Haven't noticed anybody mentioning this yet... Canon USA has updated their XL H1 specifications page (http://consumer.usa.canon.com/ir/controller?act=ModelTechSpecsAct&fcategoryid=165&modelid=12152) to include stated lux ratings (whatever those are worth... most folks here know that I am no fan of minimum lux ratings at all). Here they are:

60i at 1/60 shutter speed = 7 lux
30F at 1/30 shutter speed = 4 lux
24F at 1/48 shutter speed = 6 lux

Compared to the XL2 at 5.5, 6.5 and 10 respectively, that's not bad at all considering the much denser chips! I wonder if the latitude is going to get a bit hit?

How can that be: these disparate values?

7 <-> 5.5
4 <-> 6.5
6 <-> 10

How can that be: these disparate values?

7 <-> 5.5
4 <-> 6.5
6 <-> 10

An what are the lux ratings of the JVC GY-HD100?

This week's HDV@Work (from Video Systems magazine) has my REVISED story on the XL H1. I cover 24F and the method of recording 24F and 30F.

Missing from the HDV@Work story is this very interesting information:

The HDV@Work story has been revised to reflect Canon’s decision that all information on the process by which “24F, 25F, and 30F” video are generated is proprietary information. This has several ramifications. First, Canon takes the position that any information imparted by Canon employees, distributors, or dealers on the “process” by which these formats are generated cannot be considered valid by the press or potential customers. They claim the actual details are known only by a small number of Canon employees. Second, unless this—in my view highly unwise decision—is reversed, no information on the process will be forthcoming from Canon.

I found out Canon's "position" when they objected to 24F information in my first version.

Also missing from the HDV@Work story is this consise sensitivity data:

Canon provided sensitivity data for the XL H1 in both 1080i60 and 24fps modes. With the standard kit 20X HD lens (at Wide Angle) in Auto Mode and with a Gain of 18dB, the data are:
60i, 1/60 shutter speed = 7 lux
24F, 1/48 shutter speed = 6 lux
30F, 1/30 shutter speed = 4 lux

For the XL2, comparable data are.
XL2 with the standard kit 20X lens (measured at Wide Angle) in Auto Mode with a Gain 18dB:
60i, 1/60 shutter speed = 5.5 lux
24P, 1/48 shutter speed = 10 lux
30P, 1/30 shutter speed = 6.5 lux

These data indicate: (1) The H1, whose CCDs have about 2.5X more elements, is only slightly less sensitive than the XL2 in 60i mode, which is a major accomplishment; and (2) Canon’s 24F will deliver almost a half a stop greater sensitivity than shooting 24p using the XL2.

Progressive-scan is about a stop slower than interlace. That's why on the XL2, 24p@1/48 needs about twice the light as 60i@1/60.

On the XL H1, it's not progressive scan, so they don't take the "hit" in sensitivity that true progressive chips take. It generates its 24f and 30f off of an interlaced scan, so 60i @ 1/60 (7 lux) is about as sensitive as 24f @ 1/48 (6 lux). The difference could probably be attributed to the shutter speed differential.

Which has me all the more firmly convinced that "24F" is the same thing as Sony's CineFrame 25, just clocking the CCD at 48hz instead of 50hz.

[QUOTE=Barry GreenWhich has me all the more firmly convinced that "24F" is the same thing as Sony's CineFrame 25, just clocking the CCD at 48hz instead of 50hz.[/QUOTE]

I agree -- I think they are using Sony's next gen CCD -- because they need a far smaller volume that does Sony -- and Sony's DSP chip. These new CCDs likely can be produced only low volume and are very expensive.

When I asked Canon to confirm a "constant sampling interval" they declined to confirm this. Which, seems strange if they are really clocking at 48Hz. You would think they would want to trumpet this advantage over Sony.

Which makes me think Canon's sudden desire to confirm nothing -- which came a week after I published our preview of the H1 -- was due to pressure by Someone who is making parts for Canon.

The decision was made in Japan and does not necessarily reflect the viewpoint of Canon USA. That increases my belief this is a political decision to prevent the loss of face by another Japanese company.

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

If it is Sony who is making the CCDs, then it indicates that Sony's Z1/FX1 replacement won't have 24P either. But will they go 24F or stay with CF24?

