Gearshift and/or Connect HD necessary for laptop, and which HDV capture program?

[Alex Thames]

So far, I have been editing HDV footage down converted to SD on my laptop using Vegas 6.0d, but I'd really like to edit HDV footage, otherwise, what's the point in my camera.

First, what are the ways to capture HDV footage? I've heard of HDV Split, CapDVHS, Virtual Dub, and Vegas itself seems to have something to capture HDV? Is this correct? (Not sure if Virtual Dub can actually capture HDV footage, but I believe the rest can). So, which program is the best to capture HDV footage from my Sony HVR-A1U, and why? And what function does Virtual Dub have?

Second, is it completely necessary for me to purchase Gearshift and/or Connect HD in order to edit HDV on my laptop, an ASUS M5 laptop with:
Intel Pentium M Processor 1.6ghz
797MHz cache
752mb RAM (non-dedicated)
Microsoft Windows XP Home Edition
Version 2002
Service Pack 2
Display adaptor/Graphic card: Mobile Intel (R) 915GM/GMS,910GML Express Chipset Family (Intel integreted gfx card)
Audio card: Realtek High Definition Audio

And if so, do I need just Gearshift or both Gearshift and Connect HD, and why (or, what do each of the software do?) I've read here that Gearshift is for lower-powered computers such as my laptop, but does it actually provide intermediary codecs like Connect HD does? I've also read that Connect HD is for medium to higher powered computers, and that it produces intemediary codecs. Is it necessary to have/use intemediary codecs, or is it extremely helpful?

Can my laptop even handle Connect HD? And if not, and if Gearshift is all I need, then how do I deal with the issue of not having intermediary codecs?

Does a faster, say, desktop powerhouse computer need both Gearshift and Connect HD, or is just Connect HD fine?

[Gian Pablo Villamil]

I'm kind of going through the same discovery process myself. In terms of capture, this is what I've found:

1) HDVSplit splits HDV into scenes based on timecode as it captures. It generates standard MPEG2 transport streams (MPEG2 TS, M2T), directly usable in Vegas and playable with most MPEG2 decoders. It has very good camera control, and seems to be totally reliable, with zero dropped frames in my experience. It is what I use.
2) CapDVHS does a straight capture. Only real advantage over HDVSplit is that it can generate regular MPEG2 files (Not TS), which can be opened directly in a few more editing programs.
3) VirtualDub does not capture HD from MPEG2 devices.
4) Vegas has a rudimentary capture utility. Doesn't split scenes as far as I can tell. Captures to MPEG2 TS.

Gearshift and Connect HD play roughly similar roles in workflow - both of them give you a way to edit HD without using the MPEG2 TS files, which are awkward to work with.

To use Gearshift, you first capture using any of the options above, and build a project with the M2T clips. Gearshift then generates low resolution proxies in DV (and optionally, in Sony YUV or Cineform HD intermediate) and lets you switch quickly among them. So you would do all your editing on the low res proxies, then switch back to the original M2T clips right before rendering.

You basically end up using double (or more) disk space, but are able to edit in realtime. For a laptop user, you could just copy the low-res proxies to the laptop and leave the HD files on an external drive. Note that Vegas includes two digital intermediate (lossless) codecs, Sony YUV and Cineform HD.

ConnectHD includes a capture utility, so it captures from the camera straight to the Cineform HD intermediate codec. You then do all the work with files in that format. (You don't ever use M2T files, hence saving a big chunk of disk space). Cineform HD is very quick and easy to work with, at full HD resolution. I understand that Cineform HD files are 4x larger than the corresponding M2T, and while easier to edit than M2T, still require a hefty system.

Here's some numbers: 1 hour of HDV in M2T is about 11GB, 1 hour of DV is 11GB, 1 hour of HD in Cineform is about 40GB. So if you were using Gearshift, you'd need the M2T and the DV proxy, for total of 22GB. You could go on the road with 11GB, leaving the HD files at home and swapping them in only for rendering when you return home. If you were using Connect HD, you'd be capturing straight into the Cineform intermediate, and would need 40GB or so.

My take on it is that, if you are happy editing at low resolution, and using the HD files only at render time, Gearshift will work for you. This is the least demanding solution in terms of disk space and CPU.

If you want to edit at full HD resolution, then you need to use an HD intermediate codec. While Gearshift can do this for you as well, it ends up using more disk space than Connect HD (you still have your original M2T files + the Cineform files) and the same CPU. So if you absolutely must edit in HD and your laptop can handle it, Connect HD would be the way to go.

