I'm still using MT2 files and converting
 

I'm still using MT2 files and converting via SQUARED 5 is this a mistake? The way I look at it I have the ability to convert in any format needed and review the footage at the same time much like using a deck. It seems time consuming but I have had a lot of problems with the Dr-100s in QT mode and chose to continue shooting in MT2 converting to quick time or whatever needed. Am I making a mistake here? I have seen no loss of quality and aside from the time involved with conversion which incidentally gives me a chance to really review the footage, does anybody see this as a mistake in the process..

I am interested in this question too. Is there any loss of quality using this workflow as opposed to capturing in other methods. I asked this question of my son who is a total computer software guru type person and he says that it depends upon the implementation of the codec and that if it is pretty good there should only be minimal loss of quality.

Rob

This is the first time I've heard of Squared 5. What format are you converting to? Are you using final cut?

Hi Brian,

Squared Five are I think the people who developed MPEG Streamclip so it is a reference to that. I think the question comes down to an assessment of the quality of the MPEG StreamClip codecs. It is a pretty big step committing all your footage to this workflow without knowing the answer to that. It looks pretty good!

Rob

I generally convert to MPEG or QT. I use FC Studio Pro5.1. I bought the original OT upgrade for my first DR100 and it has never worked right so when I bought my second Dr100 QT pre installed I decided to keep with the original
mt2 work flow as opposed to integrating the two.

As I type this, I'm converting an .m2t file into a number of Quicktime movies in the ProRes 422 HQ codec. I've received footage shot by someone else (who didn't allow any pre-roll between clips) and found that capturing natively (in FCP) was losing the beginning of each shot! So I've captured the entire footage on the tape in one go with DVHSCap (a 3GB .m2t file) and am converting it in portions to ProRes.

I don't think you could go far wrong, as a workflow these days, to capture in .m2t and convert to ProRes HQ Quicktime movies (external hard drives are getting cheaper and cheaper these days).

I used to convert .m2t files to AIC (Apple Intermediate Codec), which was quite efficient for its time, but AIC only samples color at 4:2:0 as compared to the 4:2:2 of ProRes HQ. While it's true that the original .m2t file was also sampled at 4:2:0, I think it's better to get out of that as quickly as possible and into something more robust such as 4:2:2, especially if you are exporting (making round trips) to applications such as Shake, Color, After Effects, etc. Hopefully the footage holds together better with successive generations than with using a 4:2:0 codec.

Mind you, if my final destination is DVD and there's no reason to leave the NLE (FCP), I'll always capture and edit natively. But these days if I find the need to capture as an .m2t and convert, I'll only use ProRes HQ.

ProRes HQ is only available with FCP 6. On the PC platform, I believe there are excellent codecs you can buy such as Cineform, Bitjazz, Avid's codec, etc.

The conversion from .m2t to any codec, wrapped by QT or otherwise, can be a sound technique depending on the codec. To refer to Robert's question, it doesn't depend on the implementation of the codec, sure poor codec implementation can make things worse, but on the type of encoding used. ProRes, like JPEG and other lossy codecs, discards data from your footage and so should be avoided for this kind of conversion if you don't want to loose definition in your footage. The only codecs that you can use without generation loss are the ones that compress your footage in a lossless way. In a bit-by-bit lossless system, not in a "visually lossless" way. If you were capturing the footage straight out of the component out port than it would be a completely different deal. As component out is 4:2:2 you get excellent quality by using ProRes. But, if you capture to tape you already have a generation loss as the HDV compression squeezes color information out. If you convert to any lossy codec than you add a generation loss. You might not be able to see it but it's there and it might show up later on when you try to color correct and grade your footage.
Converting from .m2t to uncompressed, PNG-encoded QT or SheerVideo-QT is perfectly safe and I do it on a regular basis. This is because those codecs compress data without discarding anything, similarly to what .zip does.

