Hands on the new Canon HF11
 

See http://www.dvinfo.net/articles/camco...11overview.php -- discuss here -- thanks,

Thanks, Chris and Austin for this.

From the standpoint of features, are exposure adjustments with HF11 about the same as with the HV30, ie., without the direct aperature or shutter adjustments that prosumer cameras have, but that can be adjusted with various tricks ?

All identical; no differences in that regard. There are only three changes in the HF11: body color, internal flash memory size and the addition of the MXP mode. Everything else stays the same as before. Hope this helps,

It may go without saying Chris, but I wanted to thank you most sincerely on behalf of the entire video community for providing such a truly superb website with such totally up to date information.

You are allowing all of us a chance to share and learn in a way which is outstanding, and the quality and timeliness of information you provide such as this HF11 review and discussion is absolutely unavailable anywhere else on the Internet.

Again, a very sincere and very big "Thank You" from all of us.

Larry Horwitz

I want to raise a technical point which is worth considering when making these video comparisons:

No doubt the HF11 uses a new and presumably faster encoder to allow 24 MBit/sec video to be created from the same sensor / optics which the prior HF10 and HF100 also use. The FXP samples which the new camera records at 17 MBits/sec, despite being at the very same encoding rate as the earlier HF10/100, may themselves be superior to the HF10/100, since the new encoder is loafing along at the 17 MBit/sec rate and quite possibly may do a better encoding job.

Therefore, the FXP samples of the HF11 may indeed be superior to the FXP samples coming from the older HF10, so people who presently own the HF10/100 may ultimately want to see comparisons where the older encoder and the newer encoder are directly compared. The HF10/100 may not be capable of either the detail or motion performance of the HF11 even if both are compared at the 17 Mbit/sec rate.

I wanted to point out this caveat in case the reader assumes that the MXP (24Mbit/sec) versus FXP (17Mbit/sec) comparison of the HF11 is somehow a direct indication of how much improvement, if any, they are likely to see if they were to "trade-up" from an HF10 to an HF11. This is not a valid assumption given the likely encoder differences between the two models.

Larry Horwitz

I can arrange that. In addition to the HF11, I have an HF100 sitting here at the house. If y'all don't mind a repeat of the water sprinkler test, I'll try to knock that out later this afternoon or tomorrow morning (things are kinda busy around here though).

Thanks for your kind words Larry,

The direct comparison of image quality (24Mbps HF11 vs 17Mbps HF10/HF100) is what I'm interested in. I've been meaning to pick up an HF100, but am going to wait to see if there is a considerable difference in the image quality (due to higher bitrate with the HF11).

Note to Chris: Most the DSC chart image links somehow got mistakenly pointed back to the article page.

D'oh! In too big a hurry. Fixed -- thanks Robert!

Chris, you and Austin did a fantastic job with your test! I'm able to download the .m2tx files and don't even need to change the extension. They play in WMP with no problem on my computer. However, I do have the same problem I had with Austin's original HF10 files when downloading the image files. I just can't get them to play.

But from what I've seen, I'm encouraged by the high bitrate clips of the sprinkler pans. These are some very detailed scenes that hold together really well during the pans! I can definitely see a difference in the high bitrate. I'm also encouraged, even though it was a cloudy day, by the color balance of the clips. They seem more like the color balance I have with my HV20.

Thanks for a job well done. Any additional clips you can provide (perhaps .m2tx files) that demonstrate the resolving ability of the cam as well as the color balance, would be greatly appreciated!

Thanks Chris. This should make the upgrade decision much simpler for those of us with HF10/100s.

Shot last night; have to run up to Austin today but hope to post this eve.

thanks for the review.
I can play the mts files with no problem, but after an hour
of trying several programms to join, convert and extract the dmgs I gave up on it...

Sorry about the trouble with the DMG. I hope to have QuickTime .MOV's of that material up later tonight

here are some split and concat and par2 utilities for windows that looked the part (i can't vouch for any of them, so be forewarned)

there is also a method using windows command line involving the copy command:
http://www.monkeyjob.com/FileMonk/FA...ommand-DOS.htm


File Splitz:
http://www.download.com/File-Splitz/...dlPid=10521066

Quick Par:
http://www.quickpar.org.uk/Download.htm

there are also a lot of newsreaders that have par2 and joining capabilities built in:

Ozum:
http://www.download.com/Ozum/3000-21...dlPid=10828332

news file grabber:
http://www.download.com/News-File-Gr...dlPid=10866600

Great point! Given the way H.264 work,s it's certainly possible that the "new" encoder -- be it firmware and/or a new chip, uses more of the encoding toolset tools. I also have to wonder about both Panasonic and Canon offering 24Mbps at almost exactly the same time. It seems like both have access to a new chip.

