Hands on the new Canon HG20

[Chris Hurd]

Starting with a close look at the HG20 camcorder's exterior:

http://www.dvinfo.net/articles/camco...g20images1.php

and an overview of its menu system:

http://www.dvinfo.net/articles/camco...structure1.php

Comparison video (HG20, HG10, HF11, HF100) to follow; should have the clips posted later tonight.

[Brian Boyko]

How do you do manual focus on the new cams?

[Chris Hurd]

joystick as usual

[Steve Mullen]

Does the fact Canon uses the terms "60i" plus "PF24" and "PR30" indicated that the latter two are not really 24p and 30p?

I'm not talking about the recording system which is, of course, "24F within 60f" and "30F within 60f." I'm talking about HOW the 24 and 30 are obtained.

The HG10 was reviewed as having the worst 24p they had ever seen -- while the HF series has been reviewed as having the best. Canon has obtained 24fps in various ways. Have you learned how the HG21 does it verses the HF11?

Looking at the Canon.jp site -- it looks as though there are no 24p appropriate shutter-speeds. If true, that means a choice between 1/30th or 1/60th -- both wrong. This may mean Canon only runs the CMOS at 60p. Which means 24 frames are pulled 60p using a 2:3 pattern introducing judder. Then, the 24fps must have 2:3 pulldown added before recording recording at 60i -- which introduces another judder pattern.

The current HDreview magazine has great pix of how these two judders combine. Ugh! Plus, with the too fast shutter-speed -- it explains why PF24 on the HG10 was unuseable. At least the HG21 has PF30 which will look fine.

PS: does the VF pull-out and/or tilt-up?

[Chris Hurd]

Originally Posted by Steve Mullen

The HG10 was reviewed as having the worst 24p they had ever seen -- while the HF series has been reviewed as having the best.

Don't believe everything you read on the internet, Steve, because your source couldn't be more wrong.

And anyway it makes absolutely no sense at all, because there is no difference whatsoever in the way the HG10 records 24p relative to the HF series or any other Canon AVCHD and HDV camcorders. There was one "review" of the HG10 elsewhere on the web which made a bizarre claim of "terrible 24p" without providing any visual evidence whatsoever to back it up.

Here at DV Info Net, we proved conclusively that such a claim was completely bogus, and we validated our assertion by providing for public review the raw data which *proves* that 24p from the HG10 is no different than 24p from Canon's consumer HDV camcorders (the HV10, HV20 and HV30) or Canon's other consumer AVCHD camcorders. As expected, they all use the exact same 24p implementation. The HG10 is no different from other Canon camcorders in this regard.

See http://www.dvinfo.net/conf/showthread.php?t=101059

See also http://www.dvinfo.net/conf/showthrea...=106626&page=2

Always take DV Info Net's word over *any other online source,* because we're usually right (with the track record to prove it).

Looking at the Canon.jp site -- it looks as though there are no 24p appropriate shutter-speeds.

Incorrect. These camcorders offer 1/12, 1/24 and 1/48 shutter speeds in 24p mode.

I have not yet seen the HG21, but I believe the EVF pulls out (but not up).

[Ken Ross]

Chris, do these new cams record at the same high bitrate (25mbps) as the soon to be released HF11?

[Steve Mullen]

Originally Posted by Chris Hurd

Don't believe everything you read on the internet, Steve, because your source couldn't be more wrong.

It's interesting that beyond the "other" review site -- a google for worldwide reviews including those from owners and you'll find they report the same thing -- "bad 24p." (Which I take to mean bad 24p within 60i.)

Also, looking at the series of threads you provided,:

1) The claim made in these threads is actually "no difference from the HV20/30" which makes sense because I expect the HG series uses the same 24p system. I see no post that claims the HG looks no different than the HF10/100. And, that is the question we face today.

2) the posted explanation that the "other" reviewer was seeing 24-within-60 on a computer screen sounds valid until you realize everbody views 24within60 on computer screens. So the HF10 video should have been reviewed as bad -- and it wasn't. In fact, if that were the explanation -- then the V1/FX7 would have gotten equally bad 24fps reviews from everybody. So should the Z7. They have NOT. This explanation is simply not valid.

