|
|||||||||
|
Thread Tools | Search this Thread |
December 6th, 2006, 06:59 PM | #1 |
Regular Crew
Join Date: Jan 2006
Posts: 76
|
Problems with Z1 images - HELP -
Hi all,
I always shoot on HDV and then capture via YUV out directly from the Z1E camera. So far so good, but ... I've noticed a very strange thing. On some "colour ramps" I get a lot of blocks on the image. If they are on a calm background, it's not a big problem (frame1), but if they are on the moving part of the image it's a disaster !!!!(frame2). I'm sure that during the recording there's not this effect on the image. The uncompressed video that comes on Y/C or YUV out is ok. But the image recorded on the tape have this huge problem. ANY HELP ? - Is this produced by the MPEG HDV compression ? - Is it made by the HDV -> PAL convertion (?) - Is my camera broken ? - If I capture from FW is the image better ? ____I still use a TARGA 3000 + Premiere 6.5 + Win2K, and the FW does not work properly. ONLY during 2007 I will go to a new NLE with HDV. Any advice for the future ? Many thanks for any input on this. Carlos P.S. - I attach 2 frames with the problem |
December 6th, 2006, 09:26 PM | #2 |
Regular Crew
Join Date: Oct 2006
Location: Fort Worth, Texas
Posts: 83
|
Carlos,
First off, you should definitely capture via firewire. This is a straight digital transfer. By using the component (YUV) output you're capturing a digital source that has been converted to an analog signal. There will be a huge difference in the quality of the captured file. Those blocks you are referring to look like it may be caused by the HDV MPEG compression. BUT, it may not be as noticeable if you captured via firewire since it is transferring EXACTLY what is stored on the tape. I'm not sure, but maybe during the digital -> analog conversion in camera process may somehow cause the compression artifacts to be more prominent? I may be wrong. Another theory I may have is the format in which you're capturing the video. I noticed the images are in the 4:3 aspect ratio... are you shooting HDV and then cropping off the sides to fit into a SD 4:3 timeline? Or maybe since you're capturing via an analog connection, the quality setting on your NLE/Capture software could be turned down. So it may not even be the compression from the HDV, but the compression method being used while capturing... |
December 7th, 2006, 02:43 AM | #3 |
Major Player
Join Date: Mar 2004
Location: Toronto, ON, Canada
Posts: 479
|
I think what you're seeing is fairly normal. 1440x1080 HDV at 3.5MBPS is highly compressed so not every pixel will look amazing.
Where does your component out signal go? (ie- to a computer/laptop capture card, deck, etc.)
__________________
Mark Utley |
December 7th, 2006, 05:53 AM | #4 |
Regular Crew
Join Date: Jan 2006
Posts: 76
|
Thanks for your first answers.
I shoot in HDV and then output to PAL cropping the edges (process done inside the Z1 camera) I capture in YUV uncompressed in my TARGA3000. The same way I capture BETACAM SP or Digital, without a single block. Capturing in YUV uncompressed preserves the image quality. I believe that this is related with the MPEG compression. But when I shoot outdoors, I get NOT this kind of problem. At least, I can't see it. If I capture via firewire to the TARGA 3000 it generates only DV files, and I believe that capturing YUV could be a little better than that. Obviously that the YUV out from the Z1 it's a digital to analog conversion. But when I feed my Targa3000 with YUV, the board also makes an analog / digital conversion WITHOUT image artifacts. I will try to capture in a NLE with HDV and see if the problem is on the signal recorded or in the HDV --> PAL conversion. BTW. a simple question. When I shoot, I send the Y/C or YUV signal to a SONY 9" monitor. In this image there are NO blocks, probably because I'm using a video signal before the HDV compression. If I monitor via the FIREWIRE out (HDV>DV) I should notice the blocks? Is this a better way of working? Any advice on how to feed a monitor from a FireWire signal ? Carlos |
December 7th, 2006, 07:59 AM | #5 |
Major Player
Join Date: Mar 2004
Location: Toronto, ON, Canada
Posts: 479
|
I would guess that because you're downconverting in-camera, it's going through the HDV->DV process and being compressed. Is your system able to accept component HD?
