Music Video Link: HD100/M2/Nikons

[Taylor Wigton]

Hello to all- I'm trying to upload a .mov 720p file via the attachments options, which has a split screen of a normally captured 24p file, alongside the exact same image with what I would consider to be my "normal" grading or color adjustments. Someone chime in and let me know what the file size max is on DVinfo (Chris, Tim, etc). No luck so far.

Michael- sorry for the delay in responding.... I will give my thoughts on the subject you posed, but ultimately, I want to let people decide some of those things on their own using native files side by side or split screen to analyze the content. In general, 4:2:0 has beat out all my expectations and has led me to do just that, which is aquiring the most information, the flattest image possible as the results in post have been alarmingly good. HDCAM recorded to tape is 3:1:1 but now we're just getting into the numbers game '-) I am also a firm beleiver in the concept that being on set has far too many variables and the viewing conditions are less then ideal, as well as the pace of a set, then add in all the uncertainly of "exactly" what the director want's as his final look, and so at the end of the day, I try to get the cleanest image recorded to tape/disk/drive so that when the dust settles and the frame of mind two weeks later has altered, the director has the most image possible to work with.

More thoughts later, and some content as well. I think the DVD will happen sooner rather then later. Trying to get all my ducks in a row and then move ahead. Alot of cleareances to be had, but plenty of cleared content to use from all the different cameras so this should be interesting! best, tw

[Tim Dashwood]

Originally Posted by Taylor Wigton

Hello to all- I'm trying to upload a .mov 720p file via the attachments options, which has a split screen of a normally captured 24p file, alongside the exact same image with what I would consider to be my "normal" grading or color adjustments. Someone chime in and let me know what the file size max is on DVinfo (Chris, Tim, etc). No luck so far.

Hi Taylor,

I'll send you an email.

Tim

[Steve Benner]

Taylor - Awesome Video! This is one of the videos we should point to when people bring up those HVX vs. Canon vs. JVC debates. It looks great!

[Taylor Wigton]

(despite the name of this thread, the M2 was not used here in the below QT)

http://www.447productions.com/splitscreen.mov

HD100, stock fujinon lens, NO ND grads or pola here, 720p24 HDV, detail -3, Cinegamma off, black normal, white clip 108%. 1/48 shutter. Long end of the fujinon (70-80mm) and iris somewhere around f/T8. Image on the left is straight off the tape with ZERO grading applied. 3 way Color Corrector and RGB Balance used on the left side.

Let me know if the Qt opens properly. Thanks-

****please note: the link to this .mov is 27mb, and should open at half size when download is complete. if you click the lower right corner tab do a "save as" QT movie to your desktop, you will be able to play the file in it's "actual size" which in this case is 1280x720. Codec used here was HDV 720p24 using FCP 512.

[Taylor Wigton]

A couple months ago, a buddy of mine wanted to take some of the HD100 demo footage I had shot for JVC last spring and see if long GOP HDV was strong enough to endure ANY post production manipulation. We ultimately wanted to find out if HDV was a one trick pony. Was it as flimsy in post as everyone was suggesting? Did we really need to bake the "look" in camera? Is HDV's 4:2:0 long GOP Mpeg2 image just a chicken running around with it's head cut off? A little mouse on the praire? Or on the other hand, is this format just completely untested? Whatever the verdict, I caution you in making any steadfast conclusions be them good or bad based on the QT links here, as they are small and compressed, and there is not enough information included as to what was done to the material or if it was pushed too hard, not hard enough, etc. What do the native files look like..... on and on.

The first link is the graded footage and the second link contains several clips which are identical but "ungraded." You may want to put them side by side to give a better frame of reference.

(graded HDV) http://www.447productions.com/rwedickCC.mov ***also note that this file is 27mb and a blank white screen with a Q on the page will appear until the dowload is complete. this file was encoded h264 640x360 at 23.98****

(ungraded HDV) http://447productions.com/stock.html

[Paolo Ciccone]

Originally Posted by Taylor Wigton

A couple months ago, a buddy of mine wanted to take some of the HD100 demo footage I had shot for JVC last spring and see if long GOP HDV was strong enough to endure ANY post production manipulation.

Hi Taylor.

