XL2 Low Light Capability

[Don Palomaki]

The 5.5 Lux rating is listed at 1/60 in 60i, auto mode. Many camcorder camcorder set the lux rating at their slowest availabel shutter speed.

The XL1s was 2 lux at 1/8 shutter speed, which converts to about 16 lux at 1/60. Sounds like the XL2 could be about 1.5 stops faster.

Be interesting to hear field reports from users and what can be done in post befor ethe image becomes unusable for wedding video purposes.

[Rob Lohman]

Just to add a quote from Canon:

The three CCD image sensors are specifically designed to capture as much image detail as possible and for shooting under extremely low light conditions. In super low light, the XL2 captures crisp and clear digital data. Under extremely bright conditions, the XL2 greatly reduces vertical white streaks and smears

Now on to the real-world tests indeed!

[Steve McDonald]

Somewhere on another thread I mentioned that the 4:3 sensing area of the XL2 CCDs was only 71% as large as those of the XL1. Actually, the 71% figure applies only to the reduced width of the 4:3 area. Calculating for the reduced vertical size as well, the total 4:3 area of the XL2 CCDs is only about 51% of that of the XL1 CCDs.

This also means that compared to the Sony VX2100, the XL2 CCD size for the 4:3 area is just 51% as large. The XL2 has 345,000 pixels in this area and the VX2100 has 340,000. So, the XL2 sensing area is not only approximately half the size, but its pixels are only half as large as those of the VX2100.

Since the VX2100 has remarkably sensitive CCD pixels and is the acknowledged best performer in low light, how could the XL2 compete with this Sony camcorder in this respect? Could Canon have acquired new CCDs that somehow were so much more sensitive and effective, that despite having only half the size, they equaled
the Sony sensors' performance?

It appears to me that in order to provide a true 16:9 mode, that the standard 4:3 mode has been compromized on the XL2. Of course, the smaller CCD sensing area of the XL2, gives its lenses a lot more magnification effect or "angle of view", far outdoing the VX2100 in this respect.

With the excellent 1.6X telextender lens piece that fits under the XL2 lens, this lens power is increased even more. I wonder how much of a role the desire to produce more effective lens power, played in Canon's decision to shrink the CCD sensing area?

So, you have two of the top five semipro camcorders, the XL2 and the VX2100, one having great image quality and low-light capabilities, but only mediocre lens power and the other with a powerful lens system, 24p/30p scanning, but probably less low light performance. It would seem that the choice of using one or the other, would be easier to decide, now that the XL2 has been given these more extreme characteristics (at about double the cost of the VX2100, not including extra lenses). As I noted long ago, no one camcorder is ever going to be given all the features and performance capabilities that anyone would want.

Steve McDonald

[Barry Goyette]

Steve,

Another way of looking at this is that the sensor area of the 4:3 mask on the xl2 is identical in dimension to the gl2's effective pixel area. Due to its lower pixel density (345k vs 380k) we could expect a modestly improved low light performance, when compared to the gl2. Additionally, I have a strong suspicion that the xl2 is using sony chips..that 680k number has a ring to it and it ain't panasonic.Add to this, canon's claim of new signal processing and their improved lux ratings (5 lux vs. the xl1s's 16 lux at the same fstop), what we are seeing is a substantial improvement over the xl1s in terms of low light performance, and the "potential" for the xl2 to come very close to the performance of the sony cameras in this respect. Most user reports from people who have actually used both cameras suggest that the difference between the xl1s and the sony cam's is about 1 to 1.5 stops.

The thing is... all of these cameras do an amazing job at handling "low" light. Unfortunately there are a lot of amateur videographers that think it'd be cool to shoot in "no" light. Well it isn't...the results in no light situations are hardly ever any good, as the gain and color balance issues degrade the signal beyond hope. Better to get a handy cam and turn on the nightshot...at least you get the cool green glow.

I think this camera is targeted at the serious amateur/professional that wants to work in the minidv format, and most of us know that you can't just walk into a cave and hope the LED's from our watches will be enough.

One last thing...I do agree with you that the xl2 does appear to have compromised the 4:3 format somewhat with this camera. On the other hand, I've used a gl2 for several years, and have never considered it a compromise. It's resolution is certainly the equal of the sony cameras, and only a hair below what i've seen from my DVX100. The biggest improvement the DVX offered in terms of image quality is in progressive mode, thus I have little doubt that the 4:3 image on the xl2 will be a big jump up from the xl1s.

I think that canon made the right move...more people are watching content in 16:9, more of us own HD sets...having a camera with full resolution 16:9 seems like the right direction to me. The xl2 is an appropriate bridge product while we await the mystery that is HDV.

