step-by-step workflow for best results?
 

To anyone of you JVC HD100/HD200/ HD250 users:

Can anyone tell me what the best procedure/workflow would be to get the best color, and clearest image from these cameras in post, especifically when shooting 24fps? Someone please post a step-by-step outline or chart. Thank you.

I mentioned in the other thread about hdv editing but here it is, using FCP:

1) Acquire your footage using either your NLE or a combo like DVHSCap + MPEG Streamclip.
2) Edit your footage in the NLE. Avoid any CC, transitions and effects. Use them if you must but be prepared to re-make them later in your compositing software. The reason for this in a moment.
3) Once you locked the edit duplicate the sequence and remove all the crossfades, transitions effects etc.
3) Export your sequence using XML
4) Import the sequence in AE using either Automatic Duck or http://www.creative-workflow-hacks.c...ut-the-hassle/ The latter is free and works perfectly well for me.
5) Switch your AE project to 16 or 32 bit and use either Levels/Curves/Hue-Saturation/Colorista/Color Finesse to do your color correction.
6) Pre-Comp your effects and then use the Pre-Comps for the transition. This ensures that the CC is applied *before* the transition and not the other way around as it would normally.

The detailed process is described in Stu Maschwitz's "The DV's Rebel Guide", which I cannot recommend enough ( http://www.amazon.com/dp/0321413644?...NFTEB6SMTA4ZN& ) Stu is the cofounder of "The Orphanage" and a post-work wizard. He kindly shares his years of expertise in the industry working on high profile movies like "Sin City" in this book. The software included in the guide is worth the cover price... several times over.
If you are interested in obtaining the highest quality from your HDV images this book is a must.

Oh, Paolo :-) I knew you'd come to the rescue. Tantissime grazie per essere cosi attento.
I use a 35mm adapter, so as you know, the image is always gonna be somewhat soft. Now, I'm about to shoot a video and I worry about getting the best image possible in post. You're the authority when it comes to these cameras, so I'll follow your instructions faithfully. Thank you again.

Is there no need to use FCS2 "color"?
 

Paolo, so if you're working in FCS2, do you just ignore "Color" completely in favor of AE?

Non c'e' di che' :)

That's OK. You probably know that I advocate to avoid the sharpening in camera the things that is misleadingly labeled "detail". That's why TrueColor sets Detail to MIN. I was delighted to find in the "DVRG"" that Stu confirmed the approach. If you shoot with a 35mm adapter I assume that you will have post-work done in order to dial in a certain look and add other effects. As a last step you just add sharpening using AE's filter (Sharpen) which leads better results than the in-camera circuit and can be adjusted scene-by-scene, depending on the desired look and feel.

Regarding FCS2, I just find both Colorista and AE much simpler to deal with. Color is a great application but is very heavy, less immediate than Colorista or Levels in After Effects and you can get the same results with those tools. BTW, you should check out, if you haven't done it already, the new Magic Bullet Looks. This replaces the old MB Editors and the new interface is just amazing. The level of control over exposure, gradients etc is fantastic and Red Giant is now releasing most of its plugins multi-hosted, they work both in FCP and AE so you can use the same interface regardless on the host that you use. In this way you can apply a quick look and CC effect to your FCP footage if you don't need the extra control of AE and you'll be using exactly the same controls in AE when you need to go the extra mile for the more important projects.
I found that the applying multiple instances of Colorista, stacked on top of each other, is a very effective, fast way of doing high-quality color correction. To me it's intuitive, simple, accurate and the use of "Power Masks", in DaVinci they are called "Power Windows", gives the amount of control for selective CC without requiring to duplicate a layer just to apply a mask. This is a huge time saver. The other advantage, IMHO, is that by using Colorista you stay in the host application that you know well. Color has a completely different interface which you need to learn. I was able to use Colorista in minutes, which is really, really cool, IMO.

Wow, what priceless tips. Thanks again, Paolo.

Paolo, please forgive me for asking what may end up being a naive question (I'm a broadcast cameraman of many years new to home editing using small HDV cameras.....I'm still learning!!!!!! It's great fun!!).....but why would you switch to 16 or 32 bit for your compositing etc when your acquisition is only in 8 bit such as GYHD cameras?

Also, do you switch to AE because you find the compositing and after effects tools in FCP are simply not up to the job? I went down the Vegas route for financial reasons and find the output post compositing and effects quite satisfactory, although colours can be a bit lacking in saturation at times. Would you have any recommendations for those who use Vegas (I do read the Vegas room but can't remember reading anything on this yet)...does AE work with Vegas for example?