Eh up David!
I now use the workflow you posted many moons ago on this board (very detailed and very helpful). Works well and it's only a slight workaround over natively importing into FCP.
I was going to ask about ProRes in FCP5, but you've answered that already (soon time to upgrade perhaps) - I guess I'm stuck with m2t and AIC for the time being (absolutely no way of using component out either).
I was away from this forum for a bit so not sure whether the age old missing footage problem with HDV and FCP was ever fixed?
My camera is actually away with JVC at the moment - performing an upgrade to 'A' and the the next generation of firmware - hopefully I'll notice one or two niggles eliminated on return.

Currently working on my first short (5 min dark comedy) - it's a fast turnaround but I reckon I can handle all that's thrown my way (should my camera return!!).
Wish me luck on that :( !

I think that only running FCP/HDV and having no other post program (Shake/AE/Colour etc) I can't really worry about staying in 4:2:0?

Paulo - does the PNG encoded QT reference relate to a previous post of yours (going back a couple of weeks) mentioning that this is a good idea if you want to edit a bare minimum of frames (e.g. adjusting text on a frame)? Would that eat up a lot of power/sapce on your mac (I run on an iMac and ext drives)?

Cheers all.

No, that post was about outputting the master as a series of TIFFs, PNGs would be good too, a practice that will save you a lot of time and worries when you get in post-production. The PNG encoding that I mentioned here is a QT codec. If you look in the list of conversions available from QT you'll see that it's available. It's not nearly as small as others but it preserves the quality of the image 100% because it's lossless. The downside is a bit of sluggish performance. As it's quite well known now I'm a big fan of SheerVideo. Sheer is small enough to be manageable and fasters than many other codecs. It's not free but if you convert footage a lot it's well worth the investment. PNG is included in every version of QuickTime

Thanks Paulo - I'll definitely look into Sheer (and I think there's a free demo knocking about?), although without sending anything from FCP (aside from Motion, LiveType, Soundtrack) is it worth my while?
Ahh - same type of PNG (as in portable network graphic)? They need a new acronym if not!

Anyhow, I'll be sure to add your name should I get this short finished Paulo (TC3)!
Cheers.

Yes, they provide a 20-day fully functional, no watermark, demo and the reader is free so you can distribute the clips to other people as well.

It could because with Sheer you avoid the transcoding and data loss caused by convert HDV clips to ProRes. At the same time it's faster than PNG so, it's really a matter of workflow and convenience.

Actually that was just a parallel. They use a different algorithm and they have an additional routine that checks the color values from the original and they guarantee faithful conversion between the RGB and YCrCb color spaces.
I was just drawing parallels between lossless codecs.

Thanks for credit, I appreciate it.

Thanks for the replies Paulo (and a credit's the least I can offer!).
I do not currently own FCP6 (running v 5) so ProRes is seemingly beyond me anyhow. Currently weighing up whether I desperately require FCP6, and if something like Sheer can provide lossless HDV capture then aside from 5:1 surround in STP (which I can obtain at any rate via Logic Pro) and Color (3 way in FCP5 suffices at the moment) I'm not so sure I do.

...of course this all depends on me getting my camera back from JVC (it will be 10 days tomorrow).

Whilst I'll probably continue to capture (via squared 5) m2t files did anyone manage to successfully capture natively in to FCP in the end (after various updates to both cam and by apple?). I remember discussion on setting timecode to either FreeRun or REC (?!?) to cure this...

Hi David!

Yes, AIC gives excellent results (when Tim Dashwood first pioneered the AIC workflow, he mentioned that it is supposed to be visually lossless even though, as Paolo points out, it's not the same thing as theoretically lossless).

ProResHQ, being a broadcast-quality format (with 4:2:2 color sampling), should give an even better result.

Yes, every time I upgrade FCP, I tend to forget about the fact that not everyone else has. So, sorry about that! My simple thought was, when you are scrolling through the list of codecs to convert to in HDVxDV or MPEG Streamclip, why not go with the (superior) ProResHQ over AIC? (But this assumes that both already exist on your computer!)

Mike Curtis has made a thoughtful post just today about ProResHQ (which he calls a great codec and considers it better than Panasonic's D5 format) and is campaigning for Apple to make it available to people who DON'T own FCS 2. It's called "Let ProRes roam free!" and can be found on this link:

http://www.hdforindies.com/2008/02/y...roam-free.html

Personally, "in the future" I can see ProResHQ becoming the broadly adopted standard for people editing HD on the Mac platform. Why? Because it's a broadcast-quality HD codec and everybody has it (or will "in the future" upgrade to it).

