HDCAM 1080i60 (with 3:2 pulldown) to HDCAM 1080p25 etc

[Blue_MiSfit]

Hey guys,

I've been tasked with converting a 1080i60 HDCAM tape to a 25Psf or 50i HDCAM tape.

I don't have access to an HD standards converter like a Snell Alchemist, so I have to capture it, process in software, and then lay back to tape.

Capture via Digital Rapids and x264 to a high bitrate H.264 stream is very easy.

Thankfully, the content is all hard telecined, so I can easily use decomb or tivtc to bring things back to 24p. No need for slow mocomp or ugly blends for this project!

I get a tiny bit confused when it comes time to go to 25p.

In AviSynth, I can just to AssumeFPS(25), and I will have 1080p25. Add in PAL speedup with eac3to, and this should be compatible with HDCAM. Do I need to do anything about interlaced vs progressive chroma subsampling? I guess this reveals my actual ignorance on the technical difference between Psf 25 and "true" p25! What about fake 50i?

Obviously I've lost a lot by capturing to 8 bit 4:2:0, but the results certainly look nice.

Should I alternatively apply euro pulldown?

Also, what's a good system for laying back to tape? My best attempt has been to encode the result into MJPG for Final Cut Pro compatibility, so I can use its "edit to tape" functionality.

Any suggestions?

Derek

[Manao]

You don't need to change the chroma pixel position (they are located at the same place in progressive and in interlaced). However, if you do any processing involving chroma, you'd better do it progressively, after inverse telecining the video.

The one place you can easily mess up chroma is when you captured the video. If possible, you should try to capture it in 4:2:2 instead of 4:2:0, because odds are the 4:2:2 -> 4:2:0 subsampling that was done during the capture was done interlaced, while it should be done progressively on the inverse telecined content.

PSF 25 is just a way to store the video, it doesn't change it visually (or rather, it wouldn't change it if you were storing it losslessly. Since you're lossy, PSF will bring different artifacts than a "normal" progressive storage).

I wouldn't apply a pulldown for 24p->25p.

[Blue_MiSfit]

Indeed. I should probably capture using a 4:2:2 compatible encoder, not x264 :(

My biggest problem now seems to be getting a solid understanding of how laying back should work. I'm no expert at configuring HDCAM decks, either ;)

[Mug Funky]

yeah, capture in 422.

as far as the speedup - timestretching the audio could become a sticky point. you might actually want to lay it down in 23.98PsF, which would allow the deck itself to play in 25p if required (with it's own timestretch on the audio).

the good thing about 23.98 PsF is that it's NTSC and PAL agnostic when it's in a HDCAM/SR deck. if the client wants the audio handled more delicately, they can book some time getting their original sound-mix sped up in the DAW it came from originally (which avoids a lot of timestretching problems as it can work on solo tracks).

[hanfrunz]

Indeed. I should probably capture using a 4:2:2 compatible encoder, not x264 :(

My biggest problem now seems to be getting a solid understanding of how laying back should work. I'm no expert at configuring HDCAM decks, either ;)

If it's HDCAM (nor HDCAM SR) you don't need 4:2:2, because HDCAM uses 3:1:1 subsampling!

To change the systemfrequency on a HDW-F500 vtr do this:
1. eject tape
2. push the maitenance button with a pen (left side)
3. press SFT + mainten exec
4. press other checks
5. press system menu
6. set frequency
7. press exec
8. press exec again

[Manao]

If it's HDCAM (nor HDCAM SR) you don't need 4:2:2, because HDCAM uses 3:1:1 subsampling!I disagree, you're still better off with 4:2:2. Both 3:1:1 and 4:2:2 keeps the full vertical resolution for the chroma, the only difference is that 3:1:1 decimates chroma horizontally by a factor 3, while 4:2:2 only decimates it by a factor 2. Anyway, the key here is keeping the chroma vertically unsubsampled, until you can ensure a proper conversion, ie one that takes into account the nature of the content (progressive vs interlaced vs telecined)

[Mug Funky]

additionally, the deck's upsampling from 1440@3:1:1 to 1920@4:2:2, and the sharpening it applies, are part of the "charm" of the format.

i've enountered DP's who prefer to shoot HDCAM versus SR on a given camera simply because of this effect. of course they wouldn't use it for chroma-key work...

personally i think it's a stupid format... sony rushing a HD format to market to compete with panasonic, but not having the head tech to pull off a higher bitrate than 144Mb/s