XL H1 and 24F

[Daniel Stone]

Forgive me if this sounds ignorant, but could someone explain to me how 24p and 24f differ? Someone told me that 24f works like the XL1's frame mode, which effectively cuts resolution in half.

Why didn't they make it a true 24p?

[Nick Hiltgen]

I think there are a coupld of threads about this. The 24F mode is because (I believe) the chip is not a full progressive chip but an interlaced one. No one is 100% on how the camera does it's magic though steve mullen has a theory that checks out with me. IF you read some of the other threads you'll find that the resolution is in debate as well. My findings suggest that the camera does much better in 24F then 540 (1080 cut in half) but I suggest you do your own research.

I believe that they didn't make it true 24p because it would have been really expensive (more so then it already is.)

[Chris Hurd]

Frame mode resolution is *not* cut in half. This isn't your grandmother's Frame mode from a 1998 XL1. Canon has had seven years to improve Frame mode.

They are not progressive CCD's because this is a 1080i camera. Frame mode produces the same results as progressive scan. It's a mistake to get hung up on "true" progressive vs. emulated progressive. How the image is made is meaningless compared to how it looks. What does Frame mode look like to your eye, that is all that really counts.

[Matthew Wauhkonen]

Chris, I hate to say it but that's pretty misleading.

I've used a GL2 and its frame mode cuts resolution not in half but reduces it by about 30% (pixel shift fills in the rest). I assume the same is true of the XL1.

The XL1h uses pixel shifted interlace CCDs and there is a CLEAR reduction in resolution. Quite visible on a computer screen, possibly not as visible on a TV. I'd say at least a 20% reduction in resolution (other tests have confirmed a loss, none have thrown out specific numbers.)

Still, this beats the Sony, beats the JVC, and beats the HVX. The Canon is sharper in 60i than 24f, but the softness in 24f reduces interlace flicker, helps the codec to deal with the high amount of information, and is so minor that it doesn't matter that much on anything but the largest screen. Plus, since it's interlace and not progressive, you don't lose a stop of sensetivity.

So, yes, you do lose resolution, and it's probably similar to what is lost on the XL1. The difference is that the XL1 has low res sensors, the XL1h has an absurdly high resolution sensor.

Edit: Of course, it is TRUE 24fps even if there is a resolution drop (a very minor one, though.)

[Jim Feeley]

Canon folks still won't talk much about what they're doing, but isn't it similar to what Sony does with 24psF for their HDCAM?

<http://www.broadcastpapers.com/telecine/Evertz1080p04.htm>

It's not perfect, but my experience with 24psF has been pretty positive (it better be at the rental prices we pay)...

Happy to be enlightened about this more, though...

Jim

[Shannon Rawls]

Chris is not misleading, he's right on point. Actually the only thing misleading is the title of this thread. I don't even agree with Nick with saying "they didn't make it true 24p because it would have been expensive". Nick, THEY DID make it 24p!

Guys listen...
What does the term "Twenty Four Pee" mean to you? The sensor type? NO...THATS WRONG. That's not what it means. Please use the terminology correctly, because if you don't YOU will in fact be 'misleading' the world. This camera will record every bit of 24 frames per second (which is why the title of this thread is a lie) and it will do it PROGRESSIVELY (one single hi-res/hi-def picture at a time). Additionally, it does it in the correct cadence and fashion that a film camera would have done it.
Heck, when you think about it, then there is NO VIDEO CAMERA that records to minidv tape that is TRUE 24P. I can't for the life of me figure out why Canon called it "F" other then they decided to be truthful, just like Sony is with the F900. Which makes me think, now that I've brought up the Sony F900......
The XL-H1 is no less 24p then the $100,000.00 F900 is 24p, but I don't see anyone jumping down the F900 throat calling it "fake". It's amazing the level of miseducation people have when it comes to video cameras and terminology like the word "Fake" and "True" and "Real". *smile*

Here's a little education for those that don't know...: "The F900 CineAlta cannot record true progressive-scan, it can only record interlaced, so it breaks progressive frames into interlaced data for the purposes of recording it (which they then call PsF, for Progressive: segmented frame)." Can you guess what expert made that statement who I whole-heartedly beleive? Well, if that's what the Sony does, then what do you think the XL-H1 is doing?