Note that Vegas inludes HD intermediate codecs, and that you can do manually everything that Gearshift and Connect HD do. You could capture M2T, render them to new clips using an intermediate from Vegas, and then work with those. Gearshift and Connect HD basically simplify the workflow - considerably.

Finally - if all you are doing is simple cut edits, the Womble MPEG Video Wizard works very quickly with MPEG2 TS files.

[Alex Thames]

Thanks for that thorough response.

So say I want to edit on my laptop with no extra software. You said Vegas can manually do everything Gearshift and Connect HD can do. So does that mean I can edit at HD resolution without any of these extra software?

How would I do that?

1. Capture M2T files via HDV Split.
2. Put these M2T files onto the Vegas timeline, one at a time.
3. Render each clip individually with one of the intermediary codecs (Sony YUV or Cineform HD).
4. Use the newly rendered clips to edit in Vegas.
5. Render as whatever I want in the end.

Questions:
1. I know that M2T is a transport stream, but what does that mean? How does it differ from .mpeg2?
2. What exactly is the purpose of an intermediary codec?
3. Which is better, Sony YUV or Cineform HD codec, and why?
4. Why is an intermediary codec easier to edit (considering that, for example, Cineform HD codec is 4x the size of .m2t) than .m2t files?
5. Can Vegas simply just edit the .m2t or .mpeg2 files? And if so, which one would be better to use while editing, .m2t or .mpeg2?

Harddrive space shouldn't be a problem. I have a Seagate 300gb, Firewire 400 and USB 2.0, 16mb Cache buffer, 7200 rpm. Should be enough for a couple projects at a time. If need be, I can delete projects after I'm done.

Few more questions:
1. What do you mean HDV Split splits footage into scenes using timecode? How does it use timecode to determine where to split?
2. So, what then, does Virtual Dub do?
3. You say Vegas's HDV capturing application is rudimentary, but how so? If it can't split scenes, then wouldn't it be the same as CapDVHS? Or...what are the differences that make Vegas's capture application inferior?

[Gian Pablo Villamil]

The workflow you describe is pretty much how it would go, except that you can use Vega's batch render script to render all the clips at once. (Put all the MT2 clips on the timeline, mark each one as a region, run the batch render script telling it to render individual regions.)

An MPEG2 transport stream is like an MPEG2 file, except with some additional timing/error correction bytes added. Some programs don't like this extra stuff. CapDVHS has the option to strip it out at capture time.

MPEG2 uses intraframe AND interframe compression. This means that it stores some frames intact (keyframes) and then stores only the difference between keyframes and subsequent frames. This makes editing difficult, because when you go to a random frame, the editor has to find the nearest keyframe and then render all the frames from that point to the point where you are editing.

Intermediary codecs do not use interframe compression (ie every frame is a keyframe) and are usually lossless (ie no quality loss at all). So - MPEG2 is a lossy codec, that is difficult to navigate, but quite compact. Intermediary codecs are lossless, and easy to navigate, though bigger.

I don't know whether the Sony or the Cineform codec is better. Apparently it depends on your rendering target, because they have different ways of representing color space.

Vegas can edit M2T and MPEG2 (and DVD .VOB) files directly. On a fast machine, it can even work quite well for small projects. However, it can become quite awkward and slow for anything ambitious. No real difference between M2T and MPEG2.

MPEG2 TS contains a timecode for each group of frames. HDVSplit looks for discontinuities in the timecode and splits scenes there.

Virtualdub can capture from VfW drivers & devices (ie capture cards, webcams) and is also a very powerful conversion/editing utility.

The Vegas capture utility does not split scenes, has very little control over the camera, and seems less efficient (ie it drops frames). I can't think of a single reason to use it.

[Alex Thames]

Thanks again.

So, when you say MPEG2 is a lossy compression while Sony YUV and Cineform HD are not, I'm a little confused. Doesn't HDV encode video on the DV tape using MPEG2 compression? So even if I converted this MPEG2 into an intermediary lossless codec, wouldn't it still only be as good as the original MPEG2, since it can't add info, it just makes the file bigger and easier to edit?

Finally, so while HDV Split is a capture utility, what is HD Link? Is that even related?

[Gian Pablo Villamil]

Yes, an intermediate codec won't allow better quality than the source MPEG2.

A more useful way of thinking about it is this: a lossy codec needs to interpolate missing info on the fly, which requires computing power, a lossless codec stores the complete frame, no interpolation is necessary. (This is a simplification, treat it as an analogy.)

