Captured HDV (Mac) - conversion before editing?

[Najeeb Mirza]

Perhaps we moved too quickly on capturing the 80 hours of HDV onto a RAID via FCP, as I understand that it would have been better to have captured to an uncompressed format. That way we would be able to edit without suffering horrendously long render times. But here we are.

Now that we have already captured, what would be a good solution to convert the material to a format that can be edited without hassle? Ideally, the material would have full frames rather than the 6 GOP of HDV.

Najeeb

[Chris Hocking]

Final Cut Pro can handle HDV Natively. There is no need to convert to another format.

[Nate Weaver]

Originally Posted by Najeeb Mirza

Perhaps we moved too quickly on capturing the 80 hours of HDV onto a RAID via FCP, as I understand that it would have been better to have captured to an uncompressed format. That way we would be able to edit without suffering horrendously long render times.

Um, 80 hours of uncompressed 1080 HD would be 35 terabytes. You got that much?

You should try editing native HDV in FCP. It works fine, and if you do it right, there's no quality hit in your final product.

[David Morgan]

How is it working with the audio or multiple audio tracks when cutting HDV?
Are there any gotchas?

[Drew Curran]

Originally Posted by Chris Hocking

Final Cut Pro can handle HDV Natively. There is no need to convert to another format.

While this is true, rendering times are sloooooowwwwww!! Try exporting your edit for an SD DVD and you'll see what I mean. I even find that realtime effects are greatly reduced when using HDV in FCP.

I capture HDV and then batch export to DVCPRO HD. This can be left over night if you don't want to waste time.
When using DVCPRO HD, realtime effects are greatly improved and final rendering of your edit is greatly reduced. DVCPRO HD also gives you full frames, removing the GOP structure.


Andrew

[Rachel Oliver]

Originally Posted by Drew Curran

While this is true, rendering times are sloooooowwwwww!! Try exporting your edit for an SD DVD and you'll see what I mean. I even find that realtime effects are greatly reduced when using HDV in FCP.

I capture HDV and then batch export to DVCPRO HD. This can be left over night if you don't want to waste time.
When using DVCPRO HD, realtime effects are greatly improved and final rendering of your edit is greatly reduced. DVCPRO HD also gives you full frames, removing the GOP structure.


Andrew

Hi;

Is there any quality hit doing the conversion post capture compared to capturing via say HDMI or SDI to DVCPRO HD? It seems like a good workaround if you don't mind night renders and the slightly reduced image size compared to HDV?

Olly

[Rachel Oliver]

Originally Posted by Drew Curran

While this is true, rendering times are sloooooowwwwww!! Try exporting your edit for an SD DVD and you'll see what I mean. I even find that realtime effects are greatly reduced when using HDV in FCP.

I capture HDV and then batch export to DVCPRO HD. This can be left over night if you don't want to waste time.
When using DVCPRO HD, realtime effects are greatly improved and final rendering of your edit is greatly reduced. DVCPRO HD also gives you full frames, removing the GOP structure.


Andrew

Hi;

Is there any quality hit doing the conversion post capture compared to capturing via say HDMI or SDI to DVCPRO HD? It seems like a good workaround if you don't mind night renders and the slightly reduced image size compared to HDV? Do you use compressor for the conversion?
Cheers

Olly

[David Morgan]

what sort of system are you using?

[Drew Curran]

Originally Posted by Rachel Oliver

Hi;

Is there any quality hit doing the conversion post capture compared to capturing via say HDMI or SDI to DVCPRO HD? It seems like a good workaround if you don't mind night renders and the slightly reduced image size compared to HDV?

Olly

To be honest I don't notice any quality drop. I wasn't aware of an image size difference. The image size remains 1280x720, which is the flavour of HDV my camera produces. I could be wrong.

U are right regarding the added conversion rendering time, however, I would rather have greater speed during editing without having to constantly render the timeline. If you timetable the workflow correctly, by rendering at night, I doesn't become a problem.

Capturing direct to DVCPRO HD using a capture card would obviously save a lot of time.

Andrew

edit: The size changes to 990x720 - so there is a change

[Drew Curran]

Originally Posted by David Morgan

what sort of system are you using?

David,
I'm using a Dual 2.0 G5 with 4gb ram.

For short 5 or 10 minute projects, HDV is fine, but I recently shot a friends wedding and don't fancy what could possibly be a week of rendering required convert HDV to an SD DVD.

