The end of 35mm adapers and hv20s....

zcream said:
The chart does not look right.. Even the HV20/HV30 with its plastic lens have more lens.. 600+ lines is pretty standard among the Canon, Sony, Panasonic low-end HD camcorders..

Is it the compression ? I find it hard to believe that a DSLR sensor + real glass is worse resolution than a 1/3" sensor + plastic lens..
Click to expand...

unless the expert tester shot very underexposed causing high iso, at f 1.8 with who knows what lens

get the point? not saying anything about test conditions is as useful as not doing the test at all.
 
http://www.camcorderinfo.com/content/Canon-HV20-Camcorder-Review/Performance.htm#

>>>
Video Resolution (18.75)
To test the resolution of the Canon HV20’s video, we shoot a DSC Labs video resolution CamAlign chart under bright, even light. Resolution is measured in line widths per picture height. In 1080i, the HV20 showed a vertical resolution of 575 lw/ph and a horizontal resolution of 625 lw/ph. In 24P, the camcorder actually improved the vertical resolution, producing less break-up and artifacting in areas of high-density information, boosting the vertical resolution up to 600 lw/ph. The horizontal resolution remained unchanged at 625 lw/ph.
>>>
 
you'll wait long time for a 3000 lux test

he didn't even use a tripod....
also you realize that f/2 on tiny sensors isn't the same as f/2 with 35mm lenses as far as resolution is
 
zcream said:
Actually the total sensor size are 3x times - the D90 sensor puts 12MP in 28.4mm diagonal..The Ex-F1 puts 6MP in 9.1mm..
Pixel size on the D90 should be 50% more..
Click to expand...
Your math is off. If the diagonal is 3x as long, the image area / sensor size is 9x as big.
 
Not quite.. Diagonal is 3x but pixels are 2x. So each individual pixel is 50% more...
Let me know if this makes sense..
Diagonal is proportional to area.
Pixel * pixel size are proportional to area.

Edit :
Read Num of pixels * pixel size above
Chris_TC said:
Your math is off. If the diagonal is 3x as long, the image area / sensor size is 9x as big.
Click to expand...
 
I think a large part of the problem lies with the JPEG compression..Even though the people testing dont like the HDMI output, can we try to get a HDMI capture at 1080i/p ?? This way, we can compare the photo, JPEG compression and video sensor side by side..
Sorry to make suggestions, but I would have gone into a store if this was available in Australia..
 
ando said:
comparing to a photo camera do you think is a smart thing to do?
get your $999 video camcorder do some tests. make sure you use a 35mm adaptor and the same lens at the same f-stop (you understand why i hope).

anyway what settings did he use when shooting the chart? iso, lens, f-stop, focal length????
he didn't do this at f/1.8 in the dark at iso 1250 with his father's 1975 film nikkor lens right?
Click to expand...

I know that conditions when Joe shot his chart may have been bad. Since it was handheld resolution may have been degraded. I haven't stated anything as facts. I'm just trying to get some scientific data from the camera.

The Nikon D40 which I compared to got a 3008x2000 sensor but can be made to output a 1504x1000 picture. That means it probably skips/bins/scales the sensor by 2x2 which is what the D90 probably does as well, that will make odds more even, don't they?
I then added black bars and resampled it in a computer to get a 720 pixel height.

The lens that I used is in fact from the seventies or sixties and I got it from my dad! =). It's a Minolta 58mm f/1.4 and I used it at f/4 when I took the photo.

Here is a resolution chart from a HV20 with a homemade 35 mm adapter:
http://img169.imageshack.us/my.php?image=twoneilstatichv20achr1cw0.jpg
I found it in this post:
http://hv20.com/showpost.php?p=124172&postcount=5

Resolution on that adapter/camera combination is better. Conditions differ from ours of course.
 
Last edited:
So the resolution on a decent video on the D90 would be about 700-800 lines I'd reckon..With a decent glass..
Then the 720p could be easily upscaled to 1080p..
I'm still thinking that the 1080i on the HDMI is 1080p/24p and it need pulldown to extract the 24p signal. And I'm still hoping that its native and not just uprezzed from 720p..
 
HDMI uncompressed capture

HDMI uncompressed capture

Is anyone still up for HDMI uncompressed capture ?
The 1080i may be the ace in the hole yet..And if we bypass in-camera compression..

Toenis said:
Yes, it will skip the compression although it will compress the signal to Cineform I assume. But still usable only then if D90 could output a clear image, without metadata, via HDMI. If not the clean 1080i image can be achieved via firmware update at least as the HDMI output supports ZOOM in etc at full 1080i rez. Check the D300 specs or samples about HDMI functionality, I assume it will be close to what D90 will have.


T
Click to expand...
 
bearing said:
Resolution on that adapter/camera combination isn't that bad, is it?
Click to expand...


yes bearing .it looks good .

but your photo resize can't be compared. you don't know what sort of downsampling is inside d90. you also shouldn't think that d90 is getting 24fps of 12mp then scaling it 1mp. i don't think its cpu is fast enough for that, it's maximum speed is 4.5 fps raw. you have to wait for video on a more expensive camera, like one with 8 fps full res, to have binning and that will be very high res indeed. d90 must be skipping some horizontal lines or something like that. good thing is that such behaviour leaves room for improvement with firmware.

the test was so bad it proves nothing.
very good light, a hi-res f-stop like f/5.6-8 on a sharp lens, low iso (200 should get max IQ on d90), focusing by the math no by the eye and a tripod is the minimum for a test.

looking at the full res samples i think there will be $1000-1200 camcorders with finer 720p resolution. but no way any $1k camcorder will shoot at high iso, with 35mm dof, lenses and a 640x480 back lcd to check your focus on. so each person getting d90 just for video should balance that. on the photo side d90 will be a sure hit now that we've seen all the cameras released for this price range in 2008.
 
Last edited:
All this talk of resolution and technical specs is fine and all. But in the end only the result matters. Can you make a film with a D90 - and one that has a great look? I'm sure of it. But it'll take some work, I think we can all see that by now.
 
no post for an hour

no post for an hour

Have all the vampires gone to sleep ?

ando said:
yes bearing .it looks good .

but your photo resize can't be compared. you don't know what sort of downsampling is inside d90. you also shouldn't think that d90 is getting 24fps of 12mp then scaling it 1mp. i don't think its cpu is fast enough for that, it's maximum speed is 4.5 fps raw. you have to wait for video on a more expensive camera, like one with 8 fps full res, to have binning and that will be very high res indeed. d90 must be skipping some horizontal lines or something like that. good thing is that such behaviour leaves room for improvement with firmware.

the test was so bad it proves nothing.
very good light, a hi-res f-stop like f/5.6-8 on a sharp lens, low iso (200 should get max IQ on d90) and a tripod is the minimum for a test.

looking at the full res samples i think there will be $1000-1200 camcorders with finer 720p resolution. but no way any $1k camcorder will shoot at high iso, with 35mm dof, lenses and a 640x480 back lcd to check your focus on. so each person getting d90 just for video should balance that. on the photo side d90 will be a sure hit now that we've seen all the cameras released for this price range in 2008.
Click to expand...
 
You must log in or register to reply here.
Back
Top