HD to SD downconvert – Best Results Inquiry

[John Fishback]

I’d use my Kona 3 to do the downconversion in real time. This is discussed over and over, so do a search and you’ll find lots of answers. The prevailing opinion will be hardware downconversion is best, but if you must do it with software, then use Compressor.

John

MacPro 8-core 2.8GHz 8 GB RAM OS 10.5.5 QT7.5.5 Kona 3 Dual Cinema 23 ATI Radeon HD 3870, 24″ TV-Logic Monitor
ATTO ExpressSAS R380 RAID Adapter, PDE Enclosure with 8-drive 6TB RAID 5

Final Cut Studio 2 (up to date)

Pro Tools HD w SYNC IO, Yamaha DM1000, Millennia Media HV-3C, Neumann U87, Schoeps Mk41 mics, Genelec Monitors, PrimaLT ISDN

[John Collucci]

Thanks for the tip. I’ve done a few searches already but with nothing completely relevant showing up.

My Kona card is SD only, we’re in the process of an HD upgrade, but in the meantime I suppose that a software conversion will have to do.

Anyone else have any ideas? Is there any reason why I shoul use compressor over Quicktime Pro? I imagine that de-interlacing is best as well, but I’d really like some input from some more people who know before I decide on a workflow.

Thanks for the reply John!

[John Fishback]

The reason Compressor is superior is its scaling algorithms are better. Turn Frame Controls on and choose the best quality settings in the Encoder tab of Compressor’s Inspector.

John

MacPro 8-core 2.8GHz 8 GB RAM OS 10.5.5 QT7.5.5 Kona 3 Dual Cinema 23 ATI Radeon HD 3870, 24″ TV-Logic Monitor
ATTO ExpressSAS R380 RAID Adapter, PDE Enclosure with 8-drive 6TB RAID 5

Final Cut Studio 2 (up to date)

Pro Tools HD w SYNC IO, Yamaha DM1000, Millennia Media HV-3C, Neumann U87, Schoeps Mk41 mics, Genelec Monitors, PrimaLT ISDN

[Michael Gissing]

A few matters of confusion. Shift fields is happening because NTSC HD is upper and NTSC SD is lower. Nothing to do with centre cut versus letterbox.

Most broadcasters prefer 16:9 anamorphic (often called Full Height) This is automatically centre cut on analog broadcast and correctly displayed on digital transmission with a suitable 16:9 monitor. It is also correct for DVD.

Deinterlacing has nothing to do with downconversion. It is a look and should be thought about carefully. FCP’s deinterlacing is poor and throws away resolution. Third party plugins like Nattress are better and let you control blend parameters shot by shot. Compressors deinterlace is also better than FCP but is global and not shot specific. Given the look of the program was meant to be ‘home vid’ I would not deinterlace.

By far the best downconversion is to use an external hardware device like Teranex, Kona cards or even the built in downconvert in the Sony HDCam decks. They all look better to me than Compressor and the worst is letting FCP render down. The shift fields issue is only happening when you drop the sequence into an SD FCP timeline. All the hardware options fix this automatically on the fly in real time.

[Walter Biscardi]

AJA Kona products give you the cleanest downconversion. We do this almost daily for broadcast delivery. You have the options of Letterbox, Anamorphic or Center Cut for your downconvert.

Best of all it’s realtime so you simply lay off to SD from your HD timeline.

Walter Biscardi, Jr.
Editor, Colorist, Director, Writer, Consultant, Author.
Credits include multiple Emmy, Telly, Aurora and Peabody Awards.
Biscardi Creative Media

Creative Cow Forum Host:
Apple Final Cut Pro, Apple Motion, Apple Color, AJA Kona, Business & Marketing, Maxx Digital.

Read my Blog!

[Steve Oakley]

well I think if you are getting ready to upgrade your hardware, you should take a look at the MXO2 series product. it can down scale on output, but it can also upres on capture which works really well for getting a few stray SD shots into a HD project. the upscale is something the kona cards can’t do.

[John Fishback]

The Kona 3 up converts, down converts and cross converts in real time at broadcast quality with complete contol of letterboxing, center cut and pillar boxing as needed.

John

MacPro 8-core 2.8GHz 8 GB RAM OS 10.5.5 QT7.5.5 Kona 3 Dual Cinema 23 ATI Radeon HD 3870, 24″ TV-Logic Monitor
ATTO ExpressSAS R380 RAID Adapter, PDE Enclosure with 8-drive 6TB RAID 5

Final Cut Studio 2 (up to date)

Pro Tools HD w SYNC IO, Yamaha DM1000, Millennia Media HV-3C, Neumann U87, Schoeps Mk41 mics, Genelec Monitors, PrimaLT ISDN

[John Collucci]

Thanks for the advice John.

I’ve tried compressor to go to SD Uncompressed 10-bit NTSC with a center cut, but I can’t seem to disable the de-interlacing. Honestly, the new downconverted footage from compressor looks much worse than the simple downscale and render that I did in FCP as a working edit.

When I say “looks worse” I’m referring to what looks mostly like aliasing. Combined with the de-interlacing, it’s unusable.

For now I am working completely within FCP since it seems to be yielding better results conversion wise. If anyone has any experience that they’d like to share on how to get a good SD downconversion in compressor, I’d love to try it, and it would be greatly appreciated.

~j!

[Dan Appel]

Here’s my beef, though, with Kona 3. In the HD-SD downconvert, I see a rather noticeable halo around edges (esp in contrasty areas and GFX). When we do the same conversion through a Teranex mini, you don’t get the halo, but it feels ever-so-slightly… over-processed? A little in-organic.

And whilst on the subject, it seems as though there are options, when sending dubs/conversions out of house, that some facilities just go deck-to-deck (HDCAM SD-SDI out to Digibeta) and other route through a (full-sized) Teranex or UKON. Is there a difference? Is the in-deck downconvert any better than the convertor boxes? Is there a a best case scenario for one process over another (ie: frame rate changes; aspect ratios)? Primarily, I’m a fanatic for quality (much to my staff’s dismay), so any insight is appreciated.

And sorry for being a little off-topic. ThxD

[Joe Procopio]

do you do anything for the audio? downconvert we just did was off a frame consistently throughout output. do you have to offset the audio somewhere in the FCP prior to output to compensate?

Broadway Video, NYC
AVID/FCP editor/engineer