How PsF video from the V1 is different than "p" or "F" video

[Bob Grant]

Originally Posted by Steve Mullen

I wonder why Sony hasn't adjusted the settings for the E/P. Turning down Sharpness is a brute force solution as it kills both H. rez. (bad, and totally unneeded) as well as V. rez. (ok, but not ideal).

A flicker filter is a much more precise way of handling flicker.

There's a flicker switch but I thought it was for shooting 50Hz in a 60Hz country or vice versa, same as the Z1.

[Steve Mullen]

Originally Posted by Bob Grant

There's a flicker switch but I thought it was for shooting 50Hz in a 60Hz country or vice versa, same as the Z1.

You are correct -- the switch is for the 50Hz / 60Hz.

Not "our" twitter.

[Douglas Spotted Eagle]

Originally Posted by Bob Grant

There's a flicker switch but I thought it was for shooting 50Hz in a 60Hz country or vice versa, same as the Z1.

You're correct, Bob. The flicker switch has no bearing on what this thread is discussing.

[Steve Mullen]

Good news. The 25p timeline that I exported as 1080/60i through compressor did play -- with some dropped frames -- on my HD DVD player.

Four 25p S=8 clips:

1) no flicker filter: bad flicker -- just like original

2) minimum flicker filter -- less flicker

3) medium flicker filter -- even less flicker

4) maximum flicker filter -- very very little flicker

Looking at these results it seem with FCP it is possible to clean the signal such that when played from an hd dvd player to my HDTV -- it looks good. And, unlike the S=3, the bark on the tree is very detailed.

Sony's S=3 recommendation does not show a lot of research. It seems a panic fix.

However, I think S=8 is just too high. I think S=5 or S=6 would provide more than enough detail and enable the filter to remove all of it.

In fact, for a film look, the image needs to be a bit softer. If not 5, no more than 6.

I'm also wondering if I had made a 1080/50i disc -- would my Toshiba play it. And, would my HDTV show it. I'm wondering how multi-system these units might be.

So if you upload an S=5 clip I'll try it.

Anyway, I think it's clear there is an HD production solution for 25p.

Piotr, for all the work you've done, Sony should give you a V1. Thank you for all your help.

[Piotr Wozniacki]

Steve, thank you very much for your work! I'll upload an S=5 clip later today. It may sound dumb, but do I get it right you've engaged a flicker filter in FCP? Is the "Reduce interlace flicker" its counterpart in Vegas?

[Steve Mullen]

Originally Posted by Piotr Wozniacki

Steve, thank you vey much for your work! I'll upload an S=5 clip later today. It may sound dumb, but do I get it right you've engaged a flicker filter in FCP? Is the "Reduce interlace flicker" its counterpart in Vegas?

Yes -- and the FCP filter has the levels. With S=8 it required the Max level to eliminate most of the twitter.

I also tried to make a 50i hd dvd. Compressor will not let me -- no 50i support. When I get your new clip, I'll try a 25p. However, I suspect that the burning software may reject it.

Right now I'm going to see if the burning software will work with 24p. I fear not.

[Piotr Wozniacki]

Steve, the weather today is too dull for comparative shooting; I do have ready S=5 clips on disk but the shortest is just above the 100 MB limit of the server I've been using. Can you suggest any other host for a 107 MB file - if so, I could upload it immediately? Can your e-mail server accept an over 100 MB file?

Of course I can trim it, but this would require recompression...

[Tom Roper]

Steve,
Which HD-DVD model are you using? Does it have 1080p output?

[Piotr Wozniacki]

OK, so here is the link to a raw clip with sharpness=5:

http://rapidshare.com/files/19694265/s_5.m2t

I also encoded it using Vegas Flicker Reduction switch; please compare it with your FCP flicker switch and assess which of its levels it is closest to.

Here is the link: http://rapidshare.com/files/19712604...lter_Vegas.m2t

Finally, I encoded it again using the smallest possible amount of the Gaussian Blurr (V=0.001), and this - while showing noticeable softening of the image - eliminates all twitter, even with VLC bobbing, which we have esablished as most closely mimicking the "live" ipnut from camera.