Remember that Canon and Pana use the AVC HIGH profile which offers several encoding options NOT possible in the MAIN profile used by Sony. Which MAY suggest that Pana intends AVC-CAM to compete with both HDV and HDCAM EX -- while Sony plans no such possibility. (Canon, has hinted at a "new" codec as well. Could Canon and Pana both support AVC-CAM???)

Sony may keep AVCHD, HDV, and HDCAM EX as three very separate product lines. And, both Sony and JVC may find a way to market "HDV" on non-tape based camcorders. They really only need a new marketing name given the fact that both already support "HDV" to disk and cards.

Thanks Steve. It also will be interesting to see how the recording time compares from the HF11 to the HF10 since the newer encoder also likely draws more current, unless Canon has done some remarkable new LSI design. It may not be a coincidence that the much larger capacity new battery is being released at the same time as these cameras..... (-8

Updated firmware is indeed another possibility, in which case the battery demands are less likely to reflect any differences bwteen the old and new models.

How are people using Mac OS X viewing these files? I downloaded a few and renamed the .m2tx to .m2ts (and later .m2s), and nothing I have on my machine will play it back. VLC crashes almost instantly. Other players refuse to even open them. I'm not sure what to do.

Chris, your system specs please?

I'm using a Mac Pro, 2x 2.66 Ghz dual-cores, running 10.5.4 with 6 GB of memory.

I also tried my 2.26 Ghz Macbook also on 10.5.4 with 2.5 GB of memory and had the same results.

For what it's worth, Quicktime is up-to-date on both my machines as well.

If the HV30 is to the XH-A1,

then would part of the significance of this new encoder be that
Canon has all the relevant pieces they need for anytime they
want to do a

HF11 to the all new ??-?? (You pick a name) ?

HF11 - smooth?
 

I would like to know if HF11 gives a smoother video than the HF10/100.
I tried a HF100 a minute in a store and my impression is that the video is jerky when you pan or shoot moving objects. Ive seen video samples that are shot outside and some of them look ok but I still sometimes see this jerkiness.

Shot in 50i. 25p gives a really unsmooth look even in the built in LCD screen. Totally useless but that's not my experience with other camcorders in 25p mode.
Was I recording with crazy settings that I don't know about?

Please have a look!
The panning is pretty fast just to show you what Im talking about.

2. Download with FileFactory Basic (Free)
http://www.filefactory.com/file/955e88/n/00003_MTS

This great little player will "de-interlace" the video while playing the clip and you also don't need to install it.
Just put it on your desktop.
http://sourceforge.net/project/showf...roup_id=170561

Here is an opinion that the hf11 looks horrible at 25p which matches the posted opinion on the old thread that the hg10 looks horrible.

Of course, this conflicts with the other review site -- the one we told not to trust -- that the hg10 looks horrible but the hf11 looks great. And, conflicts with others that have posted here the hg10 looks fine.

And, there are posted videos that show both the problems and the lack of problems.

This matches my experience with the EX1. I've been sent 24p video by folks saying it looks horrible and it didn't. Others think it looks fine.

interlaced
 

My clip is interlaced, did you watch it?

Im not one of those who think 24p/25p looks bad. I just think HF100/25p looks worse than other camcorders that I've tested (miniDV) and I even managed to make it look really bad in 50i. Maybe this is an issue for all AVCHD models?

I wan't to hear others opinions after watching the clip I posted. Did I do something wrong, do I have crazy codecs on my computer or does this camcorder have problems?

The real question is if HF11 has improved.

Chippe,

The pan was too fast. All HD makes fast and not smooth panning problematic, but notice towards the tail end of the clip where you briefly managed to somewhat follow the guy with the cap.

Your pan slowed down a bit and even though that guy was moving his face and cap/clothing detail came out pretty sharp. My impression is that if there is improvement in the HF11 I can't see it to any significant degree.

With my HF100 I have done tripod mounted pans of cavalry re-enactors at a fast gallop doing charges while firing cap & ball revolvers using cine mode and 30p that I'm not unhappy with. One fact of life with these small cams is that handheld work may have to be done with some kind of stabilizer like a lightweight double grip bracket, "fig rig", or something like that. I've built myself a "poor man's steadicam" out of pipe that seems to help some.

But I'm not looking to jump on the HF11. I have a pair of HF100's I'm getting used to working with, and I believe that is the key (getting used to something and working with it).

There is no cam around (pro or otherwise) that would look sharp with a pan that fast. The fact is that no decoder could possibly keep up with that motion. I don't know why people would pan that fast other than for some kind of 'effect'. The clip proves nothing.

Amen to this, this is perhaps the most overlooked thing about shooting HD - you cannot pan like a wild man or whip the camera around and expect the footage to look good - it WILL NOT HAPPEN!!!!