Many thanks from an old dog learning new tricks!
Stuart

I'm glad you asked :) Color Correction and many effects can cause quite a shift in your original bits. We shoot at 8 bits but then when you apply color correction, especially when you use primary, secondary and tertiary CC, you can end up with clipping, which, as we know, it's irreversible loss of data.
Now, when you use 16 or 32 bit CC in After Effects the program preserves the information as the effects and colors are applied with higher precision. Believe it or not, when you render back to 8 bit your highlights are preserved and the final image will render at much higher fidelity than when you apply the effect at 8 bits. What is important to remember here is that the *processing* has to happen at high precision, the final result can still be at 8 bits per channel.

But there is an additional benefit, if you look at how fades are performed in film you'll see that they look very different from your typical fade to black in FCP, Vegas or other NLEs. Same thing if you use AE at 8 bits. The difference is that 8bit processing will lower all the levels in a linear way, turning the highlights grey. Now look at your typical Hollywood blockbuster and you'll see that the highlights fade away while staying bright, much, much brighter than midtones. When they finally fade away they are the only thing left visible. When you use 8 bit processing you leave a telltale signature that you used video :) Well, not anymore. Do your color correction at 32 bits and you'll have beautiful highlights which will fade in a way that mimics the closing of the iris of your camera.

I included a couple of photos to show the difference. They are exactly the same frame of a candle which was captured in HDV, XDCAM actually but still 8-bits with 4:2:0 compression, to which I added the effect TrapCode Shine to create the shimmering rays of light. The frame shows a fade-in from complete black, and it's at about 50% of the luminance. The difference between 8 bits and 32 is pretty dramatic! And keep in mind that you are looking at 8-bits-per-channel TIFF files, as a proof that you can acquire at 8 bits, store at 8 bits and have great looking images if you process at higher precision.

The 32 bit version: http://www.paolociccone.com/images/32bit/Candle32.tif
The 8 bit version: http://www.paolociccone.com/images/32bit/Candle8.tif

Please note that in this case the 16bit version is the same of the 8 bit one. 16 bits is better than 8 but it doesn't go as far as sometimes necessary. Nevertheless it's a big improvement from 8 so I suggest to use it as you default precision since 32 bit can be quite processor heavy. Anyway, once you know what to look for, you know when it's time to switch top 32-bit processing.


That's right. Conversely AE cutting tools are not very sophisticated and should be use only for the simples of tasks. An editor works the best at one thing: cutting. And it needs to, editing is where pacing and story-building happen. You want to be nimble and able to rethink your story and your perspective easily. FCP does an great job at that. It's just not a compositing tool. Apple has two other programs for that: Motion and Shake. I personally prefer After Effects.

I don't know vegas so I can't give you advice on how to export your Vegas projects to use in AE. There might be already solutions in place. Once thing that you might want to try is to export your Vegas project as a long clip, using a lossless codec, load the clip inside After Effects and use the free Magnum script ( http://aescripts.com/2008/01/18/magn...edit-detector/ ) to find the edit points automatically. BTW, I think that this script is just incredible and a demonstration of the power of AE. It has several uses and it can act as a kind of universal importer.

Good luck!

Paolo - first of all thanks so much for being such an incredible help to the people on this forum. I use the V3 preset all the time with my HD100 and it looks great.

Concerning Vegas, the latest version (Vegas 8) has internal 32 bit processing and also handles native m2t files very well. I'm just wondering if, for my Vegas workflow, it makes more sense to apply color correction, etc to the native m2t file, then render to the final file (mpeg-2, h264, WMV, Uncompressed, etc) using the 32 bit processing.

I guess I don't see the difference in rendering a "final file" from the original m2t file or from an uncompressed render of the m2t file in this case. Would love to hear your input and thanks again.

Hi Burk.
You're very welcome.
I don't know Vegas so my answer cannot be too complete concerning this. What I do know is that FCP has 32bit processing and it still cannot get to the level of quality that AE achieves. This is actually expected and not a disappointment, it's part of what these tools do. In fact the inclusion of Color is pretty much an admission that you need to step outside in order to get higher results.

I guess that the result is based on not just the numeric precision but also on the quality of the effects. The best answer is to do some tests of your own. Following the example that I gave with the candle you can simply take a clip with bright highlights and drop it in your timeline in Vegas. Add the same clip on top but now use the composite mode of "Add" or "Screen", I assume that Vegas has something like that. Do it another time and now compare the result when switching between 8 bits and 32 bits. The 32 bits version should not be clipping. If it does then you know that you need to use something more precise.
Once again, it depends on the type of job. For interviews and such it might be overkill to follow the workflow that I described. For anything like music videos and shorts or commercials or features I think that it's best to leave the high quality work to a specialized tool like After Effects. Note that I mention AE a lot because it's quite affordable and it has a strong following in the independent arena. There are other tools but they are much, much more expensive. AE is available for both Mac and PC, it has s ton of very high quality plugins and it's very flexible.