The practical effect is that you can edit in realtime using an intermediate, whereas editing the MPEG2 directly will be rather awkward and slow. Just try it!

HD Link is the capture utility that comes with Connect HD. Connect HD is the package of HD Link + Cineform HD codec. The version of the CFHD codec that comes with Connect HD is better than the one that ships with Vegas, or so they say on the Cineform website.

All these things have free trials, so just check them out.

[Mark Bryant]

Alex, Gian Pablo,

Firstly would like to say Gian Pablo has done an excellent job giving very detailed and accurate answers.

In terms of lossy vs. non-lossy codec, you are right of course that rendering to an intermediate codec doesn’t improve the quality. What I think was meant is that once you have rendered to the intermediate codec, from that point on it is virtually lossless. What I mean by that is: If you take the native mpeg2 file and re-render it say 10 times, each render will have a loss in quality. But with the CineForm intermediate you could render that 10 times and it will be virtually indistinguishable from the first generation render.

I have a very similar laptop (also a 1.6 Pentium M), so I can also add my own brief experiences. I bought Gearshift, I have not bought Connect HD, though I am still planning to try it. (Per CineForm’s recommended spec, our PCs aren’t up to it, but it wouldn’t hurt to try).

On My PC, I’ve found I can edit the CineForm files at full preview rate for relatively simple edits. Vegas is very responsive on the timeline. So one valid approach (which doesn’t need Gearshift or Connect HD) is as you said (with some edits):

1. Capture M2T files via HDV Split (or Vegas, if you don’t mind them not split. I capture with Vegas).
2. Put these M2T files onto the Vegas timeline, one at a time.
3. Render each clip individually with one of the intermediary codecs (You want to use the CineForm codec here… I believe the Sony YUV is an uncompressed format).
4. Use the newly rendered clips to edit in Vegas.
5. Render as whatever I want in the end.

So, on top of that… how can Gearshift help? It is useful if you want to preview more complex edits, and you find that your PC is struggling with the CineForm files. You can edit a DV proxy, and swap back and forth. Bits where you need to view the full HD resolution you edit using the CineForm file, bits where you don’t care you can use the DV proxy. Or, you can forget about CineForm altogether, and just edit DV proxy (swapping back to m2t at the end).

If you are happy editing with the CineForm flow above (steps 1-5), then I don’t think you need Gearshift. If you want to sometimes use a DV proxy then Gearshift makes this much easier.

The main advantage I see of Connect HD is that it speeds up your workflow. As I said I haven’t tried it yet, but plan to just to see.

Mark

[Gian Pablo Villamil]

Just a quick point, as I said before, if you are using Vegas alone you don't have to convert the M2Ts to Cineform individually, you can do it as a batch. IIRC this is in the Vegas help file.

Put all the M2Ts on the timeline, select each one by double-clicking, turn it into a region by pressing R+<enter> (I think you can automate this step too), then use Tools/Scripts/Batch Render to render each region as a separate clip.

[Alex Thames]

How can the same intermediary Cineform HD codec be better or worse (when you said Vegas's Cineform HD codec was worse than the one that comes with Connect HD, according to Cineform)? Unless there is a difference?

Virtually lossless - so nearly indistinguishable, but on a technical level, there is still a small degradation, right?

If I want to use Vegas on its own with no Gearshift or Connect HD on my laptop, and I have complex edits...I am wondering why my laptop would struggle with the more complex edits and Cineform HD rendered clips? Is it because the complex editing might be too much or is it because the Cineform files are 4x the size of .m2t or .mpeg2 files, or both?

If it is possible for me to get by with patience right now, I would much rather do that as I am strapped for cash at the moment. Even $50 for Gearshift is a bit pricey right now for me, especially if I don't need it. It's not so much just the cost alone either. For example, I might buy a LANC controller even though it costs far more than Gearshift costs because a controller can let me achieve things that are impossible or basically impossible to do without it, while it seems that I can do all my editing without extra software.

As far as free trials - yeah, I might try that. It's just annoying if I want to unistall and clean up afterwards if I end up not buying.

[Gian Pablo Villamil]

Originally Posted by Alex Thames

How can the same intermediary Cineform HD codec be better or worse (when you said Vegas's Cineform HD codec was worse than the one that comes with Connect HD, according to Cineform)? Unless there is a difference?

