HV20

[Tahmazo]

is the Canon HV20 a good buy nowadays?

 

[Reportage]

if its freaking cheap...yes.

 

[Bamboopictures]

HV40 is true progressive vs HV20/30 de-interlaced 25F.

 

[jaegersing]

HV40 is true progressive vs HV20/30 de-interlaced 25F.

As far as I know, HV20/30 is not deinterlaced. The image is captured as a complete frame, but then it is segmented into 2 interlaced fields for distribution as a 50i video signal, AKA PsF or progressive segmented frame. Most NLEs can recover the progressive frames back without any problem.

Again, as far as I know, the HV40 is different in that it keeps the video signal progressive all the way, similar to the 25P output from the XH-A1 (which, just to confuse matters, is derived from a deinterlaced image).

 

[brianseah]

Perhaps a question based on experience. Have you tried shooting in interlace and progressive in the same resolution and compared them?
For me, I know it is very very different.

 

[jaegersing]

Perhaps a question based on experience. Have you tried shooting in interlace and progressive in the same resolution and compared them?
For me, I know it is very very different.

I agree they are very different, mainly in the smoothness of motion (interlaced is smoother). Or did you mean something else?

 

[magicdragon]

This needs further more detailed explanation - esp on how it affects the final look?

 

[jaegersing]

This needs further more detailed explanation - esp on how it affects the final look?

If you want a lot of detail you should search on the web, there is loads of stuff written about this subject. But briefly, here is my take on Progressive vs. Interlaced.

Inlerlaced shooting (In PAL countries) captures 50 shots per second, so the time between shots is very small, and this means that when you play back the video, any movement of the subject will look quite natural. Also, when you pan the camera, provided you do it smoothly, the movement will look smooth upon video playback.

Progressive shooting captures only 25 shots per second, so if the subject moves too quickly there will be a juddery, strobing kind of appearance to the movement, instead of a natural motion. This is simply because the time between captures is too long. Also, if you try to pan the camera the same speed as you would for interaced shots, it will look very jerky, no matter how smoothly you operated the tripod.

So why do people like to shoot progressive? Main reason is that the frame rate is similar to traditional film cameras, and there is a certain "feel" that this frame rate gives to movies. There is a hope that by using the same frame rate as film, the video will also look like film. There is some truth in this, but there is a lot more to "film look" than just shooting progressive.

So what about the jerky movements? Well if you want to shoot progressive properly, you need to use similar techniques to what are used in film production. So, for example, if the subject is moving across screen, pan with him to track the movement. This will make the background judder instead of the subject, but since the background is most likely out of focus, most people don't notice it. As long as the subject looks good everybody is happy.

Another example is a camera pan across a landscape. To make it look good in progressive, you either have to pan very slowly or very fast. A normal type of pan that you would do with an interlaced videocam will be very jerky when shot in progressive.

You also can reduce the motion judder effect by choosing camera angles that have the motion going towards or away from the camera, in preference to movements across the screen.


Apart from the aesthetic quality of shooting in progressive, there are also technical reasons why you might want to. For example, if you have a scene with a lot of fine horizontal lines, or if you are shooting some graphical elements that contain fine lines, shooting in interlaced will make the horizontal lines flicker on and off which looks really horrible. It is much better in such cases to shoot progressive which does not show flickering.

 

[magicdragon]

You misunderstood me (I should have been clearer). I was referring to: differences between HV20/HV30 vs HV40 all in in 25p mode (all PAL cameras). Could you expound on what you mentioned and specifically how it affects the final video look?

As far as I know, HV20/30 is not deinterlaced. The image is captured as a complete frame, but then it is segmented into 2 interlaced fields for distribution as a 50i video signal, AKA PsF or progressive segmented frame. Most NLEs can recover the progressive frames back without any problem.

Again, as far as I know, the HV40 is different in that it keeps the video signal progressive all the way, similar to the 25P output from the XH-A1 (which, just to confuse matters, is derived from a deinterlaced image).

 

[jaegersing]

You misunderstood me (I should have been clearer). I was referring to: differences between HV20/HV30 vs HV40 all in in 25p mode (all PAL cameras). Could you expound on what you mentioned and specifically how it affects the final video look?

OK I see. Below is what I understand about 25P on the HV20/30/40.

There is no difference between HV20 and HV30 for the 25P modes. Both record their propgressive frames as PsF and output this on an interlaced video stream when you capture to computer.

HV40 records its progressive frames as progressive, and outputs a progressive video stream.

As long as the NLE is aware of the PsF format, it will combine the HV20/30 segmented frames back to progressive without any degradation or artifacts, so the visual appearance is identical to the HV40 25P video. However, if the NLE is set wrongly, it will try to deinterlace the video which will cause loss of resolution and/or deinterlace artifacts, depending on which deinterlace method is being used.

The big deal about the HV40 difference was really for NTSC users. For the HV20/30, the 24P video had 3:2 pull-down applied and was output on a 60i NTSC stream. The users had to jump through more hoops to deal with the pull-down removal properly, otherwise they ended up with loads of interlace artifacts. With the HV40, they get a progressive stream that is easier to work with.

 

[magicdragon]

Jaegersing,

Thanks for the clarification - so in essence if NLE is set to import at 1080i in a PAL version of HV40 you'll still get progressive frames (25 progressive frames per second). But on a HV20/HV30 (PAL) camcorder the NLE is set to import at 1080i you'll be importing at 25fps but since each frame is actually be made up of two frames you'll get degraded video image.

Thanks for the clarification (and yes pulldown is a pain in the rear LOL so an NTSC version of HV40) would be great in their neck of the woods.

 

[jaegersing]

Jaegersing,

Thanks for the clarification - so in essence if NLE is set to import at 1080i in a PAL version of HV40 you'll still get progressive frames (25 progressive frames per second). But on a HV20/HV30 (PAL) camcorder the NLE is set to import at 1080i you'll be importing at 25fps but since each frame is actually be made up of two frames you'll get degraded video image.

My experience with HV30 is that the NLE may or may not recognise that the footage is supposed to be progressive. If you force the NLE to identify the footage as progressive, it will work fine.

But, if you don't make sure that the clip properties are Progressive, you could have a problem. For example if you apply a speed change, the NLE will to try to deinterlace the progressive footage, and this will cause degradation. Rendering to progressive targets will also cause "deinterlacing" with similar results.