- This topic has 25 replies, 7 voices, and was last updated 14 years, 9 months ago by .
-
Posts
-
Quote:we are buying CANON XL-H1… i think i saw that output to HD… and i was asking myself if we can skip compression… by the way i didn’t know that HD compresson is worst than DV… that really suprissed me…
DV and HDV share the same data rate (25mbps) HDV compression ratio is around 22:1 (Mpeg 2 which is harder to edit and requires a lot of cpu power) while DV is “only” 5:1 (using jpeg compression).
The XL-H1 has effectively a HD-SDi output so you can record with the appropriate hardware a 1080 (i don’t know if it’s interlaced or progressive on this cam) with a 4:2:2 subsampling. You won’t be able to get 4:4:4 as it’s not a dual link.Quote:well also i have send a letter asking for a propositin to reflecmedia.com but still no answer… can anybody tell me what is the price around… i mean is it really hot or is it OK?That kind of system is extremely expensive (around 3000€ for a 3x3m screen if I’ve got a good memory). It’s interesting because you are sure to get a perfectly lit screen, but I don’t think it’s a good option for you because this system isn’t used by a lot of professionnals. It’s better to teach on a good old green screen.
Quote:well, it’s also true that there will be students lighting a stage, the “art” students that most of the time lear how to paint and draw…Don’t worry, it’s not as hard as it seems. With a good teacher behind them that will be ok.
Quote:also can anybody explain me what are the “space lights”? and where to get them???Space lights are large cylinders covered by a diffuse material with an intense lightsource inside and they produce a very diffused light on 360 degrees.
They are quite big and needs to be placed quite high. So check first if your studio can support them.
Otherwise, there are plenty of solutions to lit your screen.Quote:whith a conversion to 4:2:2 can i preserve the HD resolution???4:2:2 is the subsampling of the image. It’s completely independant from the image resolution. 4:2:2 means that the luminance channel is fully digitalized while the two chrominance channel are half digitalized (only one vertical line out of two are preserved). In fact it’s a bit more complicated but you get the idea.
In 4:1:1 (DV and HDV) only one vertical line out of four is recorded!
In 4:2:0 it’s a bit different, you get one line out of 2 vertically AND horizontally too.
So because chroma key is based on chrominance channels of the image, you can get huge artefacts such as an enormous aliasing which can be even worst if there is a compression applied to the image.
This is not anyway a problem caused by the “lighting guys” as it’s said in the horrible shake training DVD set of CMI.Quote:so many questions… i mean technology goes too fast that is for sure… yeah, and we work in AF and FUSION 5… we have DVMatte pro 2 as a plug-in. Any sugestions about keyeing software for DV/HDV??? i have read allready the “art of keying” here on FXGUIDE and it helped me so much, but maybe you have a experience with all those keyers…You can get a good key using any software and any format (except HDV). The only thing you need to know is that if you get sub-sampling artefacts, you can minimise them by going to the YUV components of the image and applying a blur to UV channels.
Quote:and just one more little question: how about neon/cool lighting for the green room??? is it better or not?Neon are not really suitable for that kind of use. The amount of light they produce is still too low for that. Don’t use normal neons, they produce a slightly green light in a very thight bandwidth so the result on the camera will be totally unusable. You need specialised neons, which produces light in a more wide bandwidth and with a calibrated color temperature. Otherwite take care with that as they can produce a flicker effect.
Dont’t mix lightsources that have a diffrent color temperature or use gels.Quote:PS: a long time a go (5years): i remember trying to get a good composite on dv with some friends and finally we decided to do a lumaKey in my living room on the white wall… what we did is: we lit the wall whith big neon ligting until it did not burn on the screen on camera, and our actor we lit with normal light but a little bit less exposed so that the highlights are not white… and for that time it was a good key… well just a little amaterish experience.Because the luminance channel of DV is not subsampled, you were able to work on the totality of the image and get a pixel-accurate key!
Quote:The XL H1 also features uncompressed HD-SDI (SMPTE 292M) and SD-SDI (SMPTE 259M) output…
what do i need more to shoot on HardDisk with no compression??? and use it in editing???You can make a live capture on a computer. I suggest you an apple G5 with FCP 5, a decklink HD card and if you want to stay uncompressed a big and very fast storage system (a huge of an xserve).
