Robert, as I probably said before, my suggestion here is based on the assumption that you need transitions and CC. This might not be always the case. Let's say that my *suggested* workflow is viable when working with material for which image quality is of the highest priority. I must also say that I picked up most of it from Stu Maschwitz's "The DV Rebel's Guide", excellent book that I suggest for anyone interested in knowing more about shooting digital. Stu is one of the founders of "The Orphanage" and one of the developers of the original "Magic Bullet".

OK, it's important to understand here that our cameras record in a lossy format. You want to capture the data from the tape and avoid loosing more. Not even a bit. As good as ProRes is, it's another lossy codec. If you capture from the camer *directly* into ProRes you end up with a generation loss. Use Propres as an intermediate codec but be sure that you captured the "raw" data from the camera. That would be the HDV data stream or the .m2t file.

Well, that's kind my point. FCP is an excellent editor, it's not a compositing suite. After Effects, Motion, Shake are the tools for the job. What I mean is that you should use a compositing suite for the very final output, the master of your sequence. Do it at 16 bit precision, 32 if you can, and you'll end up with an excellent result.

thanks for the reply,
so if i capture DVHS Cap, convert to sheer using stream clip, then open a new project in FCP using sheer as the compressor, do a straight edit, then highlight the whole sequence send to Motion for CC text, and Cross Fades.

Then export as a QT File using current settings.

If i want to take the step and use color when would i slip it in to the job.

This then i could open in such appz as Compressor or Bitvice.

I hope i have this right.

rob.

Ok up to here...

This implies that you go back to FCP and render the sequence with it which defies the purpose of using Motion for the high quality processing. Let Motion do the rendering. Export the master to high quality Sheer/Uncompressed and then drop the master into Compressor.
[/QUOTE]

I just had my first shakedown with shake, thanks to this - my most favorite workflow thread yet. I now understand the sense in this. But I have to say, "Noodles, Nodes, Squiggles, Knots," and no more timeline....? Someone is fooling with me right? I am now officially, completely lost, again.

Oh the journey

hey paolo, should I use sheer 1280x720p 24df rgb 10bt or 1280x720p 24df YDbCr as my sequence setting for FCP using hd100 footage?

Well you actually have a timeline in Shake, the nodes and noodles replace the layers or tracks of other systems, like AE. The great advantage of the node system is that it allows you to immediately see the relationship of each node. As with everything, it takes time to get used to it.

Amos, my understanding of the YCbCr is that it's an efficient way of encoding the same values that we have in 8-bit RGB. Given that most of post work will be done in RGB land I usually set the codec to RGB. it's my understanding that Sheer can map use either YCbCr and RGB without data loss.

Paolo

I have captured all ok

I have a test.mov using sheer as the compressor, it plays fine in QTPlayer.

My question is when i open FCP and setup a new seq. I can not see sheer compressor under any of the settings.

How di i set up the seq.

Rob.

Sorry mate found it.

It hepls to read the install nots. LOL.

ROB

Hi Guys,

Hope this doesn't seem like a hijack but if I'm importing into premiere can I import as m2ts files straight from tape? I already have some footage from my firestore these are obviously m2ts but I'm wondering If this is the same from tape, I'm guessing it is. Also have my doubts about Premiere as in another posts I read it clips super blacks and whites. What and how much, in terms of digital info, is this effecting.

Also can anyone advise me of the best possible workflow using Premiere. I also have about 30 hours of footage I need to edit. Would it be advantages when digitising this footage to digitise this with sheer and then import into Premiere or would this be better done post edit?

Greg

Regarding Paolo's post: I have an important project I am working on now and I could use a little advice about many things but pertaining to this thread I am wondering how I can best handle my color correction for this project.

I shot on a HD100 in 720p and captured into and edited on a FCP 5 timeline with HDV 720p presets. This is a project that I thought was dead but it somehow got chosen for a film festival so I need to polish it up and output it at as high of a level that I can manage taking into account my equipment/software, my knowledge/comfort zone and the amount of time I have (about a week.) (BTW, the festival wants it in Beta SP or SX)

I have already done a basic color correction in FCP 5 but after reading your thoughts in this thread I am prepared to redo the CC elsewhere. Problem is I only have access to the FCP coloring or perhaps Motion. For me, the coloring in Motion seems very primitive although it looks like you could get a decent result given enough time, so I would rather use FCP. I don't really have time to learn a new coloring system, but I'd like to get the coloring in a 16 or 32-bit place. I have had some banding issues with my current sequence.