But Paolo is right to campaign for the SheerVideo codec. It's not very broadly known about and, if as they claim on their web page, "SheerVideo is the fastest and most versatile codec in the world," then it definitely should be promoted and mentioned a lot more.

The one question (and reservation) I have about SheerVideo is the file sizes. For these ProResHQ conversions I've just done, the ratio is about 4.3:1, i.e. for a 1GB .m2t file you get a 4.3GB Quicktime movie file in ProResHQ.

What is the conversion ratio for an .m2t file to a SheerVideo Quicktime?

Let us know how that goes. Because, if your final output is to DVD or web (and you do your color correction, etc. within FCP), then capturing and editing natively is the way to go. And export directly from the timeline to Compressor to encode your DVD assets or web movie. It completely eliminates the transcoding step (to AIC or ProResHQ or SheerVideo, etc.).

And, supposedly, FCP will internally process your effects in uncompressed space. I made a post quite a while back giving a link to an Apple spokesman talking about it. And I think Steve Mullen made some recent posts in this forum talking about the same thing.

But a lot of people don't want to capture natively because of frustration with mid-clip breaks. Very understandable! So I'm very interested to hear of your results with native capture after the upgrade.


Absolutely! Sounds like fun. Perhaps we'll see some excerpts or stills in the "Finished ProHD Projects" sticky?

Cheers the reply David. A possible ProRes plug in for FCS1...that'd be handy!
I'll give the native workflow a go once I get my cam back and let you know, although I'm not holding up much hope for the mid clip breaks being cured.

Ta for the linkie - I'll give it a good read.
Yes - I usually export directly from FCP to compressor (then DVDSP), although in between all that I'm usually mucking about in Motion (really starting to like that) and LiveType (fancy daft Fonts aside there are some great texts to use) importing .movs back into FCP.
I'm hoping if I record 5:1 audio in Logic (cannot do this in Soundtrack Pro 1) I can still have the 5 channels of audio once in the version of DVDSP I have (FCS1).
Be good to post some finished projects up (and hopefully large filed versions and not tin pot youtube versions!).

Many thanks.

Excellent point because it gives us an opportunity to discuss a couple of topics related to HD editing that have evolved over the past two years.
First of all the cost of storage. I recently went to buy an additional hard disk, I didn't shop online, I just took my car and went top the local electronic store here in small Santa Cruz, CA. I bought a 500GB Western Digital SATA drive for $160.00 including taxes. This is approximately a cost of 33 cents per gigabyte.
Now, I know that a lot of us, including myself, used to get those FW disks. They are easy to get at the Apple stores but I had a terrible experience with those. 3 out of 4 broke down and the whole setup is really not that flexible. I use a MacBookPro for editing. I bought a 2 disk Sata Enclosure from FirmTek and later on the new 5-bay unit with port multiplication. I now have more than 2TB of storage available and can swap disks like they were floppies. Sata is the way to go, even for people with a laptop. This changes completely the approach about file size and backups.

Secondarily, ProRes, in the words of Apple, is not a delivery format, it's an editing format. Sheer can be used for acquisition, editing and, giving away the free reader, for delivery. Delivery in this case, I have to be careful around here ;), doesn't mean delivery to network and such, but it can be used for clients and delivery to post houses. ProRes, given the linkage to FCP cannot be used in that regard. This might as well change in the future and it probably will but it remains the fact that I see ProRes as a intermediate format, and not a mastering format, while Sheer can be employed in every step of the pipeline because it's completely lossless.
Regarding the speed, yes, it's freaking fast but I didn't test it against ProRes for rendering and such. It's definitely much faster than Uncompressed and more than twice the speed of HDV when rendering. Testing was done with the HD100 footage inside FCP 5.x

It's very similar over here. I can get the exact same drive at a nearby store for $132 (US$123).