Otherwise, I think that with the decklink HD you can capture and compress in realtime to a DVCPro HD codec. It’s 8bit and a bit more compressed than DV, but you still have 4.2.2 and excellent results at the end for only 100Mbps data rate!FLGB wrote:DV and HDV share the same data rate (25mbps) HDV compression ratio is around 22:1 (Mpeg 2 which is harder to edit and requires a lot of cpu power) while DV is “only” 5:1 (using jpeg compression).Weellll, DV isn’t actually JPEG, it’s more clever and more suited to video. And MPEG2 on its own should need less CPU power, but you’re right that when you start needing to edit or even just seek and scrub in it it becomes a pain in the ass and slows right down. Final Cut Pro 5 has fully native HDV support, maybe it’s easier now? Anyone used it?
I’m just nitpicking here, it’s all good advice 🙂
There are several easy ways to get where you need to go:
First thing is in the camera… make sure you get the berst color out of your camera by setting any color preferences in there. Sony has a color enhancer that really gets close to 4:2:2 after processing. It wouldnt surprise me if your camera does too.
THEN after lighting and all that jazz…
when its on DV tape.. and your using the mac… try the program hdvxdv it will bring it in and get yout to 4:2:2 dvcpro at teh end..
OR what i do, is in SHAKE, i file in , deinterlace with interpolate, (NOT BLUR) then use bits node to take it to 8, 16 or float, then output as tiffs. NOW, you can do several things here. CG on top. Color grade then CG. OR cg then colorgrade.
Its a personal preference. When scaling to 720p, make sure you use a mitchel caling algorythm found in shakes resize node.
Your going to get killer footage. By turning down the sharpness in the camera, you almost never see any artifacts. Turning on color s gets more color to play with in post also. IF YOU AR TRYING TO DO DAY FOR NIGHT shots.. leave color s off. There isnt much color at night except blue… so its easier just to bypass that.
If you capture to FCP, well export out as dvcpro 60 i then bring movie into shake and do above.
You will be surprised what you get. the reason people talk trash about hdv is that they just hear rumors about the technology. But never played with it long enough to find the sweet spots. (As with anything)
You can regraing to give it a film look and seal the effects to gether later… but id just do a pulldown on it and see what you get first.
Quote:First thing is in the camera… make sure you get the berst color out of your camera by setting any color preferences in there. Sony has a color enhancer that really gets close to 4:2:2 after processing. It wouldnt surprise me if your camera does too.The color enhancement features of the sony cameras only affects grading and can be really dangerous if the operator doesn’t take care about them. They are used to give a sort of cinelike look but that’s all. I don’t recomend the use of such features for green/blue screen shootings as these colours could be altered and cause problems later. And it’s not because you’ve set your gamma or hue settings in some fashion that you’ll get a 4:2:2 subsampling in HDV…
Quote:THEN after lighting and all that jazz…Never; always lighting first, camera settings second. Once you lighting has been set up, THEN, you’ll be able to make your white balance, iris aperture, place your flags, etc and then re-adjust the lighting if necessary.
Quote:when its on DV tape.. and your using the mac… try the program hdvxdv it will bring it in and get yout to 4:2:2 dvcpro at teh end..First, if you compress an HD footage to DVCpro HD, you’ll loose image quality as this recompression is LOSSY (it’s like if you were loosing a generation), then, HDV is 4:1:1 or 4:2:0 it means that the three quarter of the colors information is definetly lost, it’s not because you transfer it to a 4:2:2 codec that the information will re-appear by some kind of magic. Then, it’s completely unuseful to convert a HDV video to DVCpro HD format unless you want to use the benefits of the intra-frame compression of this codec. The HDV codec is perfectly handeled by quicktime and FCP.
And HDVxDV will not resample the image if you don’t change the size of the image to a much smaller one.