It also make me wonder why Canon may have gone back to Sony for CCDs and not used the new Panasonic CCDs. And, why they went 1080i rather than 720p60 at $10K. One possibility is that Sony did the entire electronics design for Canon.

If it's the case, I find Canon's position to be precarious at best! To borrow another car analogy, it's like Chrysler buying Ford engines for its cars. One can only drive a certain political correctness road in order to keep face for so long. Alas, such uneasy alliances can be very detrimental, not only financially, but consumer satisfaction-wize too.

Electronics is not the same as cars. They buy parts from each other all the time because the productions CAN be compared and sometimes for a company there is not enough reason to build all their own components if they don't use enough volume.

For example Canon makes glass so all Canons use Canon glass. Panasonic sometimes uses leica glass and Sony uses Zeisis glass sometimes. Panasonic and Sony aren't renound glass makers like Canon and therefore have no need to create their own special technology glass so they buy it from other companies, the same can be said for ccds and all kinds of other parts inside video cameras.

It's not uncomon.

iT'S DEFINITELY NOT UNCOMMON. Japanese companies tend have the strong desire to do everything. So at the Sony Showroom in Ginza there was a Sony Furniture area. There is Sony Banking and Sony Insurance.

When Toshiba wanted a pro Hi-8 camera, they used Sony designed boards made in Toshiba factories.

But the real key is that almost none of the products made by the big three are designed and built by the big three. Their marketing departments come up with a specification. Sometimes it goes to only one "loyal" subcontractor. Loyal because they have financially helped him grow his business. Loyal, because like a favor from the Mob, he is now dependent on them. This loyalty ensures the price for the new camcorder will be VERY low. And, he can be squeezed, if for example, the yen/$ changes.

Other times, the design may go out for bids amoungst the loyal sub-contracrors.

When a company is chosen, S/P/J engineers live with him as the product is developed and built. (Parts are often bought from the big company so money flows back.) Finally the finished products are put in "Sony" boxes.

In the old days, these sub contractors were only in Japan. (And they sub contracted parts even to housewives!) Now they are in Taiwan (for design) and China (production).

At the consumer level, this practice enables Sony for example to have a rolling string of new products. It orders, say 10000 units at time.

The JVC HD1 was orderd in batches of 2000 a month. If it sold well, they could order a batch more. If not, they could take the minimum contract order. If these don't sell, they are sent to discounters.

Dell has used this model to grow with almost no R&D costs and no factories. Apple is now using it.

The good thing about it is that with much/most of the design being done elsewhere, these design companies get invaluable experience while Dell/Apple helps them. And Dell/Apple saves a bundle by not having to pay for American R&D.

Eventually, the design company may design their own product. Or engineers may leave and start a company. I strongly recommend Olivia Syntax LCD HDTVs. A combination American marketing/Taiwan design that contracts to China for production. Like Dell they buy parts where-ever.

Like Sony's new Bravia LCD HDTV -- the LCD is from Samsung. Who by the way has filed many. many more patents these last years as has Sony.

Bottom line, not only is Canon dependent on Sony, increasingly Sony is dependent on Korean companies. And, they may all buy parts from China.

Which has me all the more firmly convinced that "24F" is the same thing as Sony's CineFrame 25, just clocking the CCD at 48hz instead of 50hz.Barry, You mean Cineframe 24, right? If 24F really shows the same problems, like Cineframe 24 does, then 25F (at the PAL version) would be the better choice for shooting "progressive style", I guess. I hope, Adam J. Wilt will make a test to answer this "constant sampling interval" question very soon.

Barry, You mean Cineframe 24, right?
No, CF24 is a very weird, very different animal entirely. I don't expect that 24F has anything in common with CF24 at all.

What I was saying is, Cineframe 25 is one field read out at 50hz. If Canon reclocked the CCD to 48hz and extracted one field, that could provide them with 24F. And it would be far superior to Cineframe 24, no matter how you slice it. It would not, however, validate Canon's claim that it's the equivalent of 24p -- not by any comparison of resolution. It would 100% match the "feel" of 24p, and the temporal motion sampling of 24p, but it would be much lower resolution. I figure that's what they're doing, and using the Digic II DSP to interpolate the other field.

Ah, now I understand. A very interesting idea: Canon does not interpolate from two fields, but only uses one field to generate the second. Well, we will see soon.