HDV is not different from other temporally-compressed codecs. It's a good system for encoding and decoding but should be avoided for editing professional-grade material.

Temporal compression generates intermediate frames that are deltas from a given keyframe. This means that any NLE needs to continuosly backtrack and recalculate frames every time you cut in the middle of a GOP, something that happens very often.

When outputting your footage to bring it inside a compositing program, for example when going from FCP to AfterEffects, if you edit in HDV and output in HDV you end up with recompression. At first you might think that because you use the same codec there should be no recompression. That is true for codecs that don't use temporal compression. DV is such a case. When using GOP-based codecs the only way of recreating a video stream from an edited sequence is to recreate new GOPs and this means recompression with possible loss of definition.

For this reason I usually suggest to rough cut in HDV if that is your source format but then, as soon as possible, convert your footage to uncompressed and do all the color correction/compositing in that format.

Export times are also affected dramatically. Exporting from FCP using AIC takes a fraction of the time required for doing the same in HDV. A confirmation of the need to transcode the footage when using HDV.

[Miklos Philips]

Taylor,

What did you use to pull focus on the Nikons? Did you just use the barrel on the lens manually or used some follow-focus system? Thanks,

[Taylor Wigton]

Originally Posted by Miklos Philips

What did you use to pull focus on the Nikons? Did you just use the barrel on the lens manually or used some follow-focus system?

Redrock Micro loaned me thier new follow focus setup, which worked quite well. You can learn more about it on their website, but they came up with a very smart method of attaching gears to ANY kind or size lens. A nice whip to top it off.

I removed the servo zoom/iris control on the fujinon and attached one of thier lens gears to the Fujinon (no M2) and so the Redrock follow focus system does double duty. Works excellent shooting with or without the M2.

[Taylor Wigton]

Originally Posted by Paolo Ciccone

HDV is not different from other temporally-compressed codecs. It's a good system for encoding and decoding but should be avoided for editing professional-grade material.......

Paolo- Some good fact based information. I think that there is no question that MPEG-2 compression and long GOP will be replaced by superior compression methods. JPEG200 and AVC are just around the bend. HDV is without a doubt a format that IMOP was designed to launch HD and get it into the hands of everyone in a cheap and efficient manner. I think that even the HDV consortium did not anticipate the kind of positive reaction that emerged as people got hold of HD.

To reiterate: I think that Paolo's post was sharp. However, in the case of motion pictures, the brain requires the eye as a conduit for which to make the final descision even in the face of FACT. Cinematography is not mathamatics in that math has absolute answers. I would love to hear if anyone looked at the split screen on a HD monitor and had any thoughts be them good or bad.

My thought on MPEG2 long GOP compression is that if you pull a frame grab and magnify it by 800% in Photoshop, you are certain to lose sight of what the reality on the ground is. If we constantly read about the pitfalls of one format from a scientific standpoint only, the POTENTIAL exists that people will walk around with THAT information and it will become true for them. And yes, if you step through each frame of HDV, the I-frame and it's surrounding frames are plain as day. There is no denying that there is one great frame and others not so good, but what happens when you play the clip at 24fps. Our eyes don't seem to have a good foothold on the inferior frames in the sequence. Couple that with the inherent nature of motion blur and things don't look all that bad.

I assume many of you have read the article by Walter Graff where he laid out the facts and then shot the test and let us SEE the results. 4:2:2 and 4:2:0, the HD100 and HVX200. http://www.bluesky-web.com/HDVHVX.htm

So at the end of the day, I think the best combination is scientific based facts coupled with visual content put in a proper context so that people can make a well rounded conclusion.

[Paolo Ciccone]

Taylor.
I agree in principle with you. Our eye is the best judge of the final result. The point of my post though was that we have to be careful when we make assumptions. A lot of people assume that, if you use the same codec for input and output you don't get affected by transcoding. When using HDV in your NLE you do transcode your footage, there is no way out. Transcoding is like a generation loss. Do it enough times and yes you will see it. HDV is fine for now, I got amazing images from MPEG-based codecs both with the HD100 and with the Sony F-350. Where we have to be careful is in carrying those images along the post-process. That is where the facts allow you to be better prepared and to avoid damaging that image that you so carefully crafted in the camera.

