HV20 - Progressive mode recording
 

I am considering buying HV20. The progressive image acquisition mode sounds like a great advantage to me. But I am a bit confused, because the HV20 Manual says the "video shot with the [ HDV (PF24)] standard will be recorded on the tape as 60i" anyway. So, what's the worth?

Would appreciate your comments.

I will say again what I said in this thread: http://www.dvinfo.net/conf/showthread.php?t=87445

You will be able to play it back in other cameras like you always do with 24p.
24f on the other hand can only be played back by 4 (soon 5) of Canon's own cameras - the HV10, XL H1, XH A1 and XH G1 (soon also the HV20).

The once oh so popular Panasonic AG-DVX100 and Canon XL2 also carries 24p within 60i.

Thanks for your explanation.

What I am warried about is if these all signal conversions may add much noise to the output picture?

Maybe I have missed something but I don't know of any video camera that puts out 24p as 24p. The image is still progressive since there are no fields to begin with in progressive mode. When you are seeing a progressive image within an interlaced signal on a tv, it's like you are seeing the same image twice every second. I think that's the case with PAL (50i/25p) anyway. NTSC 60i/24p needs some kind of pulldown since 60/24 is not an even number. I never had to worry about that since I'm in a PAL country. Someone else might give you more information about that.

Nevertheless, the DVX100 and XL2 seem to do fine anyway. We have yet to see more images from the HV20 but I don't think the fact that it's 24p is embedded in a 60i signal is going to add any noise. The noise comes from other factors. Don't take my word for it but I think that cameras with native progressive sensors (like the DVX100, XL2 and HV20) are more noisy in the progressive mode than they are in interlaced mode. I think it has to do with the way the sensor "sees" less light since it only grabs information from the sensors 24/25 times a second instead of 60/50 times.

There has been some discussion about the 24/25f-mode on the XL H1, XH A1 and XH G1 cameras. The sensors in those cameras are interlaced and therefore they grab information from the sensors 60/50 times a second and then the camera does some sort of fancy deinterlacing to make it into a progressive picture, and therefore it's progressive picture might be brighter and less noisy than the progressive picture from cameras with native progressive sensors.

I'm way over my head with all this tech talk so I hope that someone with more technical knowledge than me will be kind enough to either confirm or correct this. :)

Seems pretty good, Ron! Except I think that some of the newer interlaced cams that also do progressive (Canon A1, for example?) now clock themselves as 48 instead of 60, eliminating that nasty uneven motion thing. For Europe though you are right.

Bruce Allen
www.boacinema.com

I found this text very helpful. In a way, it explains how 24p gets to be 60i.

http://www.animemusicvideos.org/guid...h/video2_2.htm

Also wikipedia has tons of stuff on this.

Not quite !

The two fields (images) are not the same they differ spatially but not temporally.

That is: they are both captured from the same point in time (unlike interlaced footage) and each field is a unique spatial recording.

The JVC HD series of cameras records 24p as 24p. The higher end Canon cameras also record 24 frames when using 24F. It is the camera that adds the pulldown when playing back. The footage on the tape however is 24p. 24p DVD's are also 24 fps. There are flags set that tell a DVD player to mix it all up as fields with 3:2 pulldown so the DVD will work on any TV.

Some other cameras such as the SONY V1 record as 60i but there are flags set so a NLE will be able to pull out the 24p.

It is correct that 24p on the HV20 is really no different then 24p on the Panasonic DVX100 or the Canon XL2. DV cannot have flags so the only way to do it is to record as 60i with pulldown. The HV20 while giving us the chance to have 24p recording does not make it as easy to edit the 24p. It will be up to the capture or editing software we use to figure out how to remove the pulldown to create the 24p. I know Cineform is very good at doing this on the fly and Apple has tools to remove the pulldown after it is captured.

Capturing through HDMI isn't any different. In fact all cameras which use Component/HD-SDI/HDMI to capture end up putting out only 60i or 60p so software has to be used to remove the pulldown. This is true for even the cameras which do record 24p to tape. The first lower cost camera to actually send a true 24p through HD-SDI is the JVC HD-250 because it uses the Panasonic style of repeat frame flags used in the Varicam camera.

If I plan on reversing pulldown with Cinema Tools, how big a difference is there between originally capturing the 60i (24p with pulldown) footage through HDMI versus HDV?

I assume that the HV20 captures 24p, then performs the pulldown process, and then compresses it to HDV. So, the reverse pulldown process in Cinema Tools seems like it would be less exact on the HDV footage than on raw HDMI footage.

Is this a big deal?

I knew about the Canons, but I was thinking 24p, not 24f. :) You are absolutely right though, 24f is put out as progressive. I did not know that the JVCs put out 24p as 24p however. One learns something new everyday.

Sometimes it's good to live in a PAL country. ;)

If I have a PAL HV20, will I be getting full non-interlaced frames once the footage is in the computer?
It would (will?) be great to be able to pull any frame out of the hdv stream to use as a still.

Hi Elliot, very likely. Although the semantics of a 50i vs 25p are of course different, so NLE packages might take a different approach. You should in any case avoid regular de-interlacing such as blending the fields.

Ron, I always felt priviliged in PAL land, because of the much higher resolution than NTSC. With HD however, we're just stuck with the lower "field-rate" :(

Billy, funny, never thought of that. The pulldown might even make the MPEG2 motion detection much more difficult? However, I don't think its a big deal, considering I haven't really seen any obtrusive MPEG2 artifacts on any of the sample footage posted here.