Look here: http://www.cineform.com/products/ConnectHD.htm

for the heading called "Vegas 6 includes CineForm's codec - do I need Connect HD?" - it didn't sound *that* compelling to me.

As for complex edits, once you start compositing (using multiple layers) the workload goes up significantly.

One way to avoid this is to do intermediate renders ie. once you've finished work on a complex scene, render it to Cineform (near lossless) and then integrate that finished clip back on the timeline. You might be able to use Vegas Smart Render for the same thing.

[Mark Bryant]

Originally Posted by Alex Thames

Virtually lossless - so nearly indistinguishable, but on a technical level, there is still a small degradation, right?

If I want to use Vegas on its own with no Gearshift or Connect HD on my laptop, and I have complex edits...I am wondering why my laptop would struggle with the more complex edits and Cineform HD rendered clips? Is it because the complex editing might be too much or is it because the Cineform files are 4x the size of .m2t or .mpeg2 files, or both?

Alex,

In theory, any codec which uses compression and does not contain all of the original information is not lossless. So in theory there could be degradation, but if you can't detect it is it real?

With CineForm it isn't the size of the file which is an issue - it edits much easier than m2t or mpeg files which are much smaller. By making the video less compressed (so larger files) it is easier to edit. But still as HD has about 4 times the information compared to SD (DV), it is more taxing on the computer than ordinary DV.

You can get a free trial of Gearshift... also note you can save $10 if you use the DVInfo.net discount code "SaveNow."

Mark

[Alex Thames]

I have just attempted my first HDV project. My settings on the camera are iLink convert OFF and Down Convert LETTERBOX.

I captured footage onto my laptop using HDV Split and that worked perfectly. I came out with a bunch of .m2t files.

I opened up a new Vegas 6.0d project file template with the HDV 1440x1080 60i 29.970, I chose upper field first, chose best quality, deinterlace method: interpolate, and motion blur: Gausasian.

I put the .m2t files into the Project Media area, then pulled on onto the timeline. However, here's the problem. When I go to "Render As" there are only a few options to render the project as, and neither Cineform HD intermediary codec or Sony YUV intermediary codec are one of them.

However, if I choose Video for Windows.avi, there is a dropdown menu for template. I chose HDV 1080-60i intermediate. Under description, if I choose this intermediate, it says:
Audio: 48,000 Hz, 16 Bit, Stereo, PCM Uncompressed.
Video: 29.97 fps, 1440x1080, Upper field first.
Pixel Aspect Ratio: 1.333, using CineForm HDV codec. OpenDML compatible.

Is that right? What if I want to turn my project into true 24p, not 29.97? How do I change the frame rate? Should I open up the custom template, click the video tab, and change the frame rate to 24.000 (Film)?

Was choosing "upper field first" the right choice? Or should I have gone with lower field first or none (progressive)? In the end I want true 24p: edit as true 24p, render as true 24p, and projected as true 24p if possible.

I also see an option called "Interleave every (seconds): 0.250" and a setting under it that can be checked that says "Interleave every frame." Which one should I use and why? Or should I change the 0.250 to another number?

Also, I'm wondering about the pixel aspect ratio - what is it, and how does it affect things? I shot in widescreen 16:9 and want to have people ultimately view my project in widescreen 16:9. Is 1.333 the correct ratio to use? Why or why not? Note: on my Vegas preview window in the lower right corner, I changed it to simulate device ratio to make the preview appear widescreen. But what exactly is simulate device ratio - what does it do, why does it need to do it, etc.?

And, I can't find where to check or uncheck the "resample" thing you guys were talking about. Where is it?

Finally, can you tell me the step-by-step to batch render as Cineform HD intermediary codecs if I put all my individual .m2t clips onto the timeline in a row all at once? I want to come out with respective Cineform HD intermediaries (I guess they are .avi? yes?)

Remember, I do not use Gearshift or Connect HD, just Vegas 6.0d by itself with a few other plug-ins. Footage from camera was captured using HDV 1080i setting (I think that's 60i? not sure, please confirm). Camera was my Sony HVR-A1U.

Thanks a lot.

[Douglas Spotted Eagle]

Just choose the CineForm intermediary in the avi dialog. If you want 24p, shoose a framerate of 23.976 and a field order of "none".
Messing with anything else will likely result in errors.

[John Rofrano]

Originally Posted by Alex Thames

When I go to "Render As" there are only a few options to render the project as, and neither Cineform HD intermediary codec or Sony YUV intermediary codec are one of them.