[Bill Ravens]

I'm sure you guys have seen the generational loss tests advertised by Cineform. They claim little to no losses, with tests to support their claims. While those of us who work in the real world beleive such claims may be a bit guilded, it comes very close to truth. Used to be a cadre of people, back in the heyday of DV, who transcoded their raw DV to uncompressed, like HuffyUV, did all their editting in HuffyUV, then transcoded out to their final format. Basically two detrimental transcodes and numerous intermediate transcodes to a lossless format. I beleive that as long as the internediate processing happens in 10 bit space, or better, it's truly lossless to work in a lossless format. Of course, there's probably a little dithering loss going back out to 8 bit, but, it's probably small.

[J.R. Hughto]

Originally Posted by Taylor Wigton

I have tested the Miranda HD-Bridge DEC+ with the HD100 at 24pHDV sending out from HD100's firewire port, and it DOES uprezz signal to 1080PsF 24 via HD SDI w/ Audio and Timecode from the Miranda into an HDCAM SR deck. Exactly what I've been trying to do without using 3rd party hardware for a long time! But maybe this has been a known fact for some time time now???

I haven't seen anyone talking about this, either, but its exactly what we've been doing. We did a number of tests with Matchframe over in Burbank this summer (they have a Miranda), and the results look great.

One thing that we didn't have time to test that I'm interested in is seeing the difference between hardware and software transcodes. When shooting DV, I'm used to color correcting in DV at home so I can look at it on a monitor through the deck, and then destroying the render files and re-rendering in 10-bit uncompressed for layback to DigiBeta. I assume (careful, there) that I'd be doing the same thing with HDV - namely, editing natively, color correcting through the deck on a monitor in HDV, then wiping the render files (or not even rendering) and then taking the sequence to HDCAM for upconversion along with rendering (I do take some time to spot check each cut to make sure that I like the color correction in uncompressed as well as I liked it in DV).

However, the other idea I've had after seeing the Miranda results is to lay back the HDV sequence to tape, and take the tape in to do a hardware transcode using the Miranda. From there, do a tape to tape color correction at a post house. The difference in price shouldn't be that large - a Miranda session instead of a Final Cut one, probably only with the addition expenses of audio layback and the tape to tape (which in theory would be better anyway, with a professional colorist).

Will there be a difference in concatination? Has anyone tested this?

[Taylor Wigton]

Originally Posted by J.R. Hughto

However, the other idea I've had after seeing the Miranda results is to lay back the HDV sequence to tape, and take the tape in to do a hardware transcode using the Miranda......Will there be a difference in concatination? Has anyone tested this?

EMAIL TO SHOWREEL EDITOR, RE: HDV Output to Existing Industry Standards, and examing the denegration from transcoding HDV using NLE.
---------------------------------------
Steve,

- since both clients, broadcast, and film festivals are seeming to require
these formats, betaSP, Digibeta, HDCAM, we are now at the point of taking
the HDV format and converting it to ALL of these formats so that we can
prove (or disprove) that aquiring on HDV is possible, and delivering on
the more commonly used industry formats is a step that is viable both
technically and financially. (i don't actually know the answer to this
because we have not DONE this yet, but this is what we will discover)

- attached is the HDCAM deck that we need to record to, and in taking to
mr wedick yesterday, the HDCAM SR deck will NOT work in making HDCAM
recordings. they have different internal motors. band pro has HDCAM SR
decks but does not have HDCAM decks, and randy also indicated that using a
f900 camera head is not an option technically for which to use as a "deck"
to make the HDCAM recordings. (correct me if i am wrong here)

-therefore, we need to figure out an arrangement with a post house to
complete the task at hand. i have secured the miranda dec+ and can pick
that up anytime at a distributor in burbank which has been kindly setup by
miranda, so the company PRECO has told given me open ended (within reason
of course) use one of their miranda hd-bridge dec+ units and both miranda
and preco have given me all the tech support and contact information
should that be needed. i also have a new JVC HD100 to replace the other
HD100 that has some circuit issues. lastly i have a MacBook and FCP 512
which successfully prints to tape a FCP timeline. so we can now make an
HDV recording of the 720p24 projects (music video/short film) using the
HD100 as the VTR. i have read the there is some generation loss when FCP
does the HDV transcoding for output to tape, but this is yet to be seen. i
also read that Cineform can do a lossless HDV transcode, and Liquid, but
that is inuendo at this stage and perhaps something to look into further.