[Steve McDonald]

Barry, I have no doubt that a camera with .236-inch, 4:3 CCDs can indeed produce an excellent picture and do well in limited light. An example is my Sony TRV730, a MegaPixel Digital8. It's effective sensing area, with 690,000 video-active pixels, is only about .21-inch, on a single CCD. Yet, this camera delivers an outstanding picture, for a 1-CCD model, especially considering its price. It does quite well in limited light. Any complaints about low-light functions with it have been from people who used the J-PEG still picture capture and didn't understand the light limitations of its very fast mechanical "progressive" shutter for that mode.

So, with much larger pixels on a slightly bigger sensing area than my TRV730 and 3-CCDs to boot, the XL2 could very well produce a great picture. My contention is that relative to the VX2100, which has
twice the CCD sensing area and pixels twice as large as the XL2, it could hardly be expected to match the VX2100 in limited light. This would include most indoor situations, except where special lighting is provided.

But, we'll have to wait and see how the XL2 does in the real world, where number and size comparisons don't count much against actual, observable performance.

Steve McDonald

[Ignacio Rodriguez]

Hmm. Different companies seem to have different standards for 'minimum' light, even Sony publishes different low light ratings for the same cameras in different markets. So Canon's 5.5 lux might be close to Sony's 1 lux than expected. You also have to take into account that the 5.5 lux rating might be with the standard lens, and possibly you can attach a lens with a wider f/stop. Also, if you use 30p or 24p the camera may become 6dB more sensitive to light or more, as I believe the DVX100 does.

[Robert Mann Z.]

the low light cabality of this camera is better then the xl1s after doing some tests...

for one thing you can go 18db and use noise reduction to get a softer image but blacks stay black and the shot is saved

for more info see my frame grabs

http://www.emptyloft.com/xl2/

[Mark Grgurev]

Wouldn't the NTSC version technicly have better low-light than the PAL version? The pal version says its low-lux rating is 5.5lx at 50i at 1/50 shutter; the NTSC version says the same for 60i at 1/60 shutter. Once you turn the NTSC model to 24p shouldn't the low light be better since the shutter becomes 1/48 thus allowing more light in?

[Chris Hurd]

Actually you can shoot at 1/24 in 24p mode.

And as far as the NTSC models are concerned, why stop at 1/60. You can shoot at 1/30, 1/15 or 1/8 even.

Then there's the gain... go up to +18db.

Low light performance is highly subjective. What is acceptable to one person at 1/15 gained to +18db may not be acceptable to another. How you define "low light" might be different from how somebody else defines it, so I'm not sure how a superlative like the word "better" would have much relevance here.

Just add light!

[Don Palomaki]

To be meaningful the low light rating should include the follwing information:

incident light level (e.g., 12 lux)
aperture (e.g., f/1.8)
shutter speed (e.g., 1/60)
gain setting (e.g., 0 dB)
Output (e.g., 20 IRE)
Target reflectance (e.g., 18% gray card)
Video signal to noise ratio (e.g., 30 dB) for luma and chroma.
numbers in above example are for illust5rationpurposes only

Anyone seen those sorts of number published for any camcorder recently? I've heard that some camcorder makers have used something like slowest shutter, max gain (without regard for noise), max aperture and something like 20 IRE from a bright white (~94%) target to deliver their low light spec.

[Norman Woo]

Hi folks

In addition to the XL1, I also own the Sony PD150 used mainly in Wedding reception halls and other low light situation. It pretty well beats the Xl1 hands down. Color balance in low light is much better than the XL1. With the Xl1, the color balance icon would always be blinking indicating that it can't get a lock on the white balance.

How does the XL2 compare to the PD150 under low light conditions? If it betters it, then this is good news for me. In any case, I'm waiting with abaited breath as the XL2 is being launched here in Canada on the 16th. Woohoo!!!

Regards

[David Walding]

I really want to buy this camera but after looking around here and asking one or two people. I have still don't know if the XL2 is any better in low light than the XL1s. I don't care about lux ratings I would like someone who has used both and been able to compare the two. I shoot music videos and weddings and I need good low light performance and the XL1s was really disaponting for me and don't want to make the same mistake twice. I know the PD-170 and DVX are good low-light performers but I like the XL's form factor and would like to use it's new features. I won't buy it though if it handles like the XL1s in lower lighting conditions. Resorting to 12db or higher gains which makes my footage look like I switched to a Canon ZR-40!

[Peter Koller]

Hi, found this on a german forum:

http://www.wrc.de/xl2spuk.wmv

It is footage from a basement with no extra light but a small flashlight.

Setting was +12db and 25p (PAL model)

Looks great, almost no noise in the material!

Cheers, Peter

[Yang Wen]

Footage does look noise free but there is tremendous purple fringing around the basement windows.

[Frederic Segard]

Too bad it wasn't at full resolution. It's true; the low light is not grainy at all. I really wonder how it compared to the Sony PD170? I also noticed the purple tint, but I'd probably attribute that to the sunlight CTB, and probably some slight color correction.