3) There is another post that says "but on an interlaced monitor it will look the same as with the pulldown removed on a computer." Whatever this means -- it's not an explanation because it would apply to all cameras that use pulldown.

4) Moreover, the last comment about "panning" on the 24fps thread is "The fountain clip has panning in it. I just watched "hg10 24p-10k.mov" clip and oh boy does it stutter when the camera pans from the fountain to the table with glasses. I understand that this is 24fps and it should stutter more than 60i, but it is so visible it literally hurts my eyes, I barely can watch this, my first reaction is to turn away from the screen."

(Unfortunately, I can't find THIS clip and have no idea if it has pulldown or not -- although I suspect it did.)

5) This same post, which was NEVER answered -- asks: "Could higher shutter speed exacerbate stutter? Anyone else experienced stuttering in this episode?" None of the posts report the shutter-speed being used for 24p they shot.

So -- I do not see anything that on these threads that "proves" anything. Although I'm sure hoping you are correct!

PS: you are correct about the shutter-speeds. Odd the Canon Japan site did not list this.

[Evan Donn]

Originally Posted by Steve Mullen

I'm not talking about the recording system which is, of course, "24F within 60f" and "30F within 60f." I'm talking about HOW the 24 and 30 are obtained.

Not to get too far outside of the thread topic, but this is something I've been wondering - why don't any of these cameras encode to a truly progressive format when shooting in progressive mode? I understand the argument for this regarding tape formats but I don't see any reason why we should be stuck with interlaced files when we're recording files to a generic solid-state format.

[Steve Mullen]

Originally Posted by Evan Donn

Not to get too far outside of the thread topic, but this is something I've been wondering - why don't any of these cameras encode to a truly progressive format when shooting in progressive mode? I understand the argument for this regarding tape formats but I don't see any reason why we should be stuck with interlaced files when we're recording files to a generic solid-state format.

I think Canon knows than the vast majority of those who switch on 24p mode are never going to make a film. And, it's likely only a few are making 24p Blu-ray discs. Moreover, Japan is probably unaware of Vimeo and its desire of 720p24.

Which means mot shooters after they get their "funky" 24fps look, will edit it like they usually do -- in 60i. They will have no understanding that doing so creates broken 2:3 cadence at every cut-point. They'll be equally unaware that whenever 2 or more streams are mixed that these stream will not have the same cadence.

However, that doesn't mean they couldn't have two 24p modes! Once again we see the need for a prosumer camcorder based-upon consumer camcorders but with these extra features and controls. And, I'm unwilling to pay $3500 to get these goodies that are nothing more than a few more buttons and different firmware.

Frankly, I'd simply buy a Z7 at the best discount I could get it for -- and avoid AVCHD entirely!

[Tom Roper]

I think there actually is some factual basis to Canon 24p and 24f not being quite the same. It comes from one of Eugenia's rants, something to the effect that 24p on the AVCHD from Canon mixes an interlaced frame with a progressive one, thus the problem of removing pulldown (inverse telecine) has some complicating steps to the process. She can explain it better. Here is the link to one of her rants...

http://eugenia.gnomefiles.org/2008/0...ldown-removal/

[Steve Mullen]

Originally Posted by Tom Roper

I think there actually is some factual basis to Canon 24p and 24f not being quite the same. It comes from one of Eugenia's rants, something to the effect that 24p on the AVCHD from Canon mixes an interlaced frame with a progressive one, thus the problem of removing pulldown (inverse telecine) has some complicating steps to the process. She can explain it better. Here is the link to one of her rants...

http://eugenia.gnomefiles.org/2008/0...ldown-removal/

She's "describing, but none too well" 2:3 pulldown where some interlaced frames have fields that come from the same progressive frame while other interlaced frames have one field from one progressive frame and one field from an adjacent progressive frame. The 2-3 cadence creates judder.

Both the hf10 nd hg10 use 2-3 pulldown. But, they don't -- it is claimed by some -- look the same. In a previous post I explained that one way of getting 24 progressive frames from 60 progressive frames adds a SECOND cadence which creates another judder. When pulldown judder is added to it -- a very unpleasent studder is created.

