<<<--
Transfer the MPEG2TS file directly from the camera to a PC,
Demux it on the PC using a shareware utility called "mpgtx"
Move the file to a Mac, uncompress it in Quicktime Pro,
Make a DV clone from Quicktime Pro.
Edit in DV,
Conform to HD on your regular DV hard drives,
Move the file back to the PC
Encode to MPEG2 (HD res) using TMPEGEnc utility ($50)
Convert to MPEG2TS using the Womble MPEG2 editor ($120)
View the HD movie on your PC or Mac using the Elecard player (PC) or VLAN
shareware player on the Mac
or transfer it to a D_VHS deck or back to the camera and view it on an HD
TV.-->>>

whew... I almost fainted after reading that, this just goes to show the cam is inadequate for any serious long length work.So you have to render/encode 4 times during the entire process?

Just a reminder guys, though Robert is right in saying that miniDV looks like crap on a monitor it applies to mac DV codec wich was design as an accelerated desktop viewed codec. Most PCs will not have that problem and radiusDV on the mac is much better. It is ONLY a display thing and cannot be used as a reference (if you export the frame you will notice a very visible gain either on picture or on another codec like animation...)

Eric

I don't think this editing chain has much to do with the camera. This seems to be a workaround that avoids having to build a system that can deal with realtime HD editing. It's actually a pretty clever workaround, IMO, if slightly daunting in its complexity.

Ideally, though, it seems like (and if anyone else has a take on this I'd love to hear it) the way to deal with this format would be to remember that the MPEG2TS file is compressed the way it is so that it will fit on Mini-DV. Once you get the data out of the camera it should be converted to an uncompressed format and your master should remain uncompressed from there on out. Editing uncompressed HD, though, requires a lot of processor power and really fast hard drive access. I tend to imagine that this format will prompt a lot of people to move up to G5 systems with X-Serve RAID storage.

<<<-- Originally posted by Yang Wen :
whew... I almost fainted after reading that, this just goes to show the cam is inadequate for any serious long length work.So you have to render/encode 4 times during the entire process? Please say you're joking. This is crap. -->>>

Doesn't JVC's proprietary editor that ships with the cam edit in the MPEG2TS format?

<<<-- Originally posted by Yang Wen : Doesn't JVC's proprietary editor that ships with the cam edit in the MPEG2TS format? -->>>

It's supposed to, but how eager are you to edit MPEG2TS? I mean, it's the easiest way, I suppose, but I don't see it as being a very useful format for anything but acquisition.

<<<-- Originally posted by Robert Jackson : <<<-- Originally posted by Yang Wen : Doesn't JVC's proprietary editor that ships with the cam edit in the MPEG2TS format? -->>>

It's supposed to, but how eager are you to edit MPEG2TS? I mean, it's the easiest way, I suppose, but I don't see it as being a very useful format for anything but acquisition. -->>>

I couldn't even load it up on my roommates 2 ghz Windows XP Wintel...Paul had the same problems, but figured it out, so I'll try the same, too.

heath

Vegas in its current version will read native MPEG HD transport streams straight from the camera (at least the clips that have been posted here) and edit them normally. From that you can convert to normal MPEG2, uncompressed frames, or anything else you might want.

I expect FCP will eventually handle TS. Until then you simply demux it to MP2 and M2V.

There's nothing strange about TS. Inside its just MPEG-2 and MP2 streams.

And it can be edited with frame accuracy.

umm as far as I know, you can't get frame accuracy with any MPEG scheme, the whole idea behind it revolves around using keyframes at regular intervals that helsp to interpolate between each frame. Individual frames are not saved into the file.

Well, you could always use the software to convert to uncompressed if you really need perfect frame accuracy. I mean, I guess it won't be "perfect" but it'll be close enough.

Frame accuracy is quite possible, they just calculate the inter frames on the fly, it's CPU intensive but that's all. I believe the Womble MPEG2 editor is frame accurate.

"That's all"? that's a pretty bizarre defense for this problem. The fact is, with the MPEG2TS used on the cam, you can only cut to accurate down to 6 frames. Which is BADD

MPEG2 compression does not implie that frames are not accurate, you se all the frames on TV don't you? It just means that not all the frames are individually compressed but each frame do exist whatever the compression method. Once uncompressed or even using a MPEG2 utility, you can always have frame accuracy. The only thing is that each frame is not necessarily complete by itself in a MPEG2 stream and is dependant on others (keyframes). As long as you keep those keyframe, editing will be frame accurate. Looking at MPEG2 footage (on a DVD for instance), you can choose to advance frame by frame at any time and not just for each GOP (group of picture) of the stream. Each frame is there, even considering it's dependencies.

Eric

I suggest you read up on the MPEG2TS format used by this cam. It is "not" frame accurate.

Yang, Sony would not have chosen MPEG2 as a compression method for its proffessional cameras if it were not possible to frame accurately edit it. I have already frame accurately edited footage from this camera. I think that's about al that needs to be said on this subject.