- when me and rodney were doing the tests last year at Level 3 Post in
burbank, i got to know one of the head technical managers there and he
gave me his card and mentioned to keep in touch. since we brought in the
miranda box last year and the different HDV cameras, some of the guys were
super helpful and also very interesting in the outcomes as they realized
that these may be some of the situations they would be working with moving
forward. So Level 3 is a massive post house with millions of dollars
worth of state of the art post equipment, but i think it is worth me
making a call to them and getting in touch with the guy from last year to
see if i can come back in to do the what we were still unable to
accomplish with them last year. i will mention Showreel of course and
that is usually a big part of getting stuff for free. i will also present
to them the other issue of seeing if it is possible to do real time dubbs
of everything. for example, can you guys patch together the HD100,
miranda, HDCAM, Digibeta, and BetaSP simultaneously and make recordings to
all three formats on the fly.

- there are several post houses in LA to choose from of course.
Matchframe is another post house i've heard people talking about as
playing around with some of these new tools, the JVC/Miranda/HDCAM
specifically. So they would be good to speak to. I think the fact that
we are bringing the product owned equipment we have and support of both
JVC and Miranda should help in getting this task accomplished for little
or no money, which the incentive being a constructive article written
where they get free pub. i can also suggest working with them off hours
as rodney and i did last year.

- the music video has a deadline of friday. i have a shoot on friday but
i will make the calls tomorrow or maybe actually go into Level 3/
Matchframe because that always seems to work better.
END EMAIL
------------------------------------------------------------------

The Showreel Article along with a DVD tutorial will attempt to examine the options when aquiring on HDV and what is required to make it viable in various capacities, from web/SDTV/Broadcast/Theatrical/Festival venues. How can 35mm lens adapters be used correctly and do they enhance the value of a particular project that might lend itself to the 35mm look and feel.

J.R in the above post brings up some interesting ideas regarding HDV color correction and uprezzing the native footage and using the offline HDV CC as the guidline for a colorist that can be used after HDV exists in a new format.

The biggest question I have at the moments is how best to aquire and post HDV using which NLE and codec, should the content be exported to another codec and how then can the finished project be delivered in the most lossless, high quality, and reasonable priced way.

I think has become an interesting dialogue/thread and I hope more people (like J.R) who may just lurk feel inclined to share thier thoughts more often. It's extremely useful stuff!

[Kristin Stewart]

Taylor,

My 2 cents for the HDV post regarding codec : Cineform is a lossy one but is very efficient. Maybe it would be interesting to look at Sheervideo (www.bitjazz.com), a lossless one. It appears to support alpha channels, etc... and to be cross-platform.

[J.R. Hughto]

Originally Posted by Taylor Wigton

Matchframe is another post house i've heard people talking about as
playing around with some of these new tools, the JVC/Miranda/HDCAM
specifically. So they would be good to speak to. I think the fact that
we are bringing the product owned equipment we have and support of both
JVC and Miranda should help in getting this task accomplished for little
or no money, which the incentive being a constructive article written
where they get free pub. i can also suggest working with them off hours
as rodney and i did last year.

We were actually the first project to go through Matchframe with the HD100 - when we brought in the camera to do the transfers, they hadn't even seen one yet. I guess its a small town, in some ways! (unless of course I was the person talking about them in my previous post)

I'd be extremely interested in reading about your results. I actually asked them in August to allow me to go through these same tests with them step by step so that we would know a) the most economical means of posting HDV b) the best means of preserving the image integrity (these two not necessarily being one and the same). Unfortunately, it really wasn't in their economic best interest to do this with me, and so they declined (which is not to say that they aren't a very nice, helpful, and knowledgable staff). I think this was in large part because since we were the first through their doors with the equipment, they wanted to wait and see if this was in fact a format that they were going to need to support. From the sound of things, Taylor, you have a great package to offer them, and hopefully Matchframe or another local house will jump on this so we can get the information out!