Jump to content


Photo

SDX 900 or HDV?


  • Please log in to reply
20 replies to this topic

#1 jcadge

jcadge

    New

  • Basic Members
  • Pip
  • 4 posts
  • Cinematographer

Posted 06 November 2005 - 12:51 PM

In the near future I'm going to be shooting a doc and planning on using the
SDX 900 with the Fujinon 13X4.5 HD lens. We may go out to film but most likely will master
on SD (digi beta) and onto DVDs. How does this approach compare to shooting with
one of the latest HDV cameras say the GY-HD100 with the Fujinon wide lens keeping
in mind we would eventually down convert the HDV to SD.
  • 0

#2 David Mullen ASC

David Mullen ASC
  • Sustaining Members
  • 19769 posts
  • Cinematographer
  • Los Angeles

Posted 06 November 2005 - 01:43 PM

The SDX900 is superior in every way except it's not HD resolution. Recording DVCPRO50, you've got less compression, 4:2:2 color, you've got 2/3" CCD's and pro lenses, more advanced control over gamma, etc.
  • 0

#3 Stephen Williams

Stephen Williams
  • Sustaining Members
  • 4708 posts
  • Cinematographer
  • Europe

Posted 06 November 2005 - 02:00 PM

The SDX900 is superior in every way except it's not HD resolution. Recording DVCPRO50, you've got less compression, 4:2:2 color, you've got 2/3" CCD's and pro lenses, more advanced control over gamma, etc.



Hi,

I think the SDX900 is probably the best SD camera available. However I have not seen a working example of the Progressive Scan DigiBeta, as it was not working at IBC.

Stephen
  • 0

#4 David Ross

David Ross
  • Basic Members
  • PipPip
  • 43 posts
  • Other

Posted 06 November 2005 - 11:03 PM

I agree with the above. I would really consider the work flow issues currently with HDV. The DVC50 format is solid and proven. Cinegama is also plus. You said you were going to use a HD lens. Be prepared for outstanding images....

Good shooting!
  • 0

#5 jcadge

jcadge

    New

  • Basic Members
  • Pip
  • 4 posts
  • Cinematographer

Posted 07 November 2005 - 08:24 AM

Thanks for confirming my feeling, incidentally the JVC HD 100 does accept the P&S
Adapter for pro lens but since this project involved a lot of hand held shooting I opted for the ENG wide zoom approach. The foundation underwriting this doc is now asking for a SD delivery, and they have a relatively low budget so my next question is, if we
did go the HDV route we'll have an edited HDV master, which we would down convert
for their initial SD needs. In the future if they change their minds and want a HD version
how do you think the footage shot with the SDX 900 and HD lens upconverted to HD would compare to the HDV master? How has your experience been upconverting SDX 900 footage to HD?

Thanks for your help.

Jeff Cadge
  • 0

#6 David Mullen ASC

David Mullen ASC
  • Sustaining Members
  • 19769 posts
  • Cinematographer
  • Los Angeles

Posted 07 November 2005 - 11:18 AM

What mode are you shooting the SDX900 in? 24P or 60i?
  • 0

#7 jcadge

jcadge

    New

  • Basic Members
  • Pip
  • 4 posts
  • Cinematographer

Posted 07 November 2005 - 04:28 PM

We will be shooting 24P.
  • 0

#8 Alexis Hanawalt

Alexis Hanawalt
  • Basic Members
  • PipPip
  • 18 posts
  • Director
  • Santa Monica, CA

Posted 15 November 2005 - 05:54 PM

I have a similar question -

I'm working for a company that's about to shoot a horror feature on either the new JVC camera in 24p HDV or on the SDX900. The current lack of support for editing 24p HDV is a major holdup as far as that goes, but I'm also curious about how DVCPRO50 footage looks when upconverted to HD? Anyone have any opinions?

I would usually argue strongly for the SDX... but if this goes theatrical (it probably will) I'm not sure if HDV wouldn't be better due to the higher resolution.

-Alex H.
  • 0

#9 Peter J DeCrescenzo

Peter J DeCrescenzo
  • Sustaining Members
  • 620 posts
  • Camera Operator
  • Portland, OR, USA www.peterdv.com Blog: http://herefortheweather.wordpress.com/

Posted 27 November 2005 - 02:00 PM

... how do you think the footage shot with the SDX 900 and HD lens upconverted to HD would compare to the HDV master? How has your experience been upconverting SDX 900 footage to HD?