I remember all too well the first shoot i did with the old HC1... I was blown away by the footage when the camera was on tripod, and SOME of my handheld work, but much of it looked simply horrid... even with a rig that worked perfectly fine for SD shoots... learned quickly that stability is mandatory with HD. You need to start to think about how to operate your camera like it was mounted on a Hollywood type rig!

ALL camera movement becomes radically accentuated, so keeping the cam steady becomes "job one". To me it's easier to keep the light cameras steady (CX7, and the Canons), but even better is using the SR11 with the VF and a belt pod (3 contact points). I'll shoot "family" and non-critical stuff handheld in a pinch, but I'll use one of several approaches to stabilize if it's a shoot that I want to look good. Monopod is OK for some motion and if you need a "fixed" shot, dual bracket rig is best IMO if there's a lot of movement required, and I also use a monopod with a belt clip arrangement for some situations. I've even got a miniature handheld steadycam that will fly these small cams perfectly, but I don't use it that often, prefering the brackets most times.

Took me a while, but I've found some fairly compact brackets and monopods that don't add a lot of size/weight to the overall package and are quick to set up - still all fits in a small camera pack!

Until processors improve to handle the data/compression at the level our eyes/brains do, these things will remain "issues". Pretty amazing when you think about how OUR image processing works and how difficult it really is to recreate that!

It's funny Dave, that's an often neglected aspect of VF cams vs non-VF cams...steadiness. There's no question that while using the VF, your head/body serve as another point of stabilization. It's also far easier to hold anything steady when it's held close to the body as opposed to further away as a non-VF cam requires...especially if you use reading glasses! :)

Bruce: I agree about the HF11 from what I've seen so far.

I wasn't trying to follow the guy with the cap :-)
The HF100 was attached and I had to pull it really hard and shoot above my head.

Ken Ross: your comment is not relevant.
Im not talking about proof and sharpness, it's about the "jerkiness/flicker" and why some people see it, me included.
The panning is fast which I already wrote. It wasn't meant to be a smooth pan just to see this "stuttering" clearly.
All the objects that are unsharp by the fast movement (which they should be) are also flickering/"stuttering" in my eyes and if you don't see it I'll have to find out why I do it.

I just took a look at some scenes in Cloverfield in 1080p. We have the fast movements and the unsharp blurry picture but not the jerkiness.

OK Chippe, I see what you're saying and agree. I too see the jerkiness, but I've yet to see 24p or 30p for that matter, from any of these consumer HD cams that doesn't have this jerkiness. I tried it on the HF10 and it was bad and it's equally bad on my HV20. This is why I never use these modes.

Ken,

Thanks, I guess I have my answers. It would be nice to step down in size for my vacations.
I guess the HF100 is a good choice since HF11 doesn't seem to improve very much.

FWIW, looked at the clip frame by frame, and there was enough movement between frames to induce a pretty significant "jump" from frame to frame, and run at "real time" it was pretty bad, so it's not just your eyes... you can relax <wink>! It was almost painful.

I've looked at clips shot with my SR11 (60i), and didn't see similar issues. Some motion blur, but it was blur, not "jump"...

You'd either have to pan slowly so the "jump" is not so noticeable, or bump frame rates. I'd also suspect that if you could maintain a shallow DoF so there was less "stuff" in the frame to deal with, you'd have better results - these little cams have a pretty deep DoF, so that certainly doesn't help things...

blur versus "jump" is a function of frame rate (60i/30p) and shutter speed, not camera.

Dave,
You're right about the DOF.

Suen,
Very true. Don't forget my clip is in 50i although it was shot indoors it's still pretty bright. If we skip all these common reasons it's still a litle too "jerky" imo. Im also aware that everything is so much more significant in HD.

Thanks for all the comments.
I'll stop talking about it now, atleast in this thread :-)

Thanks Chris & Austin for the effort on this!

Chris - you raised a number of questions in your article, do you plan to add your impressions/thoughts as to how you see the answers? Would be great if you did

Chris,

i can't download this http://www.dvinfo.net/media/canon/hf11/ i obtain "access forbidden".

and i can't see in your review same footages did by hv20 to have a comparison. It's me I'm not seeing or do you have to upload yet?

Some considerations: i see feelings about this 24mbit of Hf11 are lower than the expectations and the improvement is hard to see. Maybe the reasons can be:

- size sensor of 3.2 is the limit and doesn't permit the scalability that characterizes h264 compression. So 17mbit are enough to h264 to "saturate" the sensor in the 90% of conditions. I mean that, if other conditions are equal, h264 is so efficient that needs architectural improvements to overwork the extra bitrate...

- you can see the improvement exclusively in fast motion (as fast panning or similar) or in very complex scenes...and seeing them on big plasma or on smaller monitor enlarging the image...