Ah, now I understand. A very interesting idea: Canon does not interpolate from two fields, but only uses one field to generate the second. Well, we will see soon.
Remember, that's guesswork. That's what Sony does with CF25. But we don't know what Canon does, and according to Steve Mullen they've pretty much zipped their lips and will not share details of what it does either. So we're all just guessing...

It also make me wonder why Canon may have gone back to Sony for CCDs and not used the new Panasonic CCDs. And, why they went 1080i rather than 720p60 at $10K. One possibility is that Sony did the entire electronics design for Canon.

Maybe Panasonic didn't want to work with Canon? If they gave the use of their 720p/60 chips to Canon wouldn't that steal sales from the HVX200? Panasonic has made a major point of showing the flaws of HDV. Why would they want the chips that allow the major advantage they have over current HDV cameras (motion) to be used in the HL1?

Just a thought.

Dan Weber

There seems to be a consensus among wedding videographers that HDV is almost unusable because of its inferior low light performence. Either they are exxagerating because they want an excuse not to invest in any new equipment or they have a valid point. Since I have only shot one wedding I do not have enough information to confirm or deny their objections to the HDV format.

Ah, now I understand. A very interesting idea: Canon does not interpolate from two fields, but only uses one field to generate the second. Well, we will see soon.

No -- Canon does use two fields, Simplified:

1) Objects not in otion -- like a rez chart -- will have about 800-lines because both fields can be used.

2) OBJECTS in motion have lines pulled from only one field. So there is smart software that finds areas that change positipn from onefield to the next.

Now, these objects have a V rez of only about 400-lines. So smart software takes these objects and interpolates the in between lines. Depending on HOW they do it, they can get a slight boost in V rez.

The point is that on static rez tests -- there will be no difference between 60i and 24F. Great marketing!

When objects are in motion they naturally blur and lose rez. You'll have a hard time seeing moving objects lose V rez. (Remember, moving objects naturally lose H rez too.) But, I'll bet in some situations in will be apparent to some. For example, a pure vertical pan that is slow may look like s**t.

That will lead to another flame war like SSE and CF24. :)

If Canon reclocked the CCD to 48hz and extracted one field, that could provide them with 24F.
Just had to chime in here...

Perhaps I'm totally confused, but couldn't Canon just be deinterlacing on the fly using a really good smart algorithm? They would be using both fields. When there is little motion, the deinterlaced frame would be nearly identical to standard 24P. When there is motion, there would be a hit in resolution. But, since they are using a GOP length of 15 frames I believe, wouldn't this give them about a half second to analyze what is changing in each frame over time and allow some type of smart predictive deinterlacing?

I use a VirtualDub deinterlacing filter all the time, and the result looks quite good. Couldn't something like the new Digic DV II chip the Canon has do an even better job in real time?

Sorry if I'm totally out to lunch.

Shawn Kimmel

--
Forever on DVD
shawn@foreverondvd.com
http://www.foreverondvd.com

There seems to be a consensus among wedding videographers that HDV is almost unusable because of its inferior low light performence. Either they are exxagerating because they want an excuse not to invest in any new equipment or they have a valid point. Since I have only shot one wedding I do not have enough information to confirm or deny their objections to the HDV format.

Now how in the heck would they know if they never used a HDV camera? The only time I see low light being an issue for HDV is in a totally dark reception but even if they are using a 2/3" camera they should be using lights.

I really do think it's an excuse. I think the current HDV cameras are about what the previous generation of DV cameras were at. Heck I still sometimes bust out my Canon XL1 as a backup and while it doesn't look as good as my newer cameras it is half way decent with a light. I'm sure every HDV camera is at least more sensitive than my old XL1 so I am not complaining at this point.

There seems to be a consensus among wedding videographers that HDV is almost unusable because of its inferior low light performence.

Not so: many wedding videographers are already switching to HDV. It is true that current HDV cameras are not as sensitive in low light as some of the best DV cameras, but if you boost the gain and use a small light when needed it works fine. I've shot five weddings in HDV this year and am learning to work with the limitations of the equipment, just like you would with any other camera. Used properly the image quality of HDV is impressive, and to me that's worth the trade-off in poor lighting.

In regards to the Canon XL-H1, I don't expect it to be very popular for wedding videography due to the high price. Maybe if Canon ships an HDV version of the GL2 for under $4000 they'll get more attention, but Sony's already pretty much captured that market segment.