Couldn't I just transcode my HDV sequence to a "better" codec (maybe DVC50 HD? ProRes is not an option--I don't have FCP 6) and color correct in that? Wouldn't that be in a better color space and not have any graphics or color correcting moved down to 8-bit with the whole HDV "rendering" for output? Maybe I'm kind of using the wrong terms here. I admit I don't fully understand colorspaces and what is in 8-bit and what is in 16 or 32.

Also, in the sequence settings couldn't I just check the box for "high-precision YUV" instead of "8-bit YUV" and have this keep everything in 32-Bit? (Keeping in mind that the HDV clips started in 8-bit and won't improve) Can I change this after I have captured the clips and edited the timeline? Would this be just as good as outputting to a 32-bit compositing program?

I must admit some fear of compositors. The only one I have is Motion and I have never used it as anything but a place to make my CG screens and simple GFX. Because I need to get this done and get it right I don't know how many new workflows I can experiment with in the time remaining.

Matt, since you are in neighboring SFO, send me an email to phciccone@gmail.com, let's talk a bit more about this via PM

Thanks
 

Just wanted to add a thanks to Paolo and Steve. You've both been very educational for me. I now better understand why Black Magic was telling me to avoid capturing in HDV....and why there were advantages even after the footage is laid to HDV tape. Thanks!!!

So Paolo, if I'm understanding all this correctly, I should basic edit in hdv in final cut and then export to after effects using sheer and do all the transitions/color grade/compositing/titling in after effects and lastly export from after effects using sheer? And do I need a third party program to go from final cut to AE?

BTW, do you know if apple color is compatible with sheer ?

thanks

Amos, I just finished a quick CC/grading job for a short and that's how it was performed:

- The video was delivered to me as an FCP project shot with the HD100. Sequence was edited in HDV.
- I exported the project from FCP using XML
- Imported the project using the excellent FCPtoAE script, Google it.
- Importing the project means that I don't need to render and that AE is referencing the files from the original HDV clips, no transcoding.
- Set your AE project to use 32 bits processing.
- Do all the CC/Grading
- Precompose all the scenes that need to have transitions like crossfade
- Add transitions.
- Add the comp to your render queue, select TIFF sequence, add an "Output target", select sheer and render

The reason for the last step is to render your final comp to a lossless master, Sheer, that you can then drop into Compressor and use to create your high-quality deliverable. The TIFF sequence allows you to do "surgical" updates so that if the client requires a change in the a scene you can update that scene and then render only the necessary frames without rendering the unchanged parts. After rendering the modifications you can simply import the sequence of TIFFs back into AE, add the soundtrack and the export the video without the need to re-compute all the effects. At the same time the TIFF format guarantees that you'll be working on pristine frames.

Hope this helps.

Paolo,
Is this type of work flow available to people who use Avid Liquid to edit the original HDV?

If Avid Liquid won't work, how about Edius?

What is the editing program on the PC that would work for your work flow to be able to color grade and add effects in AE without first transcoding the edited video to another format (e.g. "raw RGB avi uncompressed" is recommended to me out of Liquid, is this the same as the work flow you outline for FCP?)

Hi Jack.
I don't know either Avid or Edius so I can't give you the steps involved but I would be surprised if there wasn't way a way to move media back and forth those applications. After all editors and compositors are used in different situations all the time. Automatic Duck has a plugin that moves projects from Avid to AE. This would be the optimal solution since you skip the export to any codec altogether. Also, SheerVideo works on Windows as well, it's a QuickTime plugin and the company makes it for both platforms. You can also check Black Magic Design, they have their uncompressed codecs for Quicktime/Windows: http://www.blackmagic-design.com/support/software/

Don't know about avi, being I generally stick with technologies that work across platforms, just to keep my options open. Don't wan to get in a situation where I paint myself in a corner :)

Good luck.

