I just read an HR10 AVCHD review at another web site, and again he says that the 24p of the AVCHD cameras makes unusable footage because of stuttering. But the claim is the HV20 is fine at 24p.

I'd like to test this myself. I want to mount a HG10 and a HV20 to a board, shoot the same settings with some pans, and do a more meaningfully comparison of the matter.

I want to determine:

If the AVCHD codec drops too much information on high motion, OR
If the shutter on the two cameras is implemented differently, OR
The reviewer keeps decoding the files wrong again and again, OR
There is no difference that would contribute to a strobing effect being seen.

Anyone in the Hollywood/LA area have a HV20 that wants to help test? It would only take 15 minutes or so.

-Les

I'm guessing whatever filter he is using to get the 24p out of the 60i stream is performing deinterlacing on the source footage before the IVTC is occurring. I totally saw this when I used the InterVideo filter as a feed to CineForm HDLink, and the stuttering "went away" when I switched to the CoreAVC Pro codec with deinterlace TURNED OFF.

It would be nice if they mentioned exactly HOW they're viewing the 24p so we could figure this out. I'll bet their slam on the camcorder is keeping hundreds from buying them. Of course, if Canon had made it easier to get at this footage out of the box, it wouldn't have been a problem in the first place.

Why they are messing with telecining in an all-digital format is beyond me. I know, I know....cutting edge, right?

mike@slavis.com

I'm guessing whatever filter he is using to get the 24p out of the 60i stream is performing deinterlacing on the source footage before the IVTC is occurring. I totally saw this when I used the InterVideo filter as a feed to CineForm HDLink, and the stuttering "went away" when I switched to the CoreAVC Pro codec with deinterlace TURNED OFF.
Did you see this on a computer screen or on a TV screen? Computer screen does not count ;-)
Why they are messing with telecining in an all-digital format is beyond me.
Because this is a movie effect for otherwise standard video stream, intended to be watched on a TV, not on a computer screen. On the other hand, Canon is not helping by non-flagging telecined video properly. Some TVs are able to IVTC but only if proper flags are present. There are only a handful of TVs that do proper IVTC with badly-flagged or non-flagged material. I don't see why one should pay extra for Faroudja or Realta while others (like Canon) still produce non-standard signal.

If the actual 24p frames are non strobed when examining them on a device such as a computer display that has no trouble with 24p, then you have great 24p material.
If you then see problems on a tv set, it's the tv set's problem or fault in trying to display the 24p material in a pleasant to watch format.

To put it another way: If the 24p from the video camera looks just like the 24p from a film camera with respect to strobing and motion blur, then they would both have similar problem or no problems when brought over to a tv set.

Still another way to look at the issue: If I was going to film out the 24p from the video camera, it should look great on the computer looking at frames ( one at a time, and in motion ). I wouldn't care less what a TV set would do with it.
-Les


Did you see this on a computer screen or on a TV screen? Computer screen does not count ;-)
l.

Hi there,

Here's a busy clip at dusk of me following my 2 yr old using the HG10 at 24P.

http://www.seanmccormick.biz/outbox/HG10/run_24.mov

The file is Mpeg 2 with an mp3 tacked on as audio in a QT wrapper.

In Compressor, I just made sure the right field dominance was selected in both the initial screen and in Frame Controls.

Other than it just being shaky footage, I really don't see an unusual strobing problem.

Let me know if I'm high and I'm missing something here...

Nite Nite,

Sean

Sean:

I'm confused by your workflow. Why would you remove a field, instead doing a pulldown?

The biggest help with slowdowns in 24p is to turn the Auto slow shutter off in the set up menu. The difference is pretty dramatic. Don't own a HV20 to do a comparison however.

From the manual on enabling Auto Slow Shutter:

"The camcorder automatically uses slow shutter speeds to obtain brighter recordings in places with insufficient lighting".