That's why I was (and still am!) interested in the size ratio of .m2t to SheerVideo Quicktime. I'm editing a feature (alongside some corporate jobs) and am now considering trialling SheerVideo for the feature footage (and am wondering what size hard drive I should pick up). Even a year ago something the size of SheerVideo would have been "out of the question" storage-wise for a large job like a feature. (And two years ago I wouldn't have considered ProResHQ, even if it had existed back then.) But not today.

So I take it that this would mean I'd simply take my hard drive with the SheerVideo Quicktime to a post house and, so long as I provided them with the "free reader", they can lay it off to HDCAM tape?


That's very good to know, Paolo. Thanks.

Yes, as long as they are willing to install a QT plugin. Once installed any QT-compatible application will be able to access the clips. The reader is available for both Mac OS and Windows and it's free so you could copy it to a thumbdrive and carry with you all the time.

Looking at around £64 (incl Vat here). That's very cheap. I've recently purchased a GTech Graid 500Gb but now wondering whether I should've gone that route. I'm a little confused over how you would hook these up to your mac however (I'm on an iMac)? Do the enclosures have FW connections or are you connecting some other way - and if USB2, is this fast enough?
At these prices I can how experimenting with the aforementioned codecs make real sense.

A question for Paolo
 

Hi Paolo.

I'm running a test tomorrow with a High Definition projector. I'm trying to determine the best codec to use for digital cinema projection of JVC ProHD footage (720p24) and I'm starting to suspect that SheerVideo could turn out to be the perfect codec for digital cinema with ProHD.

I've just downloaded the trial of SheerVideo (I'll get to my question about SheerVideo shortly but first I'll give a bit of context) and made a couple of test clips. They played back perfectly in the Quicktime player. Surprisingly easy and smooth (for such large files). I also made a clip in JPEG2000 (the codec of choice for the Digital Cinema Initiative [DCI]) and it had a very stuttering, jerky playback in the Quicktime player. And this was on a six-week old iMac (2.4 GHz Intel Core 2 Duo with 4 GB of RAM). So I'm now testing purely with SheerVideo for Digital Cinema.

My question concerns verifying the best settings to convert the .m2t file into a SheerVideo Quicktime with zero loss or alteration and you are definitely our "resident expert" for best results with this codec and JVC footage.

When I go to do the Quicktime conversion of the .m2t file with HDVxDV I get about eight different codec choices for SheerVideo, and then each one contains several different options internally for fine tuning. So far (and under the time crunch of getting the clips ready for the forthcoming test) it seems to me that plain, old "Sheer" might be the best of the eight and, under the Compression settings select "Best Depth" (or should it be "Millions of Colors+"?) and under Options select:
Algorithm: Self-check
Coding: Automatic
Pixel Format Conversion: Perfect
Video Color Standard: HD (I'm not sure where HD 1035 would apply)
RGB input Range: Automatic

Will these settings give me the best results for ProHD footage? Or am I way off?

I guess the result that I'm seeking is getting the same pristine results as I saw over a year ago when some of my .m2t files were projected at a JVC event with their player and projector. I truly was stunned and amazed at how good it looked. And I guess the challenge is to replicate that all the way through post, including a trip out to Color and a couple of clips to Shake and Motion, and then to the final projection.

We're also testing a purpose-built computer box to hook up to the projector (currently with Windows but later it will run on Linux - I'm fortunate to have a computer/IT whiz helping me with that aspect as I'm pretty hopeless with such matters) and we'll download the free Windows SheerVideo reader for that. And I believe SheerVideo are working on a Linux reader as well. (It can't come soon enough!)

I really appreciate any tips on this, Paolo. Because, if this test works, I think I'll "bite the bullet" and go back to the original .m2t files of this feature-length movie and transcode to SheerVideo and insert each clip one-by-one to match the existing timeline. (I've been dreading this but, if it results in a better viewing experience for the audience, I'll have to take a week off and just grind through it full-time. But that's why I want to make sure I'm using the right settings in the first place. I wouldn't want to do this twice!)

So thanks, Paolo!

(Note to moderators: Feel free to move this if you think there's a more appropriate forum. It contains ProHD, Digital Cinema, Mac, Windows and Linux. I just took a guess on ProHD.)