Up-Converting SD for HD Projects and HD Cross-Conversion Using AJA Kona 3
COW Library : AJA Video Systems Tutorials : Shane Ross : Up-Converting SD for HD Projects and HD Cross-Conversion Using AJA Kona 3
When choosing capture cards that are compatible with Final Cut Pro, there are many options. The one you choose may be dependent on the given workflow you will be utilizing. I have used several capture cards on a variety of projects: Aurora Pipe Pro, Decklink HD, AJA Kona LH and Matrox MXO. But when it came time to choose a card for the latest show I worked on, all the research led to one card, the AJA Kona 3.
The reason I chose this card is two fold: First, we shot DVCPRO HD at 720p 23.98 utilizing the Panasonic Varicam (tape) and the HVX-200 (tapeless), and we edited on a 720p 23.98 timeline. Second, we needed to deliver 1080psf 23.98 on HDCAM. Currently the AJA Kona 3 is the only card available that will do this type of crossconversion in real-time on output. And this is a real time saver.
In order to perform this type of conversion from 720p 23.98 to 1080PsF 23.98, you simply set the Primary Format in the AJA Control Panel application to 720p 23.98 and the Secondary Format to 1080PsF 23.98.
Since this series is historical in nature, we needed to incorporate a lot of footage that is standard definition from a variety of formats to match our high definition masters. Before we started production, I consulted with TJ Ryan and Stephen Hens from Chop House Edit and together we put together a work flow to accomplish this. While Decklink and AJA both offer solutions to upconvert the standard definition footage to high definition, the only card that provides hardware-based upconversion of standard definition is the AJA Kona 3. The conversion is all done via the card's hardware, not via the computer's software/CPU processing like the competitors. In part, this is because this upconversion is a hardware-based upconversion, meaning that the conversion is all done via the card's hardware, not via the computer's software/CPU processing like the competitors. In this respect the upconversion the Kona 3 does is similar to a "mini-Teranex" hardware-type converter. Since the format I will be working with on this series is 720p, that format will be the focus of the article. But know that you can use the Kona 3 to upconvert standard definition to 1080i as well, and the quality it provides is very good.
Now while the Kona 3 can upconvert standard definition to HD, what it cannot do is a frame rate conversion. The KONA 3 performs "like Hz to like Hz" conversions; this means that you can upconvert from 525i29.97 (59.94Hz) to 720p59.94 (59.94Hz) and 525i29.97 to 1080i29.97 without issue. It can do some frame rate conversions upon output (23.98 to 29.97) even while down converting the footage from HD to SD, but it cannot convert the frame rate while it upconverts. Not to worry...there is a handy way to get the 29.97 standard definition footage to the 23.98 frame rate that we need. That handy way is via Apple's Compressor application, but we'll get into that later.
While the AJA Kona 3 has a variety of output types - SDI, component and composite - the only input type it has is SDI. This isn't an issue if you have decks that feature SDI connectivity, such as Digital Betacam, DVCPRO 50 and the higher end DVCAM decks. If you have access to these decks, then you are set. However, if you don't (like I didn't), then you'll also need to get the AJA HD10AVA mini converter or similar box. This converter takes a component or composite analog signal and converts it into a digital SDI signal with embedded audio.
The first thing to do is connect the deck to the card. With SDI (and HD-SDI), that is pretty much the only connection you need for footage. SDI carries up to 8 channels of embedded audio. If you only have component outs on your deck, then you'd use the HD10AVA to convert the signal. The converter also has audio inputs that you'd connect to the audio outs on the deck via XLR connections, so it will take component video and analog audio and convert it into an SDI signal with embedded audio. Then you connect the RS-422 cable for deck control. Done, simple. While reference/genlock/sync is generally a good idea, it is not required if you are capturing from a single deck connected to the card. Reference is recommended when performing an output to tape.
Now you need to open Final Cut Pro and go into the EASY SETUPS located in the FINAL CUT PRO menu. Choose the EASY SETUP that matches the format you intend to capture the footage as. In this case, I will be working with DVCPRO HD 720p60. For FORMAT choose HD, for RATE choose 59.94, and for USE, choose AJA Kona 3 - 720p 59.94 DVCPRO HD. This will tell FCP to take the incoming 720p signal that is being fed from the Kona card, and capture that footage using the DVCPRO HD codec.
Next open the AJA CONTROL PANEL, located in the Applications folder. In there, you need to set the FRAME BUFFER to the format that you are working with. You do that by CONTROL-clicking or right clicking on the buffer and choosing the format you need to work with.