This post has been excellent to me. I found a little more info for premiere pro users.

You can stay in the 32bit realm with premiere and after effects.

I can either do color correction in after effects right from my timeline or import my project into AE. I like and have been using color finesse as a color correction tool in AE.

Based on a previous post by Paolo, I became a beta tester for sheer's .avi codecs. I think Sheer would be a great direct replacement for the black magic codec.

good stuff, thanks

Importing m2t files to Premiere CS3 - wrong frame rate
 

I've been using AspectHD since version 4, and it's a fine code, but since version 5 I've been experiencing som wrong frame rates interpretation, ans it didn't used to happened.

When I import a m2t file into premiere (shot with HD110 24p) it assumes the file as 29,97fps. Does any one is having this problem?

In AE it is not a problem because it is possible to correct te fps interpretation with out messing the audio, but in Premiere if I try to make it assume the file as 23,976fps, the audio gets messy.

When I convet to AVI the files using HDlink the imported file is corectelly read as 24p.

Did you try the CS3 update (3.1.1)? I think it addresses it. I use the JVC presets and CS2 so I haven't had these issues...yet

Back to early in this post... there were comments about converting to AVI as an intermediary. Is there a 1280x720 version of AVI I'm not aware of? Or are you losing your resolution?

For a couple of low value projects I've converted my m2t to standard DV/AVI, but there was noticable resolution loss, visable lines, etc. I'd never do it for something I was giving a client. So I'm a little confused about the method?

Re: Don't you love it when Brainiacs get a little competitive?
 

Yes, I agree, it's awesome. Every bit of info I wanted was in that exchange. I'm with Steven on this one. I see FCP6 in my future as well. Honestly, for what I'm doing the FCP6 workflow is fine. Once I'm done I'll take it into Color and then export to SD. I'll rent an HD Deck if for whatever reason I would have to make an HD master and re-export to whatever is needed. I doubt any of this is as complicated as it's being made considering current advances. I don't think it's about right or wrong way to edit HDV, simply more choices. Your workflow is mainly based on your output (mine is anyway).

Even the Straight to Video market has gone HD. They really don't want anything shot in DV, I actually lost a sell that way to a foreign buyer. That sucks. Honestly I'm sure if you've got enough breast and blood anything will sell but for the basic martial arts action film the market is saturated with such and even that is shot on HD/HDV so they've got a lot of choices so DV is out of the question unless you're self distributing.

As for HD-DVD I knew it wasn't going to make it. It made too much sense. Hollywood and the movie industry is mainly about figuring out the worst possible way to do something and figuring out how to jump on that immediately. Plus keep in mind for a period of time this locks the little guy out of making Blu-Ray disc that can be read on your basic BR player. HD-DVD was easier to code for. I was going to start doing some HD-DVD test but I guess that's down the drain. This is the second time I've cheered for Microsoft over Sony. First was the Xbox 360 versus the PS3.

-Nate

Hi Paolo
 

Here's a workflow I thought of for a feature shot on HDV 720p on a JVC 111E:

1. Capture *.m2t
2. Edit native HDV (only cuts).
3. Once the edit is complete, which will be around 90min of movie, I export uncompressed QT or TIFF.
4. Import the movie in AE, and then finish the transitions, compositing and titling, etc. Export uncompressed TIFF (or QT).
6. Import into a CC software at a professional facility (I don't have a good monitor nor do I have the eye for color). Export uncompressed HD in TIFF (or QT). This is my MASTER, I hope. This is another 300GB of HDD space. I'm adding music and sound with this, Dolby 2.1 Stereo.
7. Encode M2V for DVD.

DVD is my primary target release...but wanted a film-out option as well. I had one basic question though:
You mentioned to transfer the timeline from FCP to AE, but then won't AE be working on the HDV footage? Or should I first render out uncompressed from AE in 16/32 bit and then reimport and do the animations? I'm a little confused here. And what would be the difference in file size for 16 bit and 32 bit? Thanks. More info can be found on the thread I posted on this subject:
http://www.dvinfo.net/conf/showthread.php?t=126419