Here is the link: http://rapidshare.com/files/19727858...1_in_Vegas.m2t

My opinion is that - with 25PsF playing fine with sharpness up to 12, provided no deinterlacing kicks in - we should try to find a way to deliver it that way, rather than reduce the resolution (even if vertical only). Sony should definitely speak their standpoint on this.

[Bob Grant]

Originally Posted by Piotr Wozniacki

.


My opinion is that - with 25PsF playing fine with sharpness up to 12, provided no deinterlacing kicks in - we should try to find a way to deliver it that way, rather than reduce the resolution (even if vertical only). Sony should definitely speak their standpoint on this.

Piotr,
from my work with this issue in SD there simply is no way to deliver 50i (or 60i I'd have to suppose) with V res greater than half the number of lines. Barry Green makes much of this in his support for progressive scan but I think the point missed is that even if you've acquired in P there's no way to force display devices to display it properly. SD CRT TVs only do I and I can assure you flicker can be a huge problem if you don't carefully watch what you're doing. HD CRT displays will do exactly the same thing. LCD displays that attempt to emulate interlace display will also manage to create flicker.

Someone correct me if I'm wrong but this 'bob' display system is close to what a real interlace display does. The difference would seem to be that bob will cause the edges to appear to roll or bob, a proper interlaced display will make them blink.

A fine line only one scan line high is always going to flicker on a 50i display. When the CRT writes that line, it's only displayed for as long as the phosphor persistance lasts. It's another 40mS before it gets displayed again. Something blinking at 25Hz is well under the point where persistance of vision kicks in, it's more of a problem where the duty cycle is very low, that's why the problem is worse on CRTs than LCDs, the poor response times of LCDs helps but as you've seen it's still not enough to totally avoid the problem. R60 displays will do a little better but still not good enough to eliminate the problem.

The only correct way to display 25p is what is effectively 50p, merge the fields and display the same frame twice. This is almost the same as what a film projector does with a two blade shutter.

None of this is really the cameras fault, it's delivering a perfectly good 25p image, you might argue it's too good but I'd rather have the res in the first place, it's going to make many things easier in post. The down side is you need to factor in at some point having to reduce the V res depending on your delivery.

[Steve Mullen]

Originally Posted by Tom Roper

Steve,
Which HD-DVD model are you using? Does it have 1080p output?

I have the A2 which only has 60i. This is no problem for me because I bought it only for my own work which, except for what I shot is Asia as a test, will always be 720p or 1080i60.

"Barry Green makes much of this in his support for progressive scan but I think the point missed is that even if you've acquired in P there's no way to force display devices to display it properly."

With 720p60 the monitor simply skips the deinterlacer and shows the P video exactly as it came from the CCDs. It looks perfect. I've used 720p since day 1 of HDV and I've never had a problem.

The fact is CineAlta is an acquistion format and I perhaps all PsF formats should be used that way. Or, at least that's true for R50 V1 video. The R50 V1 has way too much V. rez and too much aliasing in P mode. There really is a "problem" with the V1E/P. Turning S to 3 is a very bad fix for real production work. Euro Sony should get on top of this with a White Paper.

[Bob Grant]

Originally Posted by Steve Mullen

I have the A2 which only has 60i. This is no problem for me because I bought it only for my own work which, except for what I shot is Asia as a test, will always be 720p or 1080i60.

"Barry Green makes much of this in his support for progressive scan but I think the point missed is that even if you've acquired in P there's no way to force display devices to display it properly."

With 720p60 the monitor simply skips the deinterlacer and shows the P video exactly as it came from the CCDs. It looks perfect. I've used 720p since day 1 of HDV and I've never had a problem.

The fact is CineAlta is an acquistion format and I perhaps all PsF formats should be used that way. Or, at least that's true for R50 V1 video. The R50 V1 has way too much V. rez and too much aliasing in P mode. There really is a "problem" with the V1E/P. Turning S to 3 is a very bad fix for real production work. Euro Sony should get on top of this with a White Paper.

