to provide a full 4 seconds or longer should we need it, so we’ll have to plan for TWO units. Here’s the signal path:

• ISDN commentary comes in as analogue audio
• pass through a BEL delay unit (we’ll set this to 1000ms or one second for now – maximum here is 2400ms)
• delayed output needs to be converted via Proc Amp into embedded SDI audio for distribution around the building
• Proc goes into yet another Proc with option to delay a maximum of 2600ms
• this delayed Proc is onpassed to the Studio so commentary can end up on air eventually
• in addition, an UNDELAYED version straight from the ISDN unit is passed to the Studio so the presenters can interact with the commentators in real time

That’s audio sorted. Now let’s have a look at the video signal:

• The clean pictures come in off satellite in HD
• a downconverted version is sent to the Studio
• the HD feed is sent to a HD capable gallery, which is connected via TBU to the Studio Gallery (they’re in different buildings, and we’ve run out of tie lines by this point)
• we’ll create the HD graphics in one gallery, while audio gets mixed in the main studio
• we then take this dirty HD video feed and combine it with the audio from the SD Studio via an interesting bit of kit known to us as Shuffler 4
• Shuffler 4 (HD pictures plus Studio’s audio) are sent via HD SDI to ESPN TX at Arqiva Chalfont on dedicated fibres. We’re also providing an HD ASI feed via direct line to Arqiva Newman Street, and a downconverted SD backup feed is sent via BT Tower

Confused? That’s nothing! Various requests for recordings at various stages of the process require further audio shuffling and downconversions, but let’s not mention this here.

All that counts is that all went very well and our clients were happy.

Join me next week when I’ll tell you about an interesting 32APSK downlink test we did for NFL football.

Here’s a funny one…

I made this video on Xbox drive noise over a year ago. It has now had 55.000 views and 129 ratings on YouTube. Out of a sudden YouTube are inviting me to cash in on it.

It means that they’re eager to display adverts in the video and on the page. I can apply for this particular video only, and it won’t affect any of my other 130 videos in my account.

But here’s the snag:

if YouTube REJECT my application, the video and all its 462 comments video responses will be DELETED forever! It means that all my loyal fans will lose out on a smashing exposure on how bad Micro$oft and their rotten console is. It’s a bit like being in Vegas at the roulette table!

The sign-up tutorial is very nice and straightforward, and it outlines the copyright pitfalls one can get caught out by. In my case, it’s the inclusion of the Xbox 360 logo at the beginning and the end of the clip. I could argue it’s fair usage… but now that there’s money involved, bureaucrats may get a bit funny.

So what shall I do? Gamble and put everything on red, but risk losing it all? Or become a gazillionaire over the next 12 months?

Have your say below – and enjoy the video