With your Frame Buffer set to 720p59.94, now it is time to tell the card to upconvert the footage. Again, this upconvert is being done by the Kona card, so we need to make the settings changes in the AJA Control Panel. Click on the FORMAT tab in the control panel. The Primary Format should match what you wish to have in the Frame Buffer, in my case, 720p59.94. Then the Secondary (Converted) Format should be set to 525i29.97.
You will now notice a little box between your SDI IN and Frame Buffer that says UP CONVERT. Now the card will upconvert the footage from SD to HD. But how will it deal with the aspect ratio difference? Standard Definition NTSC is 4:3, whereas High Definition is 16:9. In the same FORMAT tab you will see a section called CONVERSION.
This is where you choose how you want the aspect ratio to be dealt with. Do you want the footage to be Pillarboxed, meaning you see the full 4:3 frame, but with black on the sides to fill the 16:9 frame (so you see the full image), or do you want to Zoom 16:9, meaning that you will capture the 4:3 to fill the 16:9 frame, but cropping off the top and bottom of the image in order for it to fill the frame. This will zoom in further than Pillarbox will, so it won't be as clear. But the choice you make is up to you, and what the client or network will accept. We ended up capturing most of the footage ZOOM LETTERBOX, but with a few shots as Pillarboxed, so that we could either cut it into the show that way as a stylistic choice (network OK'd that option), or we could zoom in on the footage after we added it to the timeline, and adjust the framing manually so that we controlled where the cropping of the image would occur.
Now that you have the Kona card setup, time to move back into Final Cut Pro. Open the LOG AND CAPTURE window located in the FILE menu, make sure that you are capturing the audio channels you want. Often FCP will default to capturing 8 channels of audio in this format, deselect the other channels until you get to the 2 or 4 channels that you might need. Choose your Capture Scratch drive and then start capturing your footage...or logging and capturing. I prefer to log my footage well. Make sure to add a NAME to the footage before you do a capture now to avoid those UNTITLED clip names.
Now you have all your captured footage. Captured into the DVCPRO HD 720p59.94 codec. But you need it to be 23.98 to match the rest of your 23.98 footage from your camera masters. Now we turn to the Compressor application that comes free with Final Cut Pro Studio.
Compressor is full of wonderful presets. In the ADVANCED CONVERSION folder (located in the Other Workflows folder) there are presets for pretty much every EASY SETUP you have in FCP (minus the ones that are installed by your capture card). So all you need to do is import the footage (drag and drop into the drop zone) and then drag the preset you want, in this case the DVCPRO HD 720p24 setting (in the High Definition subfolder), onto that clip. Click on the FRAME CONTROLS in the INSPECTOR and then click on the sprocket to allow you to turn them on.
In here you can set the OUTPUT FIELDS to PROGRESSIVE (leave the RESIZE alone, we already did that), DEINTERLACE to "better" or "best"...or reverse telecine if this footage is film telecined to tape...and then adjust the RATE CONVERSION to what you want. You might want to play with these settings, because you might not need to use them at all, or you might need to use all of them on BEST. Mind you, the better you make this conversion, the slower this will take to convert. For the sake of this article, I chose the Field Output to Progressive, And chose "best" for all of my options, with FIELD ORDER being set to PROGRESSIVE, as that is how we want the footage to end up. This produced good results and not too long of a conversion time. Helped to have a 3.0Ghz MacPro, and an assistant editor to do this at night and to let run overnight...But this is doable on any system. Time being your only issue.
As a side note, if you are going to be converting to a format that is not, DVCPRO HD, like ProRes (HQ) or Uncompressed 10-bit , and you are going to make custom settings, you need to manually set the frame rate to 23.976, not 23.98. 23.98 is just rounded up from 23.976. But if you choose 23.98 in those other formats, the timing will be slightly off.
And if you are working in PAL, the frame rate conversion isn't an issue, since both formats, 25p and 50i, are both 50hz formats.
So if you find that your workflow requires you to work in high definition, and you have the need to convert various formats of standard definition into high definition, be it 720p or 1080i, then the AJA Kona 3 is the right capture card for the job. It does this conversion in real time, via hardware and is comparable to stand alone hardware converters found in many tape dubbing facilities. In fact, the Kona 3 can be used as a stand alone converter as well. You need not use FCP to reap the benefits of the up converting capabilities. Just make all the adjustments you need to make in the AJA Control Panel, and then route the signal through the card via the Input PassThrough setting. Just like in FCP it won't do the frame rate conversion. For that you still need to rely on capturing and converting with Compressor.