View Full Version : GL2 resolution


Frank Granovski
July 22nd, 2002, 01:30 AM
Has anyone done a playback resolution test with the GL2?

Barry Goyette
July 23rd, 2002, 08:20 AM
Frank

I should have my gl2 in hand today, and as I have an xl1s, and a gl1 as well, I will put up some comparison shots of the 3 camcorders. I spoke with chris at macworld, and I think he will make them available on here on the site.

Anyone know of a good downloadable resolution chart?

Barry

Chris Hurd
July 23rd, 2002, 08:42 AM
By the way Barry it was great to meet you. I'll post anything you care to submit, on my forthcoming GL2 pages for dvinfo.net -- thanks,

Frank Granovski
July 23rd, 2002, 10:23 AM
Thanks. I was hoping someone would do a playback through a waveform monitor. The comparisons are welcome!

Barry Goyette
July 23rd, 2002, 01:00 PM
just a quick observation or two...I will post more in a day or so. These observations are just that, I haven't put up a test chart or measured anything, just viewing the same scene with gl1 and gl2 on a standard ntsc 9inch monitor. I've left settings at default and shot in manual and auto modes

The resolution of the gl2 is substantially better than the gl1. The contrast is slightly better (lower). The focus ring offers a little more resistance than the gl1 (and xl1s) --which makes it feel more like a manual lens when focusing--it is much easier to find sharp focus with this camera than either the gl1 or xl1s. The auto zoom speeds are smoother (this may be just a "new camera thing"..as I've noticed that my gl1 and xl1 got less smooth over time).

I miss the shooting mode dial for choosing different exposure programs, it's now an on screen menu. Several of the controls feel a little delicate.. not as beefy as the gl1 or xl1s.

Low light- With manual exposure settings the same, the gl2 exhibits quite a bit more noise in low light than the gl1, yet substantially more detail and sharpness. By turning the sharpness down, I was able to get this noise to go away, and get a result that was equal to or slightly better than the gl1. When you take advantage of the slower shutter speeds and higher gain the gl2 will give you a lot more room in low light, but you'll have to do a custom preset to deal with the noise.

thats it for now.

barry

Joshua Wachs
July 23rd, 2002, 01:05 PM
great post... as a potential future GL2 owner, your messages are very helpful... keep em coming.

CarterTG
July 23rd, 2002, 01:47 PM
Folks planning their GL2 resolution test might want to grab the EIA test pattern JPEG on JohnBeale's page just for kicks...

http://www.bealecorner.com/trv900/respat/

I'm not advocating a printout test pattern's the most scientific way, but for newbies like myself, it might be preferable to squinting at two still-lifes.

Barry Goyette
July 23rd, 2002, 07:15 PM
Here are a few more preliminary observations on the camera.

I set up a small still life under a skylight and shot it using the gl1, gl2, and xl1s. Comparing the footage from the three cameras I noticed several significant things.

1. the Gl2 is less sensitive than the gl1 and xl1s by a stop or more.

2. Both the highlights and shadows (in an normally lit scene) are handled much more delicately on the gl2 (no chalky whites or crushed blacks...even with the setup at default).

3. In moderate light levels the gl2 shows a very fine grain or noise ( that gets much more pronounced in lower light levels). This grain is much finer than the noise produced by the gl1, but it is sharper. By turning down the sharpness a couple of notches below default, it virtually disappears, and the image itself suffers little if at all at the lowered sharpness setting. This grain is not visible at normal daylight levels.

4. The gl2 appears to render noticeably more detail than either the gl1 or (especially) the xl1s. Essentially it is sharper without looking sharpened.

5. Color saturation of the gl2 is about the same as the gl1, but substantially less than the xl1s.

All observations were made (except where noted) with cameras on manual, and camera settings at default.

I really like the image quality of the camera so far.

