[Graeme Nattress]

I wish Apple would give out more info on this 🙂 In that mode, 32bit float is the data format used internally. Any FXPlug filter should be producing float data, but again, how accurate, I don’t quite know. However, in this mode, you can darken an image, and in the next plugin brighten it up without the losses you’d have if you were using an 8bit intermediate stage.

Rounding will be to 8bit or 10bit depending on timeline codec, and, right at the end.

Graeme

– http://www.nattress.com – Film Effects and Standards Conversion for FCP

[Graeme Nattress]

Actually, the YCbCr colour space is much larger than RGB, so you can make totally illegal colours in YCbCr that cannot be displayed in RGB. I think you’ve got your explanation in reverse…..

Graeme

– http://www.nattress.com – Film Effects and Standards Conversion for FCP

[Graeme Nattress]

Incorrect. FCP has had 32bit float support for rendering for ages. 10bit video uses it by default, but you can turn it on for 8bit video by selecting the “high precision YUV” option.

Graeme

– http://www.nattress.com – Film Effects and Standards Conversion for FCP

[Graeme Nattress]

Adding fake setup inside FCP is not just technically the wrong place to do it, but will lower image quality.

There’s no need at all to add setup for monitoring on a monitor that’s correctly calibrated to your NLE. None at all. After all, the Japanese who wisely ditched setup (almost, but not quite, makes up for their love of interlace – yuck) and they can monitor stuff accurately.

Digital video, of course, never has, never should have and never needs to have any setup added. If you add it, faking it in FCP, all you are doing is making your digital files incompatible with the rest of the world, and lowering quality into the bargain.

If you need to add setup, add it in the analogue domain, as it’s purely an analogue phenonmena.

Graeme

– http://www.nattress.com – Film Effects and Standards Conversion for FCP

[Graeme Nattress]

Chroma resolution of NTSC BetaSP is about half way between that of NTSC DV (4:1:1) and NTSC DigiBeta (4:2:2), however, it’s luma resolution is about 75% of that of NTSC DV, or NTSC Digibeta, which are both as good as each other in terms of luma resolution.

By analogue compression, I mean the process of whereby you take a signal from the camera head that has a certain bandwidth, and that bandwidth gets reduced enough to fit on tape. This is usually noticed by an increase in noise, and a lowering of resolution. Analogue artifacts can include filter ringing, shimmer, subtle timebase errors, noise etc.

8 bit uncompressed will capture BetaSP without making it any worse at all. DV50 would be almost imperceptibly different, and a lot smaller file size.

PAL complicates things as there, PAL DV is 4:2:0, and with the way PAL works, you don’t see the extra vertical chroma resolution of 4:2:2, so a PAL DV copy of a BetaSP original would probably be, overall, better than an NTSC DV dub of a NTSC BetaSP. I do have a couple PAL betaSP to PAL DV dubs and they’ve always looked very good to me.

As Dave pointed out, 4:4:4, even if there was such an SD capture codec, would be uber-overkill.

Graeme

– http://www.nattress.com – Film Effects and Standards Conversion for FCP

[Graeme Nattress]

FCP has had, for ages, 32bit float rendering. Making use of it might be tricky, but it’s in there. The issue was that old AE API FCP plugins were limited to 8bit RGB, but native FCP plugins didn’t have that limit (but had other constraints). Complex subject….

Graeme

– http://www.nattress.com – Film Effects and Standards Conversion for FCP

[Graeme Nattress]

Well, they are very equivalent in overall quality, but it’s a mistake to take one highly compressed format and convert it to another highly compressed format (if you’re concerned about image quality) as you’re just going to get the worst of both worlds, analogue compression mixed with digital compression. That said, depending on what the end result is, and how critical it will be looked at, it may make no practical significant difference to the output.

Graeme

– http://www.nattress.com – Film Effects and Standards Conversion for FCP

[Graeme Nattress]

It’s not the codec that has an issue though, it’s the 32bit float rendering pathway that has had the issues.

Graeme

– http://www.nattress.com – Film Effects and Standards Conversion for FCP

[Graeme Nattress]

Chroma Sharpener is 10bit safe, but the G Nicer isn’t. That’s one of the reasons I wrote the Sharpener 🙂 Sheer is uncompressed, just smaller files. No quality difference.

Graeme

– http://www.nattress.com – Film Effects and Standards Conversion for FCP

[Graeme Nattress]

I think some more bugs in 10bit got fixed in the new release, but it’s hard to keep track. There were certainly issues back with FCP5. There are still issues with some FXscript commands not functioning correct, but it’s only geeks like me that find them (or Joe).

Graeme

– http://www.nattress.com – Film Effects and Standards Conversion for FCP