And Matthew....where did you "hear" about the resolution loss of 24f vs. 60i? I know you couldn't have "seen" it, because I don't think it's there. You can't determine this on a televison "OR" a computer screen by just looking at regular ole' day-to-day footage as you would lead someone to beleive Matthew. What your eyeballs seen was 60i motion vs. 24p motion, not resolution loss. The only way to KNOW this is to TEST it with a resolution chart on a Professional High Definition Reference Monitor....something I DID with Nick. I would even like to retract part of my statements in my test as far as a resolution drop and what not, because now that I slow down and think, I kinda remember 24f being as good as OR HIGHER then 60i on the chart.

Now...can we please move on about this. The XL-H1 is a 24(pee) camera. If you wanna call it 24(eff) then fine, but it's all the same in the end. And I challenge anybody elses HDV camera to a head to head duel if you doubt me. If you still contend that this is 'fake' (lol) then let's call the f900 & jvchd100 'fake' as well. Leaving Panasonic DVX100 & HVX200 the only 'true' 24p cameras in exsistence. Fair?

*smile*

- ShannonRawls.com

[Daniel Stone]

Whoa, easy killer! I wasn't trying to put the XL H1 down, I swear. I'm just confused and not sure whether to buy it or not. I just almost bought it this weekend and didn't because of talking to bystanders at B&H.

First, thanks for the info, everyone.

I went to B&H in NYC this past weekend to play with some of the High Def cameras (and to purchase one). The XL H1 was there on display but not hooked up yet. I played with one of the other ones (don't remember which one) and switched it to "frame" mode. What I noticed was a fake strobing effect which looked remarkably like that cheesy 15fps 80's dance video effect - I'm very familiar with the 24p look (as I have the XL2 and the DVX) - but this was downright horrible. I talked to two guys who joked the "F" in 24F stands for "fake", as in "fake 24p". The sales guy said it works just like the XL1s's frame mode. Thinking about how much resolution loss this fake frame mode creates in the XL1s, I figured it creates the same percentage of resolution loss in the H1. I mean, same process... just HD. Why wouldn't it?

One of the guys I talked to has the H1. He said that there is a slight loss of resolution between 60i and 24"fake" (I'm just doing that as a joke, now) but, since we're working in HD, it's not so noticeable. He said that 24f is still a more amazing picture than any of the others in 60i even.

I left there scratching my head, wondering why Canon would do this. I did a little research (as I know very little about HD) and found that there just aren't any HDV camcorders with true 24p.

Then I found out that the H1 doesn't even record in fake progressive, but rather "flags" the footage for "progressive-looking playback". Sheesh. Now I'm afraid of that aforementioned "strobe" effect you find on 80's VHS camcorders.

As long as that 24f looks like the 24p of the XL2 only higher resolution, I'm a happy camper. Anyone? Anyone?

[Shannon Rawls]

Originally Posted by Daniel Stone

The XL H1 was there on display but not hooked up yet. I played with one of the other ones (don't remember which one) and switched it to "frame" mode. What I noticed was a fake strobing effect which looked remarkably like that cheesy 15fps 80's dance video effect

That would be the Sony FX1/Z1U you picked up. And let's not use the word "fake" for cryin' out loud. How about "Suedo" or "Wierd" or "Unpleasing".

Originally Posted by Daniel Stone

I talked to two guys who joked the "F" in 24F stands for "fake", as in "fake 24p". The sales guy said it works just like the XL1s's frame mode.

The two guys you spoke with are on Payroll from the other manufacturers to hang out at B&H and defract customers from buying the XL-H1. Job well done fellas. *smile* And the Salesman had no idea what he was talking about (odd for B&H, but possible)

Originally Posted by Daniel Stone

I left there scratching my head, wondering why Canon would do this?