Barry Goyette

Frank Granovski
July 23rd, 2002, 11:50 PM
Thanks, Barry. I sure would like to know its horizontal lines on playback. I don't know why cam makers don't post this. The double talk on Canon's site goes something like this: and NTSC is 525 lines. (Yeah but what about the cam's playback resolution.) Is it 480 lines? 490? 500? I like to know a little more about the cam before I buy it.

Sony advertised the TRV900 as having 500 lines. A resolution test revealed it only had 470 lines. The MX300 was advertised with having, well, here it is: "It provides superb horizontal resolution equivalent to 1,530 lines of a broadcast camera." A Panasonic test revealed 500 lines. XL1? 460 lines. VX2000, 500 lines. Optura PI, 360 lines. The TRV20, 485 lines....

Barry Goyette
July 24th, 2002, 07:31 AM
Frank

My earlier posts are essentially trying to relay the quality of the gl2 as it is relative to earlier canon machines that I own.

I'm probably not equipped to do the test you are looking for. I don't know what a waveform monitor is. I will try to shoot a test pattern today and see if that gives me anything.


Barry

Jeff Donald
July 24th, 2002, 08:27 AM
The 525 lines, in the NTSC spec, is vertical resolution and is fixed by the NTSC spec. Resolution charts and waveform monitors measure horizontal resolution. Horizontal resolution is affected by the lenses, chips, electrical components, tape etc.

Jeff

Don Palomaki
July 24th, 2002, 03:08 PM
NTSC is 525 scan nines, about 480 of which contain image, the rest are vertical blanking/retrace interval. Less than 480 scan lines are seen on the screen of a typical TV thanks to overscan.

Most video resolution specs are horizontal resolution, and this is usually the "limiting resolution" at that. The point where the ability to destinguish between lines and spaces blurrs to gray.

Actual horizontal resolution is determined by a variety of factors in the signal path from the target to the TV screen. These include the quality of the lens and focus, the CCD (e.g., number of pixels), how the image off the CCD is processed in the camera/camcorder, the DV signal specification (nominal 720 pixel line including overscan), the D/A converter on the playback side, and the electronics and CRT in the monitor/TV used for viewing.

There is a lot more to image quality than just a resolution number. Consider how good commercial rental VHS movies look on the tube, and how good NTSC broadcast signal looks, and they are limited by bandwidth to about 250 and 330 line respectively. But resolution claims can make interesting ad copy.

Frank Granovski
July 24th, 2002, 03:33 PM
All I'm interested in is what the playback horizontal resolution is (done with a waveform monitor). Maybe DV Mag will do such a test, and publish the findings. Is it so difficult for Canon to come clean and publish the GL2's complete specs?

Barry Goyette
July 24th, 2002, 04:13 PM
Frank

I am no expert in this, so take it with a grain of salt.

I shot a test pattern with the gl2, gl1 and xl1s today. I think there is some subjectivity when evaluating what I saw, as John Beales site seems to indicate there is more than one standard for reading resolution charts. I think the comparative numbers between the 3 cameras are more relevant, based on my method.

gl2..... 500-520 lines
xl1s.... 480-500lines
Gl1..... 450-470 lines

This is interesting because in a real life situations the xl1s looks quite a bit softer than both the gl1 and 2...but it does seem to resolve High contrast lines better than the gl1.

My tests also confirmed that the overall contrast is noticeably lower on the gl2. It also has less evidence of electronic sharpening (the radius of the sharpening is the same, but the amount is less) than the other two cameras and a higher level of noise.

Hope this is of some use.

Barry

Frank Granovski
July 24th, 2002, 04:36 PM
The lines I posted on the TRV900, VX2000, GL1 and XL1 were from DV Magazine. The MX300 res. was done by one of Pana's top tech guys. The waveform monitor does not lie nor stretch the truth regarding reading horizontal lines. Just the facts.

Don Palomaki
July 24th, 2002, 05:54 PM
Its not the limiting rsolution number, it is the frequency response curve that really counts.