David Grim,

I thought this only referred to the camcorder dropping into 1/12 s shutter if you were in a very dark situation. This means you're recording 12 fps instead of 24 fps and the video has a seriously jerky look to it (in my opinion). I've been testing with this feature on and have concluded that I'd rather have darker, grainier images than a 12 fps rate so plan to turn it off. However, for "normal" shooting this setting shouldn't come into play, is that correct? Am I missing something?

Thanks,

mike@slavis.com

Hi there,

Here's a busy clip at dusk of me following my 2 yr old using the HG10 at 24P.

http://www.seanmccormick.biz/outbox/HG10/run_24.mov

The file is Mpeg 2 with an mp3 tacked on as audio in a QT wrapper.

In Compressor, I just made sure the right field dominance was selected in both the initial screen and in Frame Controls.

Other than it just being shaky footage, I really don't see an unusual strobing problem.

Let me know if I'm high and I'm missing something here...

Nite Nite,

Sean

Looks fine to me. But, I have to say... when that kid conks himself with the horseshoe... I almost fell out of me seat laughing. Reminds me of what I habitually do to myself... daily. ;-)

It's the metallic 'bong' sound that really sells it. That kids skull has more dents in it than the Liberty Bell.

I noticed in the FCP Log and Transfer prefs that you can tick a box for 3:2 pulldown (checked by default). How does that affect the FCP-created ProRes file when the camera is using 24P mode?

My issues with strobing where medium to faster pans across stationary objects.

Did you test that?

I'll test that out tomorrow.

I was hoping to compare the HG10 and the HV20 when I started this thread.
I know the HV20 is sharper than the HG10 due to the AVCHD codec stepping on the high frequency detail to provide more compression. The differences in 24p strobing have yet to be seriously investigated.
Strapping both cameras to the same camera mount ( a board , even ) and doing rapid pans would do it.

-Les

Les,

I can't believe there isn't anyone if Hollywood/L.A. to lend you the use of their HV20 for a bit.

If anyone in the New York Finger Lakes region who owns an HV20 is willing to drop by my house for about an hour (I'm in Canandaigua), I'll try to run some of these tests myself too. Let me know, and if you're in L.A./Hollywood area, get with Les. I'm running out of time to return this cam if I decide I don't want it (I'll probably keep it), and I'd just LIKE TO KNOW for sure, y'know?

Thanks,

mike@slavis.com

HMMMM. Makes me want to go get one at Circuit City, to try out against my HV20.. Meeting in Bakersfiled...? along way to go for that.... Gotta be someone else in LA.... Anyone in Sacramento area have the HG10 yet ?

Just to cap off the subject, there is no difference in strobing or stuttering between the HG10 and HV20.

See the thread "hg10 vs hv20 vs xha1"

Another site is incorrect at slamming the HG10 in that respect, it turns out.
-Les

thanks, Les,
you guys' review is the most important reason that I decided to order hg10 instead of hv20.

I agree, I've been seriously considering upgrading from a Panasonic HDC-SD1 to an HG10 and spent a lot of time yesterday confirming that with the CoreAVC decoder installer and the latest version of Cineform HDLink that I was able to transcode absolutely smooth 24P content from the available samples posted, so I'll probably pick one up this weekend.

As stated previously, the key seems to be to have a h.264 video decoder that allows disabling of any decoder based deinterlacing, for the CoreAVC decoder, I set the deinterlace to force Weave (no deinterlacing) and disabled the "Aggressive deinterlacing" option. I also had to specifically uninstall PowerDVD, which was being used as the default h.264 renderer previously. When installing CoreAVC, when it runs the Haali Media Splitter installer, make sure to disable all default file format handling (AVI, TS, etc.) as this will cause the Haali Splitter to be brought into the filter graph as a standalone filter rather than as a combined file source and splitter.

Once you can drop your .MTS (I rename them .m2ts) files onto GraphEdit (part of the Microsoft DirectX SDK) and have the following filter graph generated, you'll probably be golden:

http://members.cox.net/eskin3/coreavc_filter_graph.jpg