Hi Jeff: I posted the following response to a similar question on another forum. (The DigiBeta footage I saw was of course originated as 60i, so I'd expect well-shot SDX-900 24p DVCPRO-50 footage to upconvert with even better results.):

=============
I've seen DigiBeta footage (SD of course) which was very carefully and professionally shot using an HD lens and then upconverted to HD via a high-end Terranex machine. When projected from HDCAM tape on a high-end HD projector it looked like very good quality HD -- but _slightly_ softer -- exactly as you'd expect.

I suppose the above scenario is one of the best case examples. If lower-quality SD footage contains more compression artifacts, or lens-induced color fringing, noise, inaccurate color, or other problems, these will of course show up in the upconverted result.

But it's also true that HD originated footage can also contain these same problems. A cheap lens on a cheap HD cam using a codec less advanced than HDCAM or DVCPRO-HD might not look as good (in certain respects) as high-quality SD upconverted to HD.

As always, it's garbage in, garbage out; same as it ever was.
==============

All the best,

- Peter DeCrescenzo

Edited by Peter DeCrescenzo, 27 November 2005 - 02:01 PM.

  • 0

#10 Mr. Macgregor

Mr. Macgregor
  • Basic Members
  • PipPipPip
  • 108 posts
  • Cinematographer
  • Los Angeles

Posted 28 November 2005 - 08:23 AM

Allthough JVC HD100 is HD resolution camera, images do not look very good. Color, definition, sharpness, very few people seem to like it.
On the other hand, try the new Canon XL H1. That really gives very good looking images.

Now, latitude and compresion is the problem. The Panasonic is better and we cannot argue that. Now what you have to consider is if a better resolution image is better than a less compressed image with better lattitude.

I think XL H1 HDV resized to PAL would be at least as good as DVCPro because color would be 4.2.2 if not better and the compression artifacts will disappear.

Edited by macgregor, 28 November 2005 - 08:25 AM.

  • 0

#11 Peter J DeCrescenzo

Peter J DeCrescenzo
  • Sustaining Members
  • 620 posts
  • Camera Operator
  • Portland, OR, USA www.peterdv.com Blog: http://herefortheweather.wordpress.com/

Posted 28 November 2005 - 11:24 AM

... I think XL H1 HDV resized to PAL would be at least as good as DVCPro because color would be 4.2.2 if not better and the compression artifacts will disappear.


I believe all HDV camcorders record 8-bit 4:2:0 color information to HDV tape. Converting this to PAL would not increase its quality.

If you're instead referring to outputting a "live" HD-SDI signal out of a Canon XL-H1 HDV camcorder (e.g.: _not_ recording/playing HDV tape) then yes, this live HD-SDI signal is likely to be 10-bit 4:2:2 digital video. Captured into an appropriate HD capture device (computer w. Kona/BlackMagic/etc. HD capture card, HD HDD recorder, and so forth) the resulting recording may contain more color info than HDV tape.

Unless you're referring to something else?

All the best,

- Peter DeCrescenzo
  • 0

#12 Mr. Macgregor

Mr. Macgregor
  • Basic Members
  • PipPipPip
  • 108 posts
  • Cinematographer
  • Los Angeles

Posted 28 November 2005 - 11:55 AM

I believe all HDV camcorders record 8-bit 4:2:0 color information to HDV tape. Converting this to PAL would not increase its quality.

If you're instead referring to outputting a "live" HD-SDI signal out of a Canon XL-H1 HDV camcorder (e.g.: _not_ recording/playing HDV tape) then yes, this live HD-SDI signal is likely to be 10-bit 4:2:2 digital video. Captured into an appropriate HD capture device (computer w. Kona/BlackMagic/etc. HD capture card, HD HDD recorder, and so forth) the resulting recording may contain more color info than HDV tape.

Unless you're referring to something else?

All the best,

- Peter DeCrescenzo



Exactly. HDV is 4.2.0. So it is pure logic, if you have an image 3 or 4 times bigger than PAL but at a lower color resolution, once you downresize you are increasing the amount of color definition for each pixel.

I am not technician, but perhaps someone around here could explain this better than me. The truth is that right now footage from a HDV sony camera shot with CF25 and downresized to PAL is almost giving 4.4.4 color sampling, and you can feel it once you start color correcting.

Edited by macgregor, 28 November 2005 - 11:56 AM.

