Multicamera footage: PMW-EX1 with HVR-Z1

[Steve Sykes]

I would like some advice about the best settings to create primarily a SD DVD using footage from both the Sony PMW-EX1 and the HVR-Z1. I am recording a 2 camera mix of a dance production under stage lighting. Ideally I would like to offer an HD alternative but but the SD DVD will be the priority. I am hoping to borrow the Z1 as my 2nd camera.

Using my PMW-EX1 has my only HD reference I have found that recording 720p 50fps has been the easiest and most flexable to work with allowing slo-mo if needed, smooth footage for action / fast dance moves and I have better results converting this to SD than 1080p footage!

I am unsure if the EX1 footage using 720p 50fps is compatible with the HDV footage from the Z1 at 1080i. The problems I can forsee if I try to transcode the Z1 footage to match the EX1 are:

1) the Z1 is interlaced (I've heard this is problematic when resizing & mixing with progressive footage)
2) the Z1 has a different a different PAR (rectangular) - again resizing interlaced footage!


Has any one resized 1080x1440 interlaced to 720x1280 without having any problems? (I would be using Virtual dub to do this)

If I successfully convert the interlaced footage to 720x1280 what happens if I place it on a 720p 50fps timeline in my NLE? Does each field now represent a frame? When I render my footage from that timeline as interlaced will the 50 frames of progressive be converted to the 50 fields of interlaced?

Alternatively I could:

Record both in HDV (1080i) - will the PAR change from square on the EX1 if I do this or do I need to resize the PAR of the Z1 footage to square?

As you can see I have many questions and concerns. How practical is this!? Any advice welcome, I have two weeks to decide what to do!

[Matt Davis]

A good friend of mine (Rick Young - macvideo.tv) regularly shoots EX1, Z7 and Sony A1 and then multi-cam edits them in FCP. HDV at 1080i being the common one format between them.

In other words, EX1 shooting at 25Mbits at 1080i is, to all intents and purposes, the same sort of technical file as the Z1. You'd shoot and edit as HDV. Don't sweat the aspect ratios of the pixels, this will be taken care of.

Before I switched to EX1, I shot HDV in my Z1s, edited using AIC at first, then HDV with ProRes, then used Compressor or DVFilm to deinterlace and then scale down to 1280x720 so I could deliver 'HD' WMVs for use in PowerPoint. Quite frankly, it was a bit of a pain then, let alone face the idea of preprocessing all your Z1 footage before you edit. Don't get me wrong - 720p25 and 720p50 are my favourite formats... for SD/web on the EX.

There's a school of thought which is along the lines of 'don't bugger around with highly compressed rushes' - I'd be tempted to stick with 1080i throughout shoot and edit, then (if you're an FCP user) use Compressor with Frame Controls enabled to deinterlace, scale down and reinterlace for SD DVD.

I'm also on a Damascus Road regarding interlacing with regards to HD. I'm not quite converted yet, but I freely accept the point that deinterlaced 1080i may lose 25% of its vertical resolution, but it's still more than 720p, AND you're still working with better horizontal resolution.

The biggest issue is going to be with the Z1 image quality compared to the EX1 - I'd enable black stretch of course, and run tests to see if pulling up mids on underexposed Z1 footage is going to match the EX1 better than adding gain (and noise). I guess you're just using the Z1 as a wide?

[Steve Sykes]

Hi Matt,

I read on the other thread that you commented that you would prefer to convert to 720p then edit. Have you changed your mind?

Originally Posted by Matt Daviss

But I would avoid interlaced right now. I'd prefer to downconvert 1080i to 720p rather than edit 1080i. Interlacing only works on CRT screens, and most of our work going forward will be on natively progressive screens, so deinterlacing will either remove 25% of the resolution or (on cheap screens) cut resolution in half. Native progressive is the way to go, period.

I have access to Sony Vegas and use virtual dub to resize. I have not tried deinterlacing. Can you you suggest a workflow that I can use? It is worth paying for a program to do a high quality deinterlace or is there a free one can use? In my thought process the footage is only going to be deinterlaced so it can be resized then it will be interlaced again. What is the best way of doing this so I dont loose the original fields!?

[Matt Davis]

[QUOTE=Steve Sykes;1022283]I read on the other thread that you commented that you would prefer to convert to 720p then edit. Have you changed your mind?

No - but I no longer get my underwear in a twist over the concept of people sticking to the lowest common denominator.

I personally will stick to progressive, and try not to use my Z1s on EX shoots - I prefer to stick to a brace of Z1s or a brace of EX cameras. I may even sell my Z1s to fund either an additional EX1 or an EX3.

But Rick is in a position similar to yours (no choice but to mix them), and he assures me his workflow is fine. Considering the volume of work he turns over in this format, I feel safe on passing that on.

Originally Posted by Steve Sykes

I have access to Sony Vegas and use virtual dub to resize. I have not tried deinterlacing. Can you you suggest a workflow that I can use? It is worth paying for a program to do a high quality deinterlace or is there a free one can use? In my thought process the footage is only going to be deinterlaced so it can be resized then it will be interlaced again. What is the best way of doing this so I dont loose the original fields!?

I've got to be careful here, as I've heard good reports of Virtual Dub, but I'm a Mac person running FCP.

To preserve the interlace, you would deinterlace to a format that could do double your frame rate, then scale down and reinterlace back to your frame rate. Sounds messy, but the sort of thing you could cook up in compressor and ProRes, and quite possibly in Virtual Dub and HuffYUV or some such. The Z1's HDV-DV conversion does something along these lines.

If you're not using FCP and multicam to cut stuff together (which requires all sources to be in exactly the same format), and that you're filming a live performance (so there's not going to be much more rushes than your final programme, per camera), then the other method is to downconvert your HDV to 720p50 - but not sure how you'd do the deinterlace to a double frame rate - I'd do it in a custom project in FCP, or cheat like buggery in Adobe After Effects, or Motion or something.

So many ways to skin the proverbial. Want to get wierd? How's this. If you filmed and and edited in HDV, you could even lay off to tape in HDV format into your Z1, then re-ingest DV using its downconvert. Use that as your master for DVD. There would be some raised eyebrows about this, but as McGuyver solutions go, it would work.

[Peter Mee]

Out of interest, why would one wish to deinterlace before scaling only to re-interlace afterwards?

Thanks

Peter

[Steve Sykes]

Matt, thanks for your advice. Its nice to know that someone is mixing footage with the EX1 and the Z1 successfully I would ideally like two EX1s!

I will have the Z1 this weekend and will try both methods:

1) As your friend Rick does - shoot both cameras in 1080i, edit in HDV and then resize. I am still concerned that the EX1 and the Z1's version have different PARs!? I may have to convert the Z1s footage to square pixels first???

The other concern I have is HDV is upper field first but DV is lower field first. Do I just keep selecting upper field first in all my options! Do you know how Rick converts from HDV to SD?

2) The second method I am tempted to try is mixing the EX1 footage at 720p50 with the 1080i from the Z1. I would have to resize the 1080i footage to 720i and then add it onto my 720p50 timeline. I assume this is where I would have to deinterlace my footage for Vegas to treat each field as a frame or will it do this without me deinterlacing?

Peter the main reason I think I would have to deinterlace the footage was if I use method 2 above so my timeline in vegas treats each field as a frame. I'm probably going to use method 1 as suggested as I don't think I need to deinterlace at any time. The other thought in the back of my head, but I think I'm wrong now, is that interlaced footage does not resize as well as progressive, if this is true then method 2 may be better!

Can anyone confirm that resizing interlaced footage looks worse than resizing progressive footage?