-
KONA 3 conversion delays in processing
I understand there is a delay in processing for certain conversions going from FCP through the KONA 3, so I wanted to do some testing so I could figure out what these are and compensate accordingly. What I found was interesting, and I would like to compare with others to see (A) if you get the same results, and (B) how you would compensate for some of these. Here are my results:
When not performing any conversion on resolution or frame rate, the KONA 3 is frame accurate in edit to tape for video and audio.
When upconverting (525i29.97 to 1080i29.97 – only resolution, not changing frame rate) the output from KONA 3 is frame accurate for audio and 1 frame late for video. To compensate for this, I would shift video in my sequence earlier by one frame.
When downconverting (1080i29.97 to 525i29.97 – only resolution, not changing frame rate) the output from KONA 3 is frame accurate for audio and 1 frame late for video. To compensate for this, I would shift video in my sequence earlier by one frame.
When converting (1080sf23.976 to 1080i29.97 – adding pulldown only, not changing resolution) the output from KONA 3 is frame accurate for both video and audio.
Now here is where it gets interesting:
When crossconverting (720p59.94 to 1080i29.97) the output from KONA 3 is frame accurate for video and audio is one field early in 29.97i.
When crossconverting (720p23.976 to 1080i29.97) the output from KONA 3 is frame accurate for video and audio is approximately .4 frame early in 29.97i.
When downconverting (720p23.976 to 525i29.97) the output from KONA 3 is frame accurate for video and audio is one field early in 29.97i.How would you compensate for these last three situations? Are my results similar to anyone else’s here? The only true solution I could come up with is to do a separate edit for video (at conversion) and audio (with no conversion), since I can’t easily shift audio accurately by less than a frame in FCP. Any advice?
OS: Leopard
Mac: Intel 1,1
AJA driver: 6.0.3
FCP: 6.0.2Decks used in output tests: Sony DVW-A500, Sony SRW-5500
-
8 Replies
-
[Mark Spano] “since I can’t easily shift audio accurately by less than a frame in FCP. Any advice? “
Actually, you can.
Jeremy
-
Jeremy, I appreciate the link, but unless you have an obvious audio sync point at that spot, that is not quite an accurate subframe edit. If I know audio in my sequence starts at 01:00:00:00 and I want to shift whatever audio is there back to exactly 00:59:59:29.5, it’s not possible to be that accurate in that audio editor. I can guess at it, but if there’s no obvious audio start (for example in cases where there’s a fade up, silent start, etc.) then I can’t be exact exact. Which, as an audio guy coming from 48000 samples, I want… 🙂
-
Sure, but it’s an option.
-
Hence the thanks… 🙂
-
If you are happy with your mix, you could export it from the timeline, duplicate your sequence, delete the audio tracks and bring back your exported mix. send that to soundtrack “multitrack” project and move the audio xx number of samples or milliseconds then export that, reimport
A lot of work for .4 frame or a field. Of course a separate edit to tape insert could be a lot depending on the TRT.
Have you looked at the delay down converting from the SR to dbeta?Doug Beal
Editor / Engineer
Rock Creative Images
Nashville TN -
Those are the options, it appears. My tests have been confirmed by AJA support. Here’s what Rudy had to say:
Hi Mark,
You’re findings are correct.
Basically there is a 1-frame delay for video only when the Kona 3 is
configured for up down or cross conversion.
The reason is the conversion is processed in the Kona 3 hardware.
The workaround is to delay the audio by 1-frame in the sequence.Cross converting from 720p to 1080i interlaced is approximately one field
because the frame rate was originally 59.94 and changed to 29.97 (59.94
field rate).Downconversion through the SR deck (in E-E) is not something I’ve tried yet, but I suppose it could come in handy. I know that if you’re playing an SR tape that the conversion output is frame accurate. I’m just glad right now that I’m not crazy and that my tests have been verified by AJA. I suppose it’s also good to have this documented here as well. Thanks for all your help.
-
Mark
Just ’cause your tests have been proven to be correct, doesn’t mean you’re not crazy. It only means your tests are correct and have been proven.
My knowledge in this regard is based on the old “it takes one to know one” and this stuff is what has driven me crazy. as a side note, just because you’re paranoid doesn’t mean they’re not out to get you.Best
Doug Beal
Editor / Engineer
Rock Creative Images
Nashville TN -
You’re absolutely right. I should clarify: I’m not crazy for finding these things out. In all other regards, however, yes.
Reply to this Discussion! Login or Sign Up