  • 0

#13 Peter J DeCrescenzo

Peter J DeCrescenzo
  • Sustaining Members
  • 620 posts
  • Camera Operator
  • Portland, OR, USA www.peterdv.com Blog: http://herefortheweather.wordpress.com/

Posted 28 November 2005 - 01:53 PM

Exactly. HDV is 4.2.0. So it is pure logic, if you have an image 3 or 4 times bigger than PAL but at a lower color resolution, once you downresize you are increasing the amount of color definition for each pixel. I am not technician, but perhaps someone around here could explain this better than me. The truth is that right now footage from a HDV sony camera shot with CF25 and downresized to PAL is almost giving 4.4.4 color sampling, and you can feel it once you start color correcting.


I don't believe that's how it works. I don't think you can gain color info by downconverting HD to a "lesser" format. I could be wrong.

Refer to this thread, one or two posts down from the top:
http://www.cinematog...olorspace&st=15

If I'm incorrect, perhaps a wiser head could interject at this point?

- Peter
  • 0

#14 Mr. Macgregor

Mr. Macgregor
  • Basic Members
  • PipPipPip
  • 108 posts
  • Cinematographer
  • Los Angeles

Posted 28 November 2005 - 03:47 PM

4:4:4 SD. While "uncompressed" is naive, here's how it goes down. First define "H" as the horizontal resolution obtained from a 1/2-pixel shifted 3 CCD system. In theory, "H" is probably somewhere between 1440 and 960, but not less than 960).

60i modes: (effective resolution "H"x1080)*
Y: 1440x1080
Cr: 720x540
Cb: 720x540
Note that the Cr and Cb information are temporally offset by 1 field.

CF30/25 (effective information "H"x540)
Y:1440x1080
Cr: 720x540
Cb: 720x540
Note in this case the Cr and Cb come from the same temporal sample. Note also that the vertical resolution available for the CF modes is essentially the same as the effective resolution of the pseudo-progressive frame. In a sense you have "H"x540 4:2:2 going into a 1440x1080 4:2:0 container - which is enough to hold it (if not a little wasteful - there are twice as many luma samples stored as is necessary).

If we're generous and say "H" is in fact 1440, then in a CF image you have 1440x540 4:2:2. Scale that horizontally by a factor of 2, and you have 720x540 4:4:4. Do a little vertical interpolation and you've got 720x480 4:4:4. You certainly do much better than 4:2:2.

And seriously - if you do this down conversion properly and watch the uncompressed SD running of your hard drive - you'll be absolutely blown away by how good it looks compared to DV.
  • 0

#15 Michael Collier

Michael Collier
  • Basic Members
  • PipPipPipPip
  • 1262 posts
  • Gaffer
  • Los Angeles, CA

Posted 28 November 2005 - 06:11 PM

And seriously - if you do this down conversion properly and watch the uncompressed SD running of your hard drive - you'll be absolutely blown away by how good it looks compared to DV.


I have seen this, its amazing. Make sure to capture HDV from the start, if you convert in camera to DV, it will drop all information in favor of the 4:2:0 720x480

The extra latitude in color correction is great, also I have noticed that the camera overshoots NTSC standard. Images that look well leveled in the computer seem to blow out a little bit on NTSC. This means that if you take the level of the video down just a tad to put it into spec, you have opened up new headroom for color correction.

I'm not big on gama selection in feild. Usually I set it to match the gama on my computer screen with adobe gama correction software. This way when I bring it into the computer no additional setup is needed and I see it with linearity. Color correct with a WYSIWYG NTSC monitor plugged in and get a good TV color correction. Color correct on the monitor for film out color.

The only problem with the HDV models that I have seen is the lame lens designs included. But what should you expect. The lens on the SDX is more expensive than the whole HDV camera. I have always been a fan of the mini-35 HDV combo. The look that is imparted to the total frame frees you up as a photographer.

Bottom line of photography is directing the eyes of your veiwer. Make them see things the way you see things. I do this with DoF and using the environment to 'draw lines' to my intended subject. With the cameras I am using now everything is way to sharp. I hate to cheat DoF but every video camera on earth is 1/3" or 2/3" and iris' only go up so high. One day we may see a T.03 with a razor thin depth of feild, but I doubt it, so its all about increasing the target size.

And for those who have questions about the compression level in HDV, it doesnt seem to be a problem. I am finnishing up a feature length movie on the ZU1 (sony, stock lens) and it looks great. I have scoured several motion frames to find MPEG compression artifacts, but they are small and not obvious. MPEG has gotten better over the years.