If you want to resample and smooth your color components, do it in your compositing app.Quote:OR what i do, is in SHAKE, i file in , deinterlace with interpolate, (NOT BLUR) then use bits node to take it to 8, 16 or float, then output as tiffs. NOW, you can do several things here. CG on top. Color grade then CG. OR cg then colorgrade.HDV is a 8bit format. It’s not because you place a bit node that it will magically become a 32bits float footage!
When doing color corrections, you can set your software to do the calculations based on a certain bit depth, but it will not make a 8bit footage a HDR footage!Quote:Its a personal preference. When scaling to 720p, make sure you use a mitchel caling algorythm found in shakes resize node.Why would you scale down a 1080p video to a 720p one?
Quote:Your going to get killer footage. By turning down the sharpness in the camera, you almost never see any artifacts. Turning on color s gets more color to play with in post also. IF YOU AR TRYING TO DO DAY FOR NIGHT shots.. leave color s off. There isnt much color at night except blue… so its easier just to bypass that.I’m not sure if after all these modifications and generation losses the footage is stil correct…
The sharpness setting in the camera should only be used to set the focus correctly.
You can’t get more color room by using any setting unless you are using a magic camera.
In fact from a human point of view a night shot is only less saturated (in the darkness, only the rods of our eyes are working). The blue color is added as a style since decade and is accepted by the audience.Quote:If you capture to FCP, well export out as dvcpro 60 i then bring movie into shake and do above.Take care, adding a new compression to HDV footage will degrade the image. It’s better to use a lossless codec.
Quote:You will be surprised what you get. the reason people talk trash about hdv is that they just hear rumors about the technology. But never played with it long enough to find the sweet spots. (As with anything)In fact HDV is not a format to use in post-production at all. There is too much compression, the sub-sampling is not adapted for, etc. See by yourself :
http://www.videomac.org/HDV1080i-Z1.zip
(It has been shot at Marrakech in Morocco).First off, this camera person has not properly set focus for this shot.
that is why its blurry and useless.
DESPITE other settings that look off.
I will be releasing footage on the 14th that will show settings that are good.
Dont worry, your properly lit green room will not have anything in common with this piss poor craftsmanship.
Hi there!
This is my first post. I want to share a few of my experiences.
genido wrote:…we have DVMatte pro 2 as a plug-in. Any sugestions about keyeing software for DV/HDV??? i have read allready the “art of keying” here on FXGUIDE and it helped me so much, but maybe you have a experience with all those keyers…I recently did an on set greenscreen test with SONY HDR-FX1. I directly fed the video to my PC through iLink (being aware of that this does not bypass the HDV compression) to premiere Pro 1.5.1 then tried to pull the matte in AE6.5 with RedGiant Software Primatte 3. It was more than I expected! I was able to pull a good key even with the Primatte’s default settings.
Primatte 3 has a built in DV/HDV video enhancer switch which dramatically minimizes the blockyness of the keyed composition.
Above all I’m satisfied with it!
Hope this’ll help u! 😮
Kapilp
Subsampling patterns are explained here but I guess there are better explanations around. It just means the “brightness” is stored with higher resolution than the “colour” part in various different and not terribly important ways.
As well as chroma subsampling like that, HDV also compresses the entire picture fairly drastically, so yes, the luminance channel is compressed.
Y Cb Cr / YUV is a form of compression in itself that splits the colour into luminance, blue and red, creating green from the subtraction of the former two.
You will unequivocally loose some info.Digital Cinema Cameras capture in Pure RGB which means they get the info for every colour natively and not through some equation…
Only ultra high-end cameras can do that such thing. e.g. the SONY HDW-900:
http://bssc.sel.sony.com/BroadcastandBusiness/DisplaySubCategory?m=10014&p=2&sp=19&s=31&cpos=3You don’t nessecarily lose information. If you go from 4:4:4 RGB to 4:4:4 YCbCr in 10 bit or higher I defy anyone to see any visual difference 🙂 Even in 8 bit, well, you can try with those Sparks I posted in the Effects forum – chain a whole bunch together in Batch, nothing much happens.
Technically yeah, YCbCr is a much larger colour space so you have poorer symbol utilisation so you add quantisation noise when you convert to it, but not, like, much 🙂
It’s the chroma subsampling that does the real damage.
- You must be logged in to reply to this topic.