ISO rating, resolution for HD100

B

Barry_Green

Moderator
Looks like 250-320 over most of the exposure curve. Measured against a gray surface, and the waveform showed a uniform 55 IRE at each stop. I'd increase or decrease the light level until the camera responded with 55 IRE at each f-stop.

In the following table, if there's enough light that you're getting an autoiris reading of "STOP", the camera responds with an ISO sensitivity of "ISO".

STOP--ISO
1.4---100
2.8---200
4.0---250
5.6---250
8.0---320
11.0--320
16.0--400

Be aware, you *never* want to shoot at f/16 -- the resolution goes to pot. Diffraction comes into play and reduces your overall resolution to less than that of a DVX! Diffraction is a no-no, don't go below f/8 for sure.

Observed resolution using an EIA1956 chart:

f/1.4--550 vert, 600 horz
f/4----620 vert, 625 horz
f/8----600 vert, 600 horz
f/11---525 vert, 525 horz
f/16---425 vert, 400 horz


Note: these numbers seem really low to me. Tim or Nate, could you run a similar test and see what you're getting on a res chart? You can download this same chart from here:
http://bealecorner.com/trv900/respat/EIA1956.pdf
 
Barry.

What about the diffraction at 2.8 and 2 ? In this HD 1/3" configuration, the sweetest spot should be around ƒ2.8.
Print three versions of the chart - RED, GREEN and BLUE - and then see what happens. I don't have a colour printer or I would do it myself.

I bet you will find that red significantly loses resolution at ƒ4. Green and Blue should hold through to ƒ8.

I plan to scope the gamma curves tonight.

Tim
 
What's the footage like...?

What's the footage like...?

Hi Barry,

Without being too technical, how do you rate the look and feel of the footage...?
compared to the DVX....and the XL2...? more filmic, sharper.....colors...?

Are you pleased with the Camera......& have you had the split screen problem..?

Cheers

Darren Wharton

www.darrenwharton.demon.co.uk.
 
I'm not prepared to comment on that yet. I haven't done enough real-world stuff to know. Everything I've done so far has been "lab testing".

And yes, I most definitely have the split-screen issue...
 
Wow - much more limited than I would have guessed if the numbers are indeed correct. Guess that's the downfall of working with 1/3" chips. ND filters and light kits are going to be [even more of] a must. I wonder how the HVX will fare in this regard.
 
For speed, it actually matches up to be just over 1 stop slower than the DVX, which is about what one would expect from the published specs (f8 @ 2000 Lux for the JVC, f11 @ 2000 Lux for the DVX). I didn't try a side-by-side to take noise into account though, so I don't know how comparatively noisy each one is.
 
THANK YOU for posting that link! I've been going crazy over here because I've got an HD100 and I've shot that exact chart and I get, at a maximum, 600 lines. I've thought there *has* to be something wrong with this camera. Now I see someone else shooting the exact same chart and delivering a full 800 lines, and that answers my question!
 
dashwood said:
Did you have the detail level turned down? What do you get when it is cranked up to max?
Click to expand...
Now that I've seen that my results were abnormal, I'm going back in. I also don't know what focal length it was shot at, but that site demonstrates that the HD100 lens is atrocious at longer focal lengths, so I'm going to set it to exactly what they did (35mm, f/1.4) and see if I can get comparable results.

So far, on my HDTV, the results have been most unimpressive. I mean, it's sharper than SD, but it's not "wow" HDTV. It doesn't have "Prison Break" sharpness, that's for darn sure. And that's been verified by the chart (600x600 ain't HD in my book). So it's possible that I have a coke-bottle lens, or that I just don't know what I'm doing... but now that I have a target to aim for, I'm hoping to figure it out and hit the mark.
 
Mediacre said:
800 lines horizontal? I thought the HD100 was real 1280x720?
Click to expand...
Like we've been trying to say since it first was announced -- pixels is one thing, glass is something entirely different, compression is something else as well. A 1/3" chip is something else too. Anyone who's trying to predict overall image sharpness by counting pixels on a chip, well, they're starting from a faulty premise in the first place.

Furthermore, you have to factor in that this is a 4:3 chart being used to test a 16:9 image; not exactly the right tool for the job. However, it's perfectly valid to use as a comparison as long as all the cameras are playing by the same rules (all are 16:9, all are shooting the same chart, all are framed the same, etc). So as an overall indicator of true measured resolution, no it's not the right tool. As a way to compare these two cameras, sure it's totally appropriate.

Maybe with a decent lens on it, it would deliver higher resolution. But in my opinion (as an HD100 owner) the lens that comes bundled with the HD100 is not, by any definition of the word, decent. It breathes like a fat guy running up a flight of stairs, it blurs the edges, and it exhibits extreme chromatic aberration. And that French site thoroughly confirms it. Although, apparently if you stay well short of max telephoto, it performs significantly better. Near the center it is reasonably sharp.

From my perspective, the lens is the weakest part of the package. I'm experimenting with getting a 16mm cine replacement hack-jobbed onto it to just see whether it's worth going the next step and getting an adapter professionally engineered. With a proper lens it may begin to excel as a camera.

Graeme said it from the beginning, the HD100 resolves the same amount of detail as the Z1, and now we've got charts to show that yes, while the Z1 may be recording twice as many pixels, it's not really recording any more actual observable detail.
 
Oooh -- interesting idea -- I think I'll take the full 3,000 x 2,500 image of that res chart and run that through HDV compression and see what it comes out as -- that should give us an idea of what the theoretical maximum of the format is capable of, and a basis to compare the actual against the theoretical...
 
Okay, here's the chart and the best that 1280 HDV can do with it. Keep in mind this is NOT coming from the camera; this is taking the original image as downloaded from BealeCorner.com, scaling it to 1280x720, and compressing it through 720p HDV.

HDV-720P-Max-Res.jpg
 
No, no camera was used. I just put the full JPG on the timeline and output it as an HDV-compliant .m2t MPEG-2 transport stream from Sony Vegas 6.
 
Barry_Green said:
From my perspective, the lens is the weakest part of the package. I'm experimenting with getting a 16mm cine replacement hack-jobbed onto it to just see whether it's worth going the next step and getting an adapter professionally engineered.
Click to expand...

The stills Dashwood uploaded shot with his Switar 16mm lens looked really sharp and nicely saturaded.
 
Switars are great glass. I've got a few here... but as it stands, they could only be used for ultra-macro, without an optical element to adapt them to the camera.

I haven't tried an Arri Bayonet mount lens, but for grins I put my M42 still-camera lens up to it, and yes it could resolve an image, so I expect we'll see still-camera mounts coming soon...
 
You must log in or register to reply here.
Back
Top