"The V1E/P has too much V res."

Maybe, depends how you want to deliver it and / or do with it. Wouldn't as much V res as possible be better for keying, display from computer systems or a film out?

Has anyone put a V1U and V1E upto a res chart to compare them?
Also lets not forget the V1U in 30p, is it having the same issues?

And we can still shoot 25p to deliver 24p very easily. Is even 24p guaranteed to be displayed correctly anyway. I realise you've got more of a shot with it than anything else but I'm under the impression it can still be problematic.

Regarding 720p, I don't think it's a big enough step up from SD in R50 countries to hold much attraction some how. That might be wrong in some ways based on what we're seeing, if 720p always gets displayed correctly then effectively the usable V res is about the same as 1080, if not higher.

[Steve Mullen]

Originally Posted by Bob Grant

"The V1E/P has too much V res."

Wouldn't as much V res as possible be better for keying, display from computer systems or a film out?

Has anyone put a V1U and V1E upto a res chart to compare them?

And we can still shoot 25p to deliver 24p very easily. Is even 24p guaranteed to be displayed correctly anyway.

if 720p always gets displayed correctly then effectively the usable V res is about the same as 1080, if not higher.

No problem with 30p or 24p so that's why I said it is only the R50 V1. While we have been focused on twitter, the other reason Sony recommended S=3 is to minimize aliasing. Same reason I use S=5/6 even though I don't have twitter. I would still not use above S=5/6 for the V1E/P.

720p always gets displayed correctly on progressive displays. That's why I'm such a fan of P. The V1 res is only VERY slightly higher than JVC 720p. So there is no real loss going to 720p. Tests have shown P "looks" 1.5X more detailed than I. So until there is 1080p, 720p is a far better format.

However, to get to 720p one has to convert files. If one is going to do this then there are two other options:

1) DVCPRO HD at 1080i50 which is 1280x1080 at 4:2:2 at 10-bits. This is a great option as it preserves full V. rez and 1280 is more than enough H. rez for a camcorder with only 960-pixels. You can export from Compressor for HD DVD at 1080p25. I'll soon know if it can be burned.

2) CineForm's new codec gives you 10-bits.

You can also convert to 720p25 AIC. This file is 317MB for 30S vs 422MB for DVCPROHD. I would go with DVCPROHD. However, it may be reduction to 720p may reduce V. rez automatically eliminating the need for a Flicker Filter.

-------------------

If the flags are set properly and one has a hidef dvd player that outputs 1080p60 -- then 24p will be perfect. The real goal in the USA is now 1080p24 into an LCD that will display it at 24p or 48p. This will duplicate a movie perfectly.

How are you easily going to get 25p to 24p without audio problems?

[Douglas Spotted Eagle]

Originally Posted by Steve Mullen

How are you easily going to get 25p to 24p without audio problems?

Too many ways to count on two hands. Without any hardware, just open a 25p file in Vegas. Set project properties to 24p, drag audio. No sync issues, regardless of the length of the project. There is always Minerva, there are also resampling algorithms in most DVD player decoders that account for this.
24p to 25p, 25p to 24p, 24p to 60i, 60i to 25p, 30p to 24p, 24p to 50i...none of these framerates offer challenges any longer on the audio side, even if they do present difficulties to some degree on the video side.

I'd bet Bob does this in his sleep, given the work he does.

[Bob Grant]

Originally Posted by Steve Mullen

How are you easily going to get 25p to 24p without audio problems?

Ctl+Drag the end of the event so it's 4% longer, you've got the option of correcting the 4% pitch change in the audio. Pretty much the reverse of what our telecines do except I'm told they don't even bother to shift the audio pitch back.

The question of how to do this in FCP came up on the Sony Vegas forum, from memory there's 3rd party utilities to do the same thing.

Probably the safest approach so you don't have a can of worms is to convert as a final step i.e. run a 25p project, render a mixed master and convert that.