Best thing I've heard in a while.
Danny, do you think by any stretch of your imagination, that Canon....of all companies....the imaging MASTERS........after successfully conquering the 24p technology in the XL2 camera in 2004........would release their brand spankin' new FLAGSHIP CAMERA in 2006.....with a $9,000.00 price tag.....
and muck it up?????? You're a smart dude, you know better then that. *smile*

The XL-H1 is every bit of an XL2. You should have bought the camera and laughed your way out the door at those 2 little devils on your shoulder. *smile*

- ShannonRawls.com

[Chris Hurd]

Originally Posted by Daniel Stone

there just aren't any HDV camcorders with true 24p.

Incorrect... the JVC GY-HD100 has 24p from progressive scan chips. Of course it is HDV1, at 1280 x 720, instead of HDV2, at 1920 x 1080, but it is still HDV.

[Chris Hurd]

If you compared 24F from the XL H1 to 24P from the XL2, I think you would be hard pressed to notice any difference (to make it a fair fight, you'd need to be in 16:9 SD mode, not HD mode, on the XL H1).

[Shannon Rawls]

Originally Posted by Chris Hurd

Incorrect... the JVC GY-HD100 has 24p from progressive scan chips.

Actually it 'ends up' with 24p in the end. It's 48p, 50p & 60p that is scans and only "60p" that records I beleive, not 24p. Even the .m2t files pulled off the tape are 59.94fps when shooting in 24p mode. It's just only 24 of the frames are encoded and we have to rely on great software like CineForm to decipher the stuff to 24p, in the end.

So theorhetically, if you go by all the naysayers hypothesis', then the JVC HD100 IS NOT a "TRUE" 24P camera either! This is what Douglas Spotted Eagle is meaning when he says Canon is telling the "truth" when using "f" instead of "p".

(now ofcourse I'm being silly here), but I'm saying that to illustrate a point and hope you see my relevance of terminology like 'true' and 'real' and the way haters can misconstrue things to those that don't know any better.

We know the HD100 is in fact a 24p High Definition Video Camera. But don't give it that title because the chips are progressive. Give it that title because thats what you get when you shoot with this camera and end up with the footage.
Going along those same lines, then we should also know that the XL-H1 (in every honest sense of the nomenclature) is in fact a 24p High Definition Video Camera as well and not get hung up on the "f" that Canon decided to use in stead of "p". (in DV mode, it can be 3-2 or 2-3-3-2, your choice) and all that jazz.

It's all about what comes down the firewire pipe and what you get from that MiniDV tape! That's how we look at it for the Sony F900, then rightfully so, we need to look at it for these sub $10k cameras as well.

- ShannonRawls.com

[Daniel Stone]

So is there a huge difference in resolution between HDV1 and HDV2.

And does that mean that even when in SD mode there is no true 24p mode on the H1?

I guess that, all in all, I may just not be ready for HDV yet... it seems that there are still so many "buts" (I'm talking all HD cams, not just the H1): It films 24p... BUT not real 24p (and it doesn't even really film in anything other than 60i). It has so much resolution... BUT you lose a lot of it in MiniDV compression (unless you spend even more money to be teathered to an external acquisition device). Great picture all-in-all... BUT horrible audio.

The biggest thing is just that "un-real" (kinda like "un-dead") 24p. If it's not called "fake", then what does the "F" in 24f stand for? I'm joking.

Until I get to experience the 24f myself, I'm definitely not biting my nails to own it.

Plus, I'm not convinced that the H1's HDV displayed on an SD display looks any better than the XL2's SD.

[Barry Green]

Shannon, you realize when you make these proclamations, you're directly contradicting Canon themselves, right?

Canon says that they will not call it 24p because it is not a progressive chip. It is scanned interlace. The 24F effect is created by the DIGIC II processor.

There is a difference. It may not make a difference to you, but that does not mean that the difference doesn't exist. You yourself used to try to tell us that the Sony CF24 was the same thing, it just needed to be "processed in post" -- and now that you've had some experience with something that looks and acts like the real deal, you now say that the Sony's CF24 is not the real deal, but instead it "looks like Rerun from What's Happenin' doing the Pop-Lock."