__________________
Mark Utley |
December 7th, 2006, 08:11 AM | #6 |
Major Player
Join Date: Feb 2005
Location: San Diego, CA
Posts: 216
|
I must admit that I also occasionally notice blocking in dark, monochromatic colored areas with minimal luma gradients. Even an underexposed blue sky gradient will sometimes show this happening also. Luckily, we're not usually underexposing the sky... :)
As a result, I've taken to trying to avoid these types of colored areas. You can do so by lighting it more brightly, adding texture (either with light patterns and/or by changing the shot and/or wardrobe), etc. If you're shooting interviews, try using a background with more visible texture to it if you can. I noticed this blockiness most on one of my early shoots with a green wall: the wall had all kinds of blockiness. On that shoot in particular, I used Cinegamma2 with Black Stretch On and I've always wondered if that had anything to do with it. What color settings are you using for your shots? Also, my blockiness problem on my shoot was exacerbated by my moving between After Effects Std (which operates in 8 bit RGB) and PremierePro with the Cineform AspectHD codec (which operates in 8 bit YUV) a few times. That simple conversion added noise/blockiness in the backgrounds even more. This probably isn't your problem, but it's something for you and others to possibly look out for also. Finally, I notice this more on my LCD editing monitors than on any CRT output (e.g., comparing DVD output). I suspect that my monitors are doing some color shifting that makes low level quantization steps more noticeable. Bill |
December 7th, 2006, 09:09 AM | #7 |
Regular Crew
Join Date: Jan 2006
Posts: 76
|
Mark Utley - "Is your system able to accept component HD?"
No. I can only import composite, Y/C and YUV in PAL (SD) The FireWire inpout from the Pinnacle TARGA3000 never worked properly. I will ask a friend that has Premiere Pro 1.5 and see if it can import directly from the FW port the m2v stream to the timeline. Maybe the signal recorded on tape is a little better. Not very confident on this... ------------------------------------------------------------ William Gardner I usualy use Cinematone OFF or 1 and Black Strtch OFF. Never the Cinematone 2. Feels to dark for me. Interviwes always done with MANUAL on all the items. I notice that the blocks are on the image but when I do come color correction (a bit more colour or I touck the levels) they appear in "full glory". Because of the TARGA3000 i still have to edit with Premiere 6.5. Really not a big problem for me. Specialy with Betacam tapes. I import on DV, edit and in the end re-import the used clips in YUV. The final quality is great. I use Combustion 4 for CC. With the Z1 camera, some problems have appeard. I can not control the camera with 9-pin control (tried some IEEE/9pin converts but they are not frame accurate) So I have to import directly to YUV (70G / hr)... In general I am very pleased with the image from the Sony Z1 camera. But Yesterday I've noticed these blocks .... urgggg This week-end I will try to shoot some images for test. William Garner, please a question: When you do interviews How do you do for monitoring the image? I use a 9" SONY monitor (PVM-9041) in Y/C or YUV. I know that these signals are taken from the camera BEFORE the HDV compression. If I monitor with the FireWire after the HDV>DV conversion, I should be abble to notice these kind of things. What do you think ? Should I start using a Laptop and DV Rack ???? Maybe. Carlos |
December 7th, 2006, 02:21 PM | #8 |
Major Player
Join Date: Feb 2005
Location: San Diego, CA
Posts: 216
|
Well, I often monitor using just the LCD on the Z1.
However, I have monitored using a laptop with DV Rack (w/ HDV support). If you have a full res monitor, this is great for focus. However, at least on my laptop, it wasn't the full frame rate, so it actually made the HDV artifacts more noticeable because I didn't have the full frame rate to create the temporal smoothing that makes some of this stuff less noticeable. You can try the fully functional DV Rach w/ HDV support demo for a month or so for free, so try that out and see if it works for you. I've also monitored using HDVsplit. This gave me a full frame rate, full res monitor on my laptop and was free (!!!). It doesn't have many of the useful bells and whistles that DV Rack has (like waveform monitors, etc.), but for focus and exposure monitoring and looking for blockiness it's pretty good. Good luck, Bill |
December 8th, 2006, 02:13 AM | #9 |
Major Player
Join Date: Mar 2006
Location: Helsinki, Finland
Posts: 393
|
Yes, it's the mpeg2 encoder in the camera. Basically, it doesn't look good under certain conditions.
|
| ||||||
|
|