Just to make it totally clear, I have no connection with Adobe, I just happen to like their tools a lot :)

The discriminating factor in your choice of workflow is to know if you will need at some point to use a compositing program. For a lot of work that I do I end up in AE anyway because the masking and keyframing tools in any NLE are generally primitive at best. If you add things like motion path, graph editor, 3D space and effects like fractal noise and expressions, there is just too much power there to try to emulate those features in the NLE. Also effects and transitions are not as high quality. If I need to do Chromakeying there are just too many factors that require manipulation of the alpha channel before and after the keyer that an NLE is not suited for. AE is the natural habitat for that. So, since I know that I have to land there anyway, I just avoid doing CC in the NLE. If you tend to "orbit" around Vegas exclusively than you have to adjust your workflow accordingly.

Hope this helps.

Paolo, thank you so much for a fantastic explanation. The visual example was quite staggering! That's really helped. So, if I have this correct, keep your video at 8bit and any additional work, such as effects and CC should be done at 32 to keep the 8bit footage looking as good as it was when recorded!?

So, would you use AE for all your transitions too? such as dissolves? I'm now interested in this as we've just completed 2 high end corporate films and have another coming up soon, and I'd really like to increase the quality (or not lose it!). I'll be looking into the possibility of importing the EDLS from Vegas into AE for example. I find this fascinating so I'm really keen to learn more on this!

As for Vegas 8. It has 'floating 32 bit' ,which, as far as I understand isn't true 32 bit processing. Is that correct?

Thanks again
Stuart

It is true 32-bit math processing.

Yeah, I was quite impressed by the visual difference myself. The bit depth to use depends on your hardware. Since AE doesn't set this parameter by comp but by project, it's a all-or-nothing situation. Fortunately changing it is very simple so it can be adjusted at will. If you machine is not very fast I would suggest to set your bit depth at 16 bits to give you more precision in normal situations. If your comp has more extreme uses of effects, like in the case that I demonstrated, switch to 32 bits.

Yes, if the job must be realized at high quality, I perform all the transitions in AE. Besides the bit precision there is another reason. If you do your CC and you need to crossfade from two clips that use different looks, a plausible case since crossfade usually signal passage of time, doing the transition in AE will guarantee that you apply the CC first and the transition after.

32 bit is floating point. Colors channels are expressed in infinite values from 0 to 1. If you switch to 32 bits in AE and open the Info panel and move the mouse over your image you'll see that the colors are expressed as a floating point number. Where it gets interesting is that you can set colors to go past the value of 1.0, creating areas of your footage that have two or three time the maximum value allowed. 32 bit processing will still preserve those values, no clipping will happen and you will be able to render those clips at 8 bits and still obtain a good image.

Very helpful thread. I've been needing something like this for a while now, desperately - apparently! Thanks for offering up so much of your time Paolo.

It's always the little things that get me though. Like, what is "XML" and how do you export using it?

Dumb enough question? Cause I've got more where this one came from. But I definitely know where the on/off switch is..

Hey Eric.
XML is the Xtensible Markup Language, it's an open-ended file format to describe any kind of storage using a simple text file. The language is very similar to HTML but instead of defining a specific type of "grammar", the type of tags used in it, XML allows you to define any set of tags in order to describe your type of data.
For example, in HTML you have things like <table>...</table> to describe a table or <p>...</p> to introduce a paragraph. In XML you can describe the entire FCP sequence in textual mode. Here is an excerpt:

<clipitem id="SC02-T03-C2 2">
<name>SC02-T03-C2</name>
<duration>3200</duration>
<rate>
<ntsc>TRUE</ntsc>
<timebase>24</timebase>
</rate>
<in>811</in>
<out>887</out>
<start>1474</start>
<end>1550</end>
</clipitem>

Even by naked eye we can pretty much make sense of it.
By using XML FCP opens up a lot of possibilities because now you can export a sequence and another program can read the XML sequence and find the in and out points of every clip and recreate the sequence inside the other program, After Effects for example, without having to re-render your footage. It's also incredibly faster as creating XML is very quick and even long sequences will be export if a few seconds.

To use it just go in the menu File/Export, it's the last option in the menu. Export the sequence, close FCP, open AE, use the FCPtoAE script, point at the same sequence file, voila' , your FCP sequence is now an AE project!