Lower field rate (or frame rate) but better compression.

HDV = 25Mbps for both NTSC and PAL.

Therefore:

NTSC has a higher frame rate but each frame is more compressed.
PAL has a lower frame rate but each frame is less compressed.

Yes. :)


.


.


.

With PAL 25p from the HV20 you do have to be carefull how it is dealt with in your NLE. To us the video looks progressive but to a NLE the video will look like interlaced video and it will process it in that way.

Any animated effects you may add in that timeline such as moving titles or any type of a wipe will end up rendering as 50i. This may make your final video look a little funky because the video will be progressive but the stuff added will be interlaced.

You will have to trick your NLE into thinking the video actually is progressive and use a 25p timeline. This way everything should work out fine. The methods used to adjust the timeline properties are different for every NLE.

The same is true for 24p editing from the HV20. If you edit the 60i in a 60i timeline any effects you create will be rendered as 60i and not 24p with 3:2 pulldown. Lucky for us goofy NTSC users we end up having to use a 24p timeline anyway so everything 24p renders correctly. By the time we process the 24p out of the 60i file we don't really have to worry anymore unless we want to add the pulldown back in to record back to tape.

That's the real beauty of working with 25p HDV as opposed to 24p HDV. All you have to do with 25p is take a matching HDV preset and set the fields to none or progressive.
I think 24p HDV needs some kind of additional support by the NLE?

You in fact leave the project as HDV 50i interlaced as the 25 progressive frames are held in a 50i stream.

So I should leave it at Lower Field First? But I thought there were no fields? I always set it to progressive and that worked for me. What would be better about leaving it at Lower Field First for instance?

Lee see my post above as to why this can give bad results. A 50i timeline will mean everything added will render as 50i and not 25p. You have to use a 25p timeline in order for things to render correctly. Of course if you only ever do cuts only then you will not have much to worry about but even a cross dissolve will end up rendering as 50i. Now maybe mixing 50i effects and 25p raw video isn't that big if a deal to some people but if somebody shoots progressive they may want to keep everything progressive so it is good for them to know to choose a 25p project instead.

...that's what I meant with semantics: it seems the same but it means something different. Another thing you want to avoid by not using a 50i timeline is a HDTV displaying the odd and even fields of the (once progressive)frame alternatively, deinterlacing them on the fly, thus decreasing vertical resolution and actually introducing flicker. Whoa...

So while in this case NLE input is (or seems) interlaced, output /must/ be progressive.

(That is, I have yet to produce my 1st progressive MPEG2 project, so for me this is theory until I receive my HV20...can't wait!)

You will not be able to go back to tape as true progressive with the HV20. The 24p video would have to add the 3:2 back in and then encode to HDV.

Then again the HV20 does playback 24F so maybe a NLE could feed 24F if it was able to output 24F to the HV20. Even then the way 1080i works as soon as you hook up the HDMI cable the HDTV will treat it as 60i.

The only option I can see would be to create a true 24p HD-DVD or Blu-Ray disk. The HD disk players like progressive scan DVD players would be able to output true progressive frames via the HDMI or component output.

The other option of course is to create a 24p DVD by down converting your project to SD. If you can remove the pulldown in your NLE then 24p video from the HV20 should create some earth shattering 24p DVD's. I think the same should be true for 25p DVD's as well. I have never made a 25p DVD but I am pretty sure it works fine. Of course this no longer HD but it is one of the only ways to deliver progressive for interlaced worlds.

Could I put 25p back on tape with the HV20 (since no pulldown needs to be reversed for 25p)?

At this point I'm starting to wonder whether there are flags in something like a lead portion of the streams on tape / hard drive, that indicate the nature of the fields. There must be... Those are likely to be read and used by different elements of the production chain, including print-to-tape.

I did a brief scan in other area's here, but couldn't come up with something like that. Elsewhere I've found indication of such a flag being embedded in MPEG2 streams, that are used by DVD players and receivers to determine what to do with the fields and whether to deinterlace, inverse telicine or what.

It would be nice to be able to learn how to use those flags and not have to worry about these things...

Since the HV20 plays back the F modes it is hard to say if F mode footage could be encoded from the 25p and then sent to the HV20. This may or may not work and we really will not know until the camera comes out.

As for normal HDV the video being sent back to the camera must be 50i. That means your 25p project will have to be encoded as a 50i HDV stream.

We also do not know if there will be any flags or not. Some HDV cameras have progressive flags while some do not.

We do have to remember that the HV20 is a consumer camera. Canon may have mostly put progressive recording in for mostly a look and not so much a format that is designed to have a end to end solution for progressive editing. That isn't to say that solutions will not come up in NLE's in the near future. In fact I'm sure Cineform will directly support removing the pulldown on the fly during capture. In this case if you shoot 25p or 24p with the HV20 you will end up with captured Cineform AVI files that really are progressive.

It would be nice if the HV20 adds flags to the video but at this point I doubt it. If it does then of course that makes an already sweet camera even better.

That can be changed with one click in the settings of any modern nle. The computers don't "see" anything.

Hi Mikko,
true, but after configuring your project settings right, how do you know that the NLE won't deinterlace a 25p file assuming it's 50i? That's where flags might come in...

Again: can't wait. Is there anyone here who happens to have some 25p material we can experiment on?