I think an HDV with mini 35 and some prime lenses will look better than the stock SDX-900 in both HD and SD deliveries (after a little color correction).
  • 0

#16 Phil Rhodes

Phil Rhodes
  • Sustaining Members
  • 11943 posts
  • Other

Posted 28 November 2005 - 07:18 PM

Hi,

Yes, you could shoot HDV, downconvert it, and be able to claim that it's 4:4:4 standard def, or very nearly. You're not improving the absolute colour resolution, you're increasing the amount of colour information there is per pixel - and there's less pixels in the SD frame. You're effectively creating a lower resolution result so you can't see the flaws in the high res image. If the U plane of a 4:1:1 image is 640 pixels wide, representing a 1280-wide DVCPRO-HD image, and your output format is 720 pixels wide - well, you have a colour sample for more or less every output pixel.

You are throwing away a lot of luminance information, though. If I were to do this, my procedure would be to uncompress the HDV and post purely in that, then blow it down as the very last output step.

Theory would claim that a blown-down HDV image would actually contain less information than DVCPRO50 or digibeta, both of which record at a much higher data rate than HDV. In practice, however, I have seen it look much better. I think HDV looks vastly better than it has any reason to, and blowing it down allows you to average out noise and supersample everything twice, removing nasty edge artifacts and other undesirable processing, which is an enormous help quite apart from the improvement in colour reproduction per pixel.

Phil
  • 0

#17 Daniel J. Ashley-Smith

Daniel J. Ashley-Smith
  • Guests

Posted 28 November 2005 - 07:52 PM

Straight answer, I'd go SDX-900. I've seen footage from these HDV cameras and from the SDX-900, the HDV basically looked like the standard miniDV consumer camera just with a huge resolution. The SDX-900 had a different quality to it, film like.
  • 0

#18 Simon Wyndham

Simon Wyndham
  • Basic Members
  • PipPipPip
  • 146 posts
  • Camera Operator
  • UK

Posted 29 November 2005 - 04:42 AM

Hold on just a cotton pickin' moment.

4:4:4 will give you a higher chroma RESOLUTION etc, and is better for keying, but I cannot see how it gives that much more latitude. Ie you are still working with 8-bit colour. Certainly things might be smoothed out as a result of the downconversion, but if you took a grey graduation for example there are still only 235 shades available for legal video. You should still only be able to push it the same distance before you start to get banding.

Anyone wish to put up some extreme colour corrections to actually illustrate the limits between the two?

Now if the colour was sampled and recorded at source in 10-bits then I'd be impressed.
  • 0

#19 Phil Rhodes

Phil Rhodes
  • Sustaining Members
  • 11943 posts
  • Other

Posted 29 November 2005 - 06:31 AM

Hi,

> 4:4:4 will give you a higher chroma RESOLUTION etc, and is better for keying, but I cannot see how it gives
> that much more latitude.

It doesn't. I don't see where someone's claimed it does?

> Ie you are still working with 8-bit colour.

Yes, but that's got nothing to do with latitude.

What you'll get if you colour correct a blown-down frame, from my experience of this sort of thing in the past, is much lower levels of noise and compression artifacts which will make it easier to achieve clean results, although I'd still caution about the extremely high compression that's applied to HDV material in the first place.

Phil
  • 0

#20 Simon Wyndham

Simon Wyndham
  • Basic Members
  • PipPipPip
  • 146 posts
  • Camera Operator
  • UK

Posted 29 November 2005 - 08:26 AM

"The extra latitude in color correction is great, also I have noticed that the camera overshoots NTSC standard. Images that look well leveled in the computer seem to blow out a little bit on NTSC. This means that if you take the level of the video down just a tad to put it into spec, you have opened up new headroom for color correction."
  • 0


Metropolis Post

CineLab

Media Blackout - Custom Cables and AKS

Aerial Filmworks

rebotnix Technologies

FJS International, LLC

Willys Widgets

Gamma Ray Digital Inc

Technodolly

Tai Audio

The Slider

Opal

Glidecam

CineTape

Paralinx LLC

Rig Wheels Passport

Wooden Camera

Broadcast Solutions Inc

Ritter Battery

Visual Products

Abel Cine

rebotnix Technologies

FJS International, LLC

CineTape

Metropolis Post

Ritter Battery

CineLab

Abel Cine

Rig Wheels Passport

Aerial Filmworks

Tai Audio

Opal

Technodolly

Glidecam

Willys Widgets

The Slider

Gamma Ray Digital Inc

Media Blackout - Custom Cables and AKS

Visual Products

Paralinx LLC

Wooden Camera

Broadcast Solutions Inc