|
|||||||||
|
Thread Tools | Search this Thread |
June 2nd, 2006, 10:48 PM | #16 |
Great DV dot com
Join Date: Sep 2003
Location: Lewisville, NC
Posts: 78
|
Hi, all --
Since I was the one behind both the footage and the filmout, let me jump in here: 1. There was no deinterlacing of the hd100 footage, it is native 720P. We did "smart" deinterlacing of the Z1u footage which does not lose rez but creates simulated motion blur. 2. Since I have the luxury of doing some of these test on someone else's dime, I thought the test was worth it - and indeed it was interesting. I intentionally included some "perfect" shots and some less than perfect shots in the mix to see how it held up -- and the green screen shots were done in a single pass, as most folks will do. Many times pro compositing must create a multi-layered key, at least core, edge and garbage, to get a truly clean key even with 4:4:4 footage. Having seen a bunch of filmout samples of low-budget video-originated films, the original material is often less than perfect. The less-than-perfect video shots (underexposed and stretched in post) held up better than the Green screen shots. Just picked up grain/noise in dark areas. The green screen shots looked noticeably worse than on an HD projector. You can do really pretty nice green/blue screen with HDV, much much better than DV - in fact, pretty near to many 4:2:2 cams. I have been able to pull pretty near perfect keys with Keylight (no uprez of chroma) for some shots or with PriMatte (good chroma uprez, though we turned it OFF for the shot Stephen saw). I recommend PriMatte over UltiMatte at this time for HDV. 3. The "soft focus shot" Stephen refers to was a long shot of a large pond with grass and trees, lots of detail. While it is true that pulling precise focus is tough with the cams (unless you have a true HD monitor along) we're doing some more testing on that, because it appears that the softened focus was the result of higher compression applied due to the amount of detail. Since HDV has a constant data rate, higher detail requires heavier compression. In DV's DCT schema, this results in "mosquito netting," but in MPEG2 it results in softening of the pic. I will be testing a direct-to-disk recorder in the next couple of weeks and we will do the same shot to compare. In fact, this latter issue appears to me to be more of a critical issue than the lens -- unless you are doing all medium shots and closeups. Panoramic shots with lots of detail seem to be the downfall of the format no matter what the lens. I have already suggested to the manufacturers that one of them needs to do the same thing Sony and Panny did with DV - come out with an HDV-Pro version that runs the tape faster and uses lower compression.
__________________
John Jackman www.johnjackman.com |
June 3rd, 2006, 12:12 AM | #17 | |
Major Player
Join Date: Apr 2005
Posts: 512
|
Quote:
|
|
June 3rd, 2006, 09:47 AM | #18 |
Regular Crew
Join Date: Feb 2006
Posts: 184
|
Actually that is thinking backwards. I'm a big fan of tape for ease of storage and price but the future of hiqh quality recording is non-tape systems not running tape faster. That is part of the problem with this industry, we always try to cram new technology and formats into old habbits.
Think about it, right now we can take a stream out of the camera to a Hard Drive (portable and less than portable), why worry about some type of new compression running to faster tape when we can just use something like the Cineform Codec (10 bit even) to hard drive. The only thing missing is the processor to do the math on the fly. It's not hard to do, in fact the Silicone Imaging ("eve") camera is going to Cineform RAW. I've been testing the CitiDisk HD drive with the HD100 and have been impressed that this HD is taking the firewire stream and saving it to files on a lap top drive and doing a little computing along the way to break up files, wrap them in a file format, etc. With just a bit more logic and computing power added to the box you could easily have this save out in a much better format than anything running to tape. Ok, so I know the most important part of the puzzle. The cameras need to spit out raw data from the firewire port and not just to the component out otherwise we can only take the mp2 stream and convert it. |
June 3rd, 2006, 09:52 AM | #19 | |
Trustee
Join Date: Apr 2005
Location: Santa Cruz, CA
Posts: 1,116
|
Quote:
|
|
June 3rd, 2006, 12:46 PM | #20 | |
Great DV dot com
Join Date: Sep 2003
Location: Lewisville, NC
Posts: 78
|
Quote:
In the meantime, in a couple of weeks I will be testing the Wafian recorder to do just what you suggest -- although it ain't gonna work for run-n-gun docs, it would be fine for movie production work wher you are already toting lots of gear. But the proof is in the pudding, and a comparison of the exact same shot from HDV tape and Cineform 4:2:2 will tell the story.
__________________
John Jackman www.johnjackman.com |
|
June 3rd, 2006, 11:31 PM | #21 |
Major Player
Join Date: Apr 2005
Posts: 512
|
There is an excellent reason I would rather keep recording to tapes. Hard drives and P2 are great when you're working on something in-house with the infrastructure to handle it, but hell when you're working with a new client every day as a freelancer, especially in the field. I can record something on a tape, give it to the client and all that's left is to collect my paycheck. Tapes are cheap, practically disposable. You buy a ton and put the ones you use on your invoice. The problem with P2 and hard drives in this sort of environment is that they need to stay with you, but the data contained on them needs to go with your client. It literally goes against the entire point of having a portable storage medium. You can talk about bringing a ton of firestores along or having the client provide their own P2 or hard drives (hoping they're compatible with your camera), but I just think that a tape is less hassle in the end.
|
June 4th, 2006, 12:08 AM | #22 |
Regular Crew
Join Date: Feb 2006
Posts: 184
|
With clients where you give away a tape at the end of day you really can't compare - I agree.