Ultimately I would like a capture card from AJA or Blackmagic, but my G5 is the old PCI flavour and most cards are now PCIe.


Andrew

[David Morgan]

thx for the reply. I just got a Canon A1 HDV camera and I'm wadding into the idea of shooting/editing HDV. However, I primarily do live multi-camera theatre or music shows which last anywhere from 1-1/2hr to 3 hrs. I sure can't find myself in a situation where I've got all this HDV footage and run into a nightmare trying to edit it. (using firewire 400/800 capture drives)

Intel Mac macbook pro
Final Cut Suite 5.1

[Chris Hocking]

I don't see why it would take much longer to convert HDV to MPEG-2 than it would for DVCProHD to MPEG-2, especially on a new Intel-based machine. I would have thought it would take longer to convert HDV to DVCProHD and then to MPEG-2, than it would from HDV straight to MPEG-2 for DVD. I've edited long HDV timelines before on an eMac without much of a hassle. Sure you have to render some of the filters/transitions, but it's ALMOST as fast as DV. And as Nate said, if you do it right, you won't loose any quality. Converting the timeline to another format (i.e. from DV to MPEG-2 or HDV to MPEG-2) is always time consuming, but I would have thought (and I could be wrong on this one) that as HDV is MPEG-2 already, the conversion process would be a lot easier on the Mac as it only has to transcode from one variety of MPEG-2 to another. However, the resolution is a lot bigger on HDV than it is on DV, thus more pixels, and therefore, an obvious increase in render time. However, that taken into consideration, I can't imagine there would be that much difference in time in comparison.

Related Article: http://larryjordan.biz/articles/bk_hdv_convert.html

[Drew Curran]

Chris

All I can say is do some tests and see for yourself. You will see a big difference.

HDV has a 6 or 12 GOP structure, which is very different from SD Mpeg2, which has individual frames. The GOP structure of HDV makes it very processor heavy. So during conversion of HDV to Mpeg2, the GOP has to be removed and significant scaling has to take place. This is my reading of the process.

DVCPRO HD is also individual frame and removes the GOP structure. Its relatively quick to convert from HDV to DVCPRO HD as less scaling has to take place.

As I said above, editing with DVCPRO HD is a breeze compared to working with HDV. The RT effects in FCP increase when using DVCPRO HD = less timeline rendering.

Go to the Panassonic HVX200 forum and they will shout the advantages of DVCPRO HD over HDV to anyone who'll listen.

Andrew

[Evan Donn]

Originally Posted by Drew Curran

Chris
HDV has a 6 or 12 GOP structure, which is very different from SD Mpeg2, which has individual frames. The GOP structure of HDV makes it very processor heavy. So during conversion of HDV to Mpeg2, the GOP has to be removed and significant scaling has to take place. This is my reading of the process.

SD MPEG2 does not have individual frames, however it also doesn't have an enforced, consistent GOP like HDV does - I believe it just has a maximum of 18 frames.

Conversion from HDV to any other format will require reconstruction of the individual frames from the GOP, so this should take the same amount of time no matter what format you are subsequently re-encoding to. If you encode to another intraframe codec like MPEG2 (or back to HDV) the overall process is going to take significantly longer than if you convert to an interframe-only codec like DVCProHD. However - if you then encode your DVCProHD to MPEG2 you'll take the time hit there.

Even simple playback requires full reconstruction of the frames from the GOP, which incurs some processor overhead and therefore will reduce the resources you have available for things like real time processing - that's why you'll see better realtime performance working with DVCPro, in addition to a small increase due to the lower res (~12% less pixels for 1080, ~25% less for 720).

The biggest rendering hit I've experienced with HDV is in rendering back to HDV for a print to tape. I just don't master anything on tape anymore. There seems to also be a huge performance hit going directly from your timeline to compressor though, regardless of the format, so I generally export>quicktime movie to another preset such as AIC or DVCProHD and then use that as a master in compressor or MPEGStreamclip.

[Drew Curran]

Originally Posted by Evan Donn

However - if you then encode your DVCProHD to MPEG2 you'll take the time hit there.

Of course, but I've noticed this process to be a lot qicker than converting HDV to Mpeg2. The time comparisons speak for themselves.

I've now abandoned the HDV editing process completely in FCP is because of the known timecode/dropout problem when capturing from the JVC HD100. I now use the DVHS-cap/MPeg Streamclip method, converting directly to DVCPRO HD.


Andrew