Why would canon allow this? It means the camera doesn't have to run at both 60p and 24p. Much simpler circuit. We'll see if the hg21 work like the hf10 or the gh10.

[Tom Roper]

Originally Posted by Steve Mullen

She's "describing, but none too well" 2:3 pulldown where some interlaced frames have fields that come from the same progressive frame while other interlaced frames have one field from one progressive frame and one field from an adjacent progressive frame. The 2-3 cadence creates judder.

Both the hf10 nd hg10 use 2-3 pulldown. But, they don't -- it is claimed by some -- look the same. In a previous post I explained that one way of getting 24 progressive frames from 60 progressive frames adds a SECOND cadence which creates another judder. When pulldown judder is added to it -- a very unpleasent studder is created.

Why would canon allow this? It means the camera doesn't have to run at both 60p and 24p. Much simpler circuit. We'll see if the hg21 work like the hf10 or the gh10.

If I took 24p from a Sony EX1, I could use the utility dgpulldown.exe to add repeat flags to go from A-B-C-D to AA-BBB-CC-DDD. When decoded it would playback with typical 2:3 pulldown judder, but the frames would have no interlacing artifacts or combing because they were not re-rendered.

on the other hand...

If I took 24f HDV from a Canon XH-A1 which has the repeat flags already embedded, I could use the same utility dgpulldown.exe to remove the repeat flags to go from AA-BBB-CC-DDD to A-B-C-D, losing the pulldown judder. But just like the Sony EX1 above, there are no deinterlacing artifacts or combing since the Canon digic processor encoded them as progressive frames, and the dgpulldown removal tool merely stripped away the repeat flags.

In the second case (above), the point I am making is that 24f from the Canon HDV XH-A1 is not the same as 24p from the Canon consumer camcorders, which interleave temporal fields into some frames, and non temporal ones into other frames. While it's still 24fps and gives it that look, it's not progressive encoding in my judgment. And my only point in stating that observation, was a counterpoint to the notion that all Canon 24p/f is handled the same way. They aren't, as Eugenia pointed out "none too well."

Your point is really directed at a different observation, that hf10 and hg10 don't have the same look at 24fps, due to a different cadence, perhaps something like 3:2:2:3 or perhaps something else. It's the judder from the cadence that accounts for the difference.

[Chris Hurd]

Originally Posted by Tom Roper

... a counterpoint to the notion that all Canon 24p/f is handled the same way.

Sorry, Tom, but that assertion was never made here. Of course there is a difference between 24F Frame mode from Canon's three-chip HDV camcorders and the 24p-within-60i scheme offered by their single-chip consumer camcorders. That fact has never been the point of contention here.

Instead, the assertion made here is that 24p is handled the exact same way within all Canon *single-chip consumer* HD products (a line currently composed by a couple of HDV and several AVCHD camcorders)

We proved this once before, by offering for download a number of comparison clips which clearly reveal this assertion to be true.

And, I'm about to do it again, by offering for download a number of comparison clips showing the same scene in 24p from the HG10, HF100, HF11 and HG20. The clips have different encoding rates, but the 24p cadence is clearly identical (it wouldn't make sense from them to be any different, as I've pointed out several times before).

Your point is really directed at a different observation, that hf10 and hg10 don't have the same look at 24fps...

Actually they *do* have the same look at 24fps -- as we have proved before and will prove again -- and that is not a different observation; it's always been the primary observation of this particular thread. Thanks for understanding,

[Chris Hurd]

In order to keep *this* thread on topic -- the topic being the new HG20 -- I have started a separate discussion for comparing a couple of downloadable 24p clips, M2TS files taken straight out of the camcorders, one from the HG10 and one from the HF100. That link is http://www.dvinfo.net/conf/avchd-for...g10-hf100.html

[Eugene Fong]

Originally Posted by Chris Hurd

In order to keep *this* thread on topic -- the topic being the new HG20 -- I have started a separate discussion for comparing a couple of downloadable 24p clips, M2TS files taken straight out of the camcorders, one from the HG10 and one from the HF100. That link is http://www.dvinfo.net/conf/avchd-for...g10-hf100.html

Any updates on the review of this Camcorder (HG20)? I'm interested in the low-light performance of this video camera and how it compares to the the HG10, HF10, and HF100.