PS i give you my congratulations for the choose of test condition: infact the "water proof" in one of harder test to try a codec (in particular an inter-frame compression...).

ciao!

i am looking to upgradefrom Hv20 verses Canon hf11 -data please .
 

We primarly shot 35mm film
But we do special effects shots in digital for quick refrence work .
We do try to get the most out of these consumer cameras and they are generally impressive
I am almost ready to buy an hf 11 .We presently use a canon hv 20 on our 35mm filmed prductions as a quick refrence footage for our post effect department .
I like the hf 11 stated features -
however
as a buyer of the hf 11 am i about to get into the typical problem where
solid state out board memory cards are as expensive as this prosumer -consumer camera ?
Is thsi really the case or does the camera have good on board memory at hi rez?
with the canon hv 20 we only need to buy more tape .
We are trying get the highest resolution video image with some mionimal color space .
Thanks for a reply
Regards
Bruce

Sorry about that -- I'll try to get this resolved in a day or two.

Bruce, I'm not sure if I understand your question, but here goes...

The built-in flash memory is certainly good enough for full-resolution 24mbps video.

These days, SDHC cards are cheaper than dirt. You'll need a Class 4 or Class 6 card for full-resolution 24mbps video. They are laughably affordable. They are practically being given away. They are so incredibly common as to be ubiquitous. And they pay for themselves in no time. Hope this helps,

This is a myth that keeps being repeated. Yes, shutter-speed in relation to frame rate determines the amount of "judder" that is present. However, HOW it is SEEN is dependent on 5 more factors:

1) The way the edges of objects are "represented." Film, with it's very high resolution can display an edge, even a "sharp" edge, as the edge -- no more and no less. Expensive HD camcorders can do almost the same because they have a flat frequency response from DC out to the limit needed for recorded resolution (typically 1440-pixels).

Inexpensive cameras do not actually capture the detail you are lead by marketing to expect. The lens has a very low MTF -- compared to the $20,000 HD lens -- which means fine detail never makes it to the CMOS. Although the CMOS may have what sounds like high resolution, by the time the luminance is obtained by debayering RGB from the chip -- actual fine detail is lost.

In order to make the image look good, the mid-frequencies which carry EDGE DETAIL are boosted to a point just before where ringing starts. (Ringing makes the awful double edge that is called Edge Enhancement.) So now SHARPNESS is carried by the edges of objects -- not by fine detail.

You now have HARD edges, which the eye perceives TOO WELL. You now see the edges jump rather than see a scene with motion. In audio terms, you hear the singer pushed toward you and no longer part of the orchestra. Too much mid-frequency boost.

Note the BALANCE between lens and sensor resolution can be upset when three 1920x1080 chips are placed behind a prosumer class lens. This is why 24p judder can be seen on the EX1, but won't be seen on a $75,000 CineAlta. Or, a CineAlta with prime lenes. It's also why Panasonic DVCPRO HD Varicam wins praise from filmmakers. A high MTF lens passes information down to a relatively low rez recording format. And, no boost is used.

2) Real HD cameras have 2/3" sensors which minimize DOF. Our HD camcorders have tiny chips that cannot blur-out the background. So the background edges are easily seen thereby giving the eye extra information about edge movement. And, when video shooters don't pan WITH a moving subject -- the subject edges stand-out. The same thing happens when we pan at the wrong speed.

3) Larry Thorp, "inventor" of the CineAlta camera, adds a third variable: screen brightness. Unless you are a real videophile, you set your HDTV to a VERY high brightness. The higher the brightness, the more the edges stand-out and are perceived. In a movie theater, brightness is controlled.

4) To this I'll add that inexpensive cameras also try to create apparent sharpness by setting gamma to give a bit of extra contrast. Extra contrast hardens edges making them EZ for the eye to lock on. The expensive cameras have operators who know how to set gamma.

5) Most ALL monitors cannot pass through the full bandwidth need for 1920x1080. So once again, mid-frequencies are boosted to create apparent sharpness. And, once again -- edges are further hardened making them more visible. And, I'll bet the vast majority of HDTV owners have never calibrated their units and are running Sharpness at 50%. The correct value -- depending on the set's design, is ZERO to about 15%. So, HOW we watch can affect our perception of judder.

Bottom-line -- there is a reason why film and expensive HD cameras are used when "pros" need the frame rate to be 24p. And, there is a reason why VIDEO has almost always been shot at 50i/60i. Simply dialing in 24p and 1/48th doesn't turn any of our cheap cameras into film cameras.

Nicely put.

I'm stuck between choosing:
HV30/HF100/HF11

Sometimes shoot the camera solo (monopod/glidecam) when Im at track days, but when Im working on more serious stuff I use my Brevis35 - previously with the HV20.

I want to use the solid state recorders as its soo much more convenient, especially when I'm in racecars but if the image quality isn't on par with my old HV20, not sure what to do?
Need to purchase within a week really.