So let's stick with what reality is. Reality is that the Canon does not use a progressive scan chip. It uses an interlace chip. That's what Canon says. That's what the truth is. You may not feel that it makes a difference as far as your work goes, that's fine and dandy -- but you should at least specify what IS a difference, vs. what you weigh as a deciding factor.

The next step of reality is that there is a noticeable resolution loss that happens when going to 24F mode. This can easily be seen in Pete's comparison charts on his Scratchpad site. Pete shot the exact same chart at the exact same framing using the exact same settings, changing only the mode from 60i to 24F. The drop in resolution is very easy to see. I've extracted a patch from the identical segments of the chart to show the difference:


Is it night and day? No. Is it obviously lower in res? Yes. How much lower? Depends on your particular interpretation; I'd say that in this circumstance, using this framing, the 60i is resolving cleanly to around 750, whereas the 24F crosses lines right around 600. A difference of about 20% lower res. Not too far off from what "frame mode" used to do.

In true progressive scan, that's not what happens. True progressive scan causes an INCREASE in resolution, not a dropoff, because it doesn't have to do interline filtering to reduce interlace flicker. I'm talking about scanning the CCDs progressively, not the recording format. The recording format is the least of our concerns; the Sony CineAlta records as interlaced but it scans as progressive, and that's what counts.

Is there a difference between 24F and 24P? Yes. Absolutely. Obviously. Unquestionably.

Does that make 24F "bad"? No, not by a long shot. 24F is probably the best HDV recording mode going. It provides genuine 24-fps motion rendition, unlike CF24. Its lower detail helps the MPEG encoder handle it easier. And the fact that it records 24 discrete frames, rather than a stream of 60i, gives it 20% more bandwidth to apply to the frames, as well as letting it use a more efficient progressive encoding rather than the less-efficient interlace encoding. 24F should be more resistant to mosquito noise and more resistant to macroblocking than 60i. 24F, and HDV's 24P, are likely to prove the most robust, most resilient ways to record HDV material.

But it is about 20% lower resolution than 60i. It's a good simulation of 24p progressive scan, but it is not the same.

[Barry Green]

Originally Posted by Daniel Stone

So is there a huge difference in resolution between HDV1 and HDV2.

Depends on how you define resolution, and what you're recording.

There's a huge difference in the # of recorded pixels between the two formats. But as far as resolvable definition from the camera heads, there's much less difference.

If you're talking about recording computer-generated imagery, yes HDV2 could retain much more of that detail. But if you're talking about a complete camera system and lens and chips and the final image that gets recorded to tape, there's not much difference in recorded/resolvable detail between an HDV1 JVC HD100 and an HDV2 Sony Z1. However, for temporal resolution HDV2 does hold a substantial edge.

And does that mean that even when in SD mode there is no true 24p mode on the H1?

Well, yes and no. Yes, in that there is never actual progressive scanning going on. But no, in that by the time you get to SD, the loss of fine detail from 24F scanning is going to be moot because when it gets scanned at SD resolution it's likely going to have just as much detail as a true SD progressive-scan chip would deliver. I haven't tested this, but I'd guess that as far as resolved detail goes, SD 24F from an XLH1 and 24P from an XL2 are probably going to be indistinguishable.

The biggest thing is just that "un-real" (kinda like "un-dead") 24p. If it's not called "fake", then what does the "F" in 24f stand for? I'm joking.

"Frame". The DigicII processing appears to be an enhanced Frame Mode.

[Barlow Elton]

Originally Posted by Matthew Wauhkonen

Plus, since it's interlace and not progressive, you don't lose a stop of sensetivity.

Actually, you do lose some sensitivity in 24F. It's quite obvious when you switch the dial. This is the mystery to me...is it actually doing a true PsF, ala F900? Whatever it's doing, it's phenomenal. I can't believe how filmic 24F is.

...not that I showed anything overwhelmingly cinematic. I just know that if that's the WORST that it can do, imagine it in production scenarios.

The mind boggles