Use HDVSplit (freeware) that does use scene detection for capture. Then use EDL to mimic your DV edits on the HDV material.
We tried HDVSplit for capture, but it didn't seem to handle the scene detection too well (ended up with way too many seperate files compared to the capture from within Premiere). Apparently it handles an already captured single MPG (result of an HDV capture from Premiere) fine, so I will try that.
If it does split the source into matching HDV parts is the process of using them;
1. Create a new Premiere Project (at HDV res)
2. Import the split HDV quality files into Premiere
3. Import the EDL file into Premiere
As long as the HDV sized parts have the same filenames and (hopefully beginning and end points) then Premiere should output the same movie, but at the higher resolution?
We tried HDVSplit for capture, but it didn't seem to handle the scene detection too well (ended up with way too many seperate files compared to the capture from within Premiere).
Not true. It handles scene detection perfectly. Many more files is logical, since PR only captures a tape at a time. If you are comparing the number of files versus DV with scene detection, yeah, well it is different material, one is DV and the other is MPEG2, it is like comparing melons and grapes. One kilo of each will give different numbers.
Why is the number of separate files an issue? You want scene detection, don't you?
If your file names from the DV capture were with date and timestamp, the same way HDVSplit does by default, you should be OK. If you used another naming convention, try to replicate that in HDVSplit.
OK, so there will be a different number of extracted HDV scenes compared to DV scenes.
Will this matter to Premiere? Once I start a new project, import all the HDV extracted scene files and then import the timeline (EDL) is Premiere going to be smart enough to realise the different files relate to the existing timeline and edit points?
Initial HDVSplit results were along the lines of 150 seperate MP2 files compared to the 40 or so from Premiere DV Capture.
So if HDVSplit captures the HDV format scenes, but creates many more separate files, how does Premiere know which one releates to which scene on the timeline?
I assumed the split detection works the same for both DV and HDV (splits the tape at the moments when the filming/recording was stopped/started).
Sorry if all this is so low level. This is the first time we have used Premiere for any high end stuff. Next time we start a new Project in Premiere it will be using HDV source from the start. Our mistake was to capture at DV and then spend days on the edit. If there is anyway to replace the DV parts with HDV without having to re-edit it will be so much easier.
One thing that may cause this difference in number of files between DV and HDV capture with scene detection is that PR DV capture may have had to use optical scence detection instead of date/timestamp, for the reason that the camera when downconverting to DV may not include the date/timestamp information. This is only a guess since I never tried it, but it seems the only logical explanation for the difference in number of scenes.
If this is indeed what has happened, then you can as well start from scratch.
When using two cameras, one HDV and one DV and capturing with HDVSplit for the HDV and Scenalyzer for the DV and using the same naming convention, it was easy to mix them, because the date/timestamp info gave accurate scene detection with both applications. I have never tried your workflow.
If you are comparing the number of files versus DV with scene detection, yeah, well it is different material, one is DV and the other is MPEG2, it is like comparing melons and grapes. One kilo of each will give different numbers.
Proper scene detection should produce the same number of files for any media type. Every time the camera was stopped is a new scene. There is no other proper determinant for creating a new clip. If you started and stopped the camera 10 times during recording, then you should end up with 10 clips after capture using scene detect. Any other number of clips is not a proper functioning of scene detect, regardless of the media format.
Proper scene detection should produce the same number of files for any media type.
Incorrect. With 50 scenes DV captures 50 AVI files. With HDV with the same 50 scenes, you have 100 files, 50 M2T and 50 XMP files.
Thanks for all the tips guys. Turns out it can be done through CS4.
You run a batch capture using the EDL file. This then prompts for the tapes again and imports at HDV resolution.
We did get a few failed imports, but the time to fix them is minor compared to having to re-edit everything.
Hopefully this helps someone else with this issue in the future. But from now on we will definately start by capturing at the full resolution using HDVSplit first and then use Premeiere for the edit process.
you have 100 files, 50 M2T and 50 XMP files.
OK then. You should end up with the same number of clips.