Jump to content

Varicam vs. PAL Digital Beta


Guest Pete Wright

Recommended Posts

Guest Pete Wright

Varicam records 960x720 pixles, 8-bit, compressed about 6.5:1 (DV is 5:1)

 

Digittal Beta PAL records 720x576 pixels, 10-bit, compressed about 2:1

 

Which one would be better for film production, to be projected in a theater.

 

Is the slight extra resolution more important, or is it the overall greater picture quality, or is it about the same.

 

Let's assume that same HD lenses would be used on both cameras.

Link to comment
Share on other sites

  • Premium Member

Wow:

 

Should Varicam be called HD?

Low Cost HD and Near HD Quality

SDX900 vs. F900 for Film Production

Newly Announced Canon XL2 vs. Varicam

And now Varicam vs. PAL Digital Beta

 

Whats next, Varicam vs. Mothra?

Link to comment
Share on other sites

Guest Pete Wright

Lumier's info that I included in another (later) thread makes this debate unimportant. The next generations HDV camcorders that will be available soon supposedly have excellent image quality and true 1280x720 pixels.

 

Pete

Link to comment
Share on other sites

  • 2 weeks later...
Guest Frank Miller

That's right. It is the Lumier guy at the dvinfo HDV thread that has the inside info on the new HDV cameras, that was his name.

 

Digi Beta (PAL) is compressed 2x, Varicam 7x, the recorded resolution is about the same. (Varicam cuts down 25% horizontal resolution when recording.) I've seen footage from both. Digi Beta picture is a lot better. And you can't even get Varicam in Europe.

Link to comment
Share on other sites

  • Premium Member

Hi,

 

In the interest of factual correctness - wow, what a new concept around these parts:

 

> Digi Beta (PAL) is compressed 2x, Varicam 7x

 

The comparison is not relevant because the mathematics used in each case is wildly different. Notice that Sony call their "compression" bit-rate reduction. It's based on the Huffman algorithm. Theoretically Huffman allows accurate recovery of the orginal data; you do theoretically lose something in the YUV conversion and the image is subsampled 4:2:2 but the compression is practically transparent. I posted extensively on Huffman a while ago.

 

> the recorded resolution is about the same

 

Not necessarily. Varicam uses a compression algorithm based ultimately on standard-def DV (as does HDCAM.) The algorithm uses a truncated discrete cosine transform with, critically, variable quantisation tables across the image. The degree to which picture detail is affected by this scheme is dependent on the entire image content and no simple comparison can be made. In general, DVCPRO-HD footage will appear to have considerably more resolution than Digital Betacam.

 

> And you can't even get Varicam in Europe.

 

There are three in London, considerably more in Paris. Panasonic UK certainly don't market them for narrative film production, which may be where this misconception comes from.

 

Phil

Link to comment
Share on other sites

  • Premium Member

What?s next: 16mm is better than 35mm?

 

I think the conversation here is far to simplistic. Slinging around resolution and compression schemes is ignoring the unique niche each format has. Sony?s progressive HDCAM is a logical choice for feature film makers who need maximum resolution for projection. It is, however, expensive to shoot and edit. The Varicam, and DVCPROHD offer a lower cost solution and the unique ability to over and undercrank. For commercial shooters like myself, variable frame rate is far more important than a few extra lines of resolution. If I was only concnerned with resolution, film would be the obvious choice.

 

I think it is time to look at each format and discuss how to maximize its potential. Remember, filmmaking is a business and budgets dictate the format you will shoot. As cinematographers, it is our job to maximize what each format is capable of, whether it be Film, HD, SD, or a Fisher Price pixel vision camera.

 

Chris Bell

Cinematographer

www.spotshooter.com

Link to comment
Share on other sites

  • 2 weeks later...

PAL Digi Beta is an interlaced format, and its vertical resolution (in Field mode) is about 400 lines PPH on a static subject. A Varicam will shoot 720 progressive, so it's vertical resolution should be much better than Digibeta.

 

There is a way to test this. Shoot a slant edge, or ISO 12233, target and analyse the resulting image via a frame capture with free software, called SFRwin. the resulting spacial frequency response is an accurate way to measure the resolution of the whole camera system.

 

To get a higher resolution progressive picture from a Digibeta, you need to use a mechanical shutter. The shutter would need to be closed for one field time, and the camera set to "frame" (EVS) mode in the menu. The resulting image would be, effectively, a full progressive frame recorded as 576-25PsF, with 100 or so additional lines of vertical res.

 

If you look into the engineering menu section of the Digi Beta manual there is a reference to an item called FRM Shutter. Sony told me they did not end up making this accessory.

Link to comment
Share on other sites

  • Premium Member
Are they storing charge on the CCD between fields?

You could think of it that way. What they're doing is integrating over full progressive frames instead of fields, clocking data out at the frame rate instead of the field rate. In effect, it's 576p, which is not so very far from 720p.

 

As for whether 720p can be called high definition, if 1080i is HD, then so is 720p. Because of the interlace loss, 720p is actually slightly higher resolution than 1080i. It just doesn't look like that on 1080i CRT's.