Alex, HDV 1080-60i intermediate under Video for Windows (avi) is the CineForm codec. This will get you an intermediary file for editing.

Originally Posted by Alex Thames

What if I want to turn my project into true 24p, not 29.97? How do I change the frame rate?

If you want you project to be 24p then start a new project using the NTSC DV 24p Widescreen (720x480, 23.976 fps) template. In Vegas 6 it is not that critical that you use this template for your project because it does an awesome job at rendering 60i to 24p but I like to use it because it makes sure any generated media you create are generated at 24p so what you see is what you get in the final product. (i.e., no surprises)

Originally Posted by Alex Thames

Should I open up the custom template, click the video tab, and change the frame rate to 24.000 (Film)? Was choosing "upper field first" the right choice? Or should I have gone with lower field first or none (progressive)? In the end I want true 24p: edit as true 24p, render as true 24p, and projected as true 24p if possible.

You do not want to use 24.000 (film) unless you are physically scanning in celluloid film. You also want your interlace to be set to none. The (p) in 24p means progressive (i.e., no interlacing). Just use the standard 24p template. It is set up properly.

When you render, use the Main Concept MPEG-2 / DVD Architect 24p NTSC Widescreen video stream template.

Originally Posted by Alex Thames

I also see an option called "Interleave every (seconds): 0.250" and a setting under it that can be checked that says "Interleave every frame." Which one should I use and why? Or should I change the 0.250 to another number?

Forget about that stuff. Just use the 24p template.

Originally Posted by Alex Thames

I'm wondering about the pixel aspect ratio - what is it, and how does it affect things? I shot in widescreen 16:9 and want to have people ultimately view my project in widescreen 16:9. Is 1.333 the correct ratio to use? Why or why not?

Pixel Aspect Ration for HDV is 1.3333 but if you are delivering DV Widescreen the PAR is 1.2121. You may have to go into event Pan/Crop and use Match Output Aspect to avoid slight black bars on the sides due to the difference is aspect of HDV. Leave the PAR in the template to 1.2121 if you are delivering DV Widescreen.

Originally Posted by Alex Thames

Note: on my Vegas preview window in the lower right corner, I changed it to simulate device ratio to make the preview appear widescreen. But what exactly is simulate device ratio - what does it do, why does it need to do it, etc.?

Simulate device aspect is just a guide for you. It has no affect on the actual footage. It makes Vegas compensate for the fact that your PC monitor has a pixel aspect of 1.000 and your video does not. In other words, it removes the distortion of viewing 1.3333 PAR video on a 1.000 PAR device and lets you see what your final video will look like on the appropriate output device.

Originally Posted by Alex Thames

And, I can't find where to check or uncheck the "resample" thing you guys were talking about. Where is it?

That is in the event properties.

Originally Posted by Alex Thames

Finally, can you tell me the step-by-step to batch render as Cineform HD intermediary codecs if I put all my individual .m2t clips onto the timeline in a row all at once? I want to come out with respective Cineform HD intermediaries (I guess they are .avi? yes?)

Here are the steps:

1. Place all of the M2T files on the timeline
2. Create a region around each one if you want separate files (i.e., double-click each event and press the ‘R’ key)
3. Go to Tools > Scripting > Batch Render
4. In the Batch Render dialog, choose your render directory
5. Check Video for Windows / HDV 1080-60i intermediate as the render template
6. Check the Render Regions radio button so that each event/region will be rendered as a separate file
7. Click OK and go for coffee (I like mine dark, no sugar) ;-)

Originally Posted by Alex Thames

Remember, I do not use Gearshift

:(

Originally Posted by Alex Thames

Footage from camera was captured using HDV 1080i setting (I think that's 60i? not sure, please confirm). Camera was my Sony HVR-A1U.

Yes your HVR-A1U shoots 1080-60i.

~jr

[Gian Pablo Villamil]

Originally Posted by Alex Thames

...

Finally, can you tell me the step-by-step to batch render as Cineform HD intermediary codecs if I put all my individual .m2t clips onto the timeline in a row all at once? I want to come out with respective Cineform HD intermediaries (I guess they are .avi? yes?)

...

I think you can use the script from here:

http://vasst.com/resource.aspx?id=16...6-fc723f15e6c2

to make a region for every selected event. This will save you a bit of time. Just add all the .M2Ts, select all, run the script, then batch render choosing the appropriate format and the regions option.

Also, on the same site, John Meyer has an improved batch render script that puts the event name in the rendered file, not just a sequence number.