However, tape is just a medium, soon other forms of media will be just as easy to "give away" and this is what we want to work towards. Not only for their quick access (instant access) ability but the fact that we can stream high quality footage to these "new" mediums by just changing compression schemes. In other words once a piece of hardware has the ability to spit out data it's just a matter of changing software to get various formats, compression ratios, etc. Trying to do this the current DVCPRO or HDCAM way would mean changing the entire tape drive system, tape formula, construction, etc. Just for fun I looked up an average price of DVCPRO Tape: $15.00 average for 60 min and HDCAM runs an average of $35.00 for 40min and $50.00 for 60min (large format tape). Assuming I was going to shoot 7 hours of footage It would cost me $105.00 for DVCPRO or $385 for HDCAM (11 40 min loads). Well that is about the number of hours of high quality "compressed" HD you can fit on a 80gig drive. Depending on what make or model, these can be had for $80-$150. One of the things I love about the HD100 is that I can go to tape and Hard Drive at the same time. This gives me instand backup. In time though, I'll be ok with tapeless formats. |
June 4th, 2006, 01:58 AM | #23 | ||
Major Player
Join Date: Apr 2005
Posts: 512
|
Quote:
Quote:
To be fair, once you get the drives back to the studio you can archive them on 3.5" drives which are much cheaper per GB (about half the cost) than the laptop hard drives you see in these portable recorders. Last edited by Stephan Ahonen; June 4th, 2006 at 02:40 AM. |
||
June 4th, 2006, 10:43 AM | #24 |
Regular Crew
Join Date: Feb 2006
Posts: 184
|
If you want to be right, then be right. You are missing my point or just not seeing it in order to support your tape based opinion so I'll drop this silly argument (difference of opinion).
1. Tape can store anything, but if part of the storing process means having to change the tape speed, tape size, or tape forumlation then it's not just like a hard drive. Remember earlier in this thread it was suggested that we just slap a DVCPRO Tape in the HD100 and run it faster. That is not the solution to our HD100's or any other low cost camera to provide the quality recording we need. 2. I never suggested that we record DVCPRO HD or HDCAM to Hard Drive. There are plenty of compressed formats that will give you better than HDCAM quality and the 7 hours I was talking about. Take CineForm RAW for example: 10 Bit vs HDCAM 8 bit. Full image saved to 4:2:2 vs HDCAM 3:1:1. (To be fair I believe at 1920x1080P Cineform RAW will give you 4 hours on 160gb but lessen the rate a bit and you could easily fit more time plus keep quality high). If you want to see Cineform RAW in action: http://indiefilmlive.blogspot.com/20...some-more.html Why should Sony push for you to record on to a simple Hard Drive when they can sell you HDCAM equipment, expensive HDCAM tape, entire HDCAM workflows, including 50K decks? It's time for a change, the major players are holding on to anchient worflows because they have huge roots here and they know it's a big businees for them. Why doesn't RED, SI, Arri D20, etc. record to tape? And my simple point that seems to be lost is that you can give away a hard drive just as well as tape stock to your client at not much more cost. Here is what I would tell my clients "Here ya go, just plug this dirve in and start editing". Client "You mean, I don't have to take this over to Post House X and have them ingest it for me on that 50K deck which cost me X numbers per hour?" or "You mean I don't need that deck that I bough for 50K anymore? Well that's good because my heads were just about due for replacement". Anyway, I hope you see my tongue and cheek humor here. |
June 4th, 2006, 01:55 PM | #25 | |
Wrangler
Join Date: Jan 2002
Location: Los Angeles, CA
Posts: 6,810
|
Quote:
Case in point: I'm currently working with the Genesis and our first AC is a 30 year veteran focus puller who has a full arsenal of tools working--Panatape with remote readout, remote lens controls so he can sit to the side and triangulate the distance, 12" Astro monitor with magnification mode and peaking turned way up so he can see critical focus--and yet he is regular fed focus notes ("come just a pinch closer" "you're just a tad deep" etc) from the DP who is watching from the big CRT back at video village. And we are generally shooting between 2.8 and 4...! The AC's contention is that for whatever reason, depth of field is just a bit less than expected vs 35mm, somewhere between spherical and anamorphic. We theorize that it may have something to do with the depth of the film emulsion vs the truly flat plane of the image sensor, i.e. the circle of confusion is "cheated" a bit more on film, but who knows for sure. One reason that I can theorize that you are seeing more out-of-focus closeups on TV in general is that with the current HD cameras (except for the Dalsa and the D-20, neither of which is yet being used for any network shows as far as I know) the operator is not able to give much focus feedback to the 1st due to the electronic viewfinder vs the optical viewfinder that we are accustomed to. In a multiple camera situation, the DP may not be able to properly monitor both cameras and give them focus notes, so it's a bit of a crap shoot.
__________________
Charles Papert www.charlespapert.com |
|
| ||||||
|
|