 

Then along comes Tiffen with some fractional pro-mists, and we should all see how silly this whole debate is.

 

 

 

-- J.S.

Link to comment
Share on other sites

If they're integrating over the entire frame period, why d'you need the shutter?

 

Sorry, this isn't making sense to me.

Hi Phil,

 

When you put the Digi Beta into "frame" mode it reads out alternate lines to alternate fields, unlike field mode where every line is read out , and alternately combined, every field.

 

In frame mode the vertical res goes up, but so does the blur because each line is now "exposed" for a 25th of a second, and the fields overlap for 1/50th of that time. The electronic shutter (extended clear scan) isn't of use here because that shutters each field separately - it's the 1/50th overlap that we're interested in.

 

To get a 25PsF image out of the Digi Beta you need to capture the frame all at once in 1/50th of a second and then record it as interlace. You can do that by physically stopping any light from reaching the CCD at the point the first field is read out (with the camera in frame mode) by using a mechanical shutter. When the second field is read out it consists of pixels that haven't accumulated a charge since the first field time, they've just been sitting there in the dark. The only down side is you lose a stop.

 

End result, you've got an interlaced image where all the lines were captured together in 1/50th of a second - 576PsF

 

The next question I know you'll have is, "where do I get the shutter?" Unfortunately Sony don't make it.

 

But, it should be possible to put a mechanical shutter in the lens, perhaps in place of the 2x convertor. A simple 180 degree shutter timed to the camera sync would do the job.

Link to comment
Share on other sites

  • Premium Member
If they're integrating over the entire frame period, why d'you need the shutter?

I really don't know that you absolutely have to do it with a mechanical shutter. It seems like it should be possible to do it by reading a full frame to memory and dumping charge, then reading interlace from memory to tape. Video cameras already achieve very short exposure times by clearing the CCD, then reading out what it's picked up only a few thousandths of a second later. This is getting to the feeble fringes of my knowledge, we really need somebody like Larry Thorpe to explain it.

 

Perhaps you could do a cheap quick and dirty progressivization by putting a mechanical shutter in front of the lens and syncing it to black out one field. But even that would require that the camera not dump charge between fields. Too bad Larry's retired now, does anybody know where we can find a Larry Thorpe type?

 

 

 

-- J.S.

Link to comment
Share on other sites

  • Premium Member

Hi,

 

I'd try it, other than that I'd need a shutter of at least 8" diameter to cover the front of my lens, spinning at around 750rpm assuming a two-bladed type, and I value the integrity of all my fingers!

 

Phil

Link to comment
Share on other sites

It seems like it should be possible to do it by reading a full frame to memory and dumping charge, then reading interlace from memory to tape.

 

If it did that it would be a PsF progressive scan camera. The Digi Beta doesn't do it, that's the problem.

 

Perhaps you could do a cheap quick and dirty progressivization by putting a mechanical shutter in front of the lens and syncing it to black out one field.

 

This is exactly what you need to do, set the camera to "frame" and blank one field time. To prove it works I just shot through the gate of an Eclair ACL with the lens and mag off (and yes, it was a very dark picture with soft edges but it proved the point). A 16SR would work too.

 

To do the test you need the shutter to run at 1/50th synced to black one field. Leave the camera in field mode and just start and stop the shutter until you get maximum flicker, then switch to frame mode via the menu and there it is - no flicker and "film" type movement.

 

The other advantage with a real shutter is that you get the same temporal image capture as film, i.e. each frame blanked by the action of a mechanical shutter rather than an instantaneous electronic readout of all the pixels.

 

By the way, I love the term "progressivization" We could call it the "physical optical path interruptor that offers progressivisation to location interlace aquisition".

Link to comment
Share on other sites

  • Premium Member
I have somewhere an old silent Simplex projector which uses a shutter after the lens, two blades, and about that size.

 

 

Or maybe an old Simplex E-7, which had a double shutter, one behind the aperture, and one in front of the lens?

 

e7.gif

Link to comment
Share on other sites

  • Premium Member

Hi,

 

> rather than an instantaneous electronic readout of all the pixels.

 

Most CCDs don't read out anything like instantaneously. Eventually all the columns end up coming out of one output amplifier, and it takes time. The exact characteristic of it, and therefore the spatial rendering of temporal changes, is dependent on the CCD.

 

Phil

Link to comment
Share on other sites

Most CCDs don't read out anything like instantaneously. Eventually all the columns end up coming out of one output amplifier, and it takes time. The exact characteristic of it, and therefore the spatial rendering of temporal changes, is dependent on the CCD.

 

I understood that the transfer of charge from the photosites to the optically masked image storage area occurs simultaneously and almost instantaneously. If the charge transfer was to have any effect on spacial rendering of temporal changes these would most likely manifest themselves as smear and lag. True, the readout takes a while, but the CCD charges travel down the columns and out the amp etc. in the dark, i.e. without accumulating additional charge.

 

If you shutter the image just before readout the characteristics of the CCD don't matter anyway. A mechanical shutter will result in a more "filmic" motion capture than a 25P camcorder.

Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
×
×
  • Create New...