Jump to content

I'm trying to figure out the cause of this CCD imager issue.


Mike Krumlauf

Recommended Posts

Hello fellow cinematographers,

 

I have a very strange situation going on that has made itself seen between 3 different models of the Sony HDW-F900 camera. The first 900 was the original model with first version software, the second one I had was the slash 3 (/3) version, and the final one i am investing my time and money is the (once) beloved Panavision HD900F. I am getting small clusters in what seems to be bright high contrast areas of vertical line smearing. I have used only 2 lenses between the three cameras which was a POS Fujinon SD Lens and an OPTEX modified Canon HJ9x5.5 HD Wide Angle. I have spoken to severial camera engineers on the phone all over the country and none of them can pint point what this is, as if they almost have never seen it before which, given this camera's age and lifespan i find hard to believe. The Panavision model has shown it the least and its so minimal i can actually control it and it does not affect my image but in the back of my mind im thinking.. "this camera was well over 100 grand at one point, it should not be doing this". I have no idea if it is a CCD problem or a lens and CCD alignment problem as i never properly had the lenses on any of the 900s properly white shaded with a lens file made, etc. I've searched up and down google for weeks with not one close example of my problem. Some people over at DVXuser mentioned it might be "CCD Clock Noise" which is gibberish to me as im not an engineer trained in these chipsets. I have included some examples of the problem. If there is something i can do menu setting wise or with an optical filter in a mattebox to get rid of this please let me know.

 

Also, I am not shooting HDCAM tape on these cameras. I am using a Miranda DVC800 box and recording via SDI to a AJA KiPro Mini. The problem shows up on the monitor component out as well.

 

If i could just figure out what it is 110%, i could better understand how its happened. It just blows my mind that this problem has carried itself over 3 cameras.. does not make sense...

 

I posted a full frame shot of the image, then two cropped pictures showing the areas. Pay attention to the street lights and the areas right next to where the lights end. I am happy to post more examples if needed.

 

Thanks,

 

Mike :)

post-74343-0-70030200-1536807150_thumb.png

post-74343-0-29403500-1536807292_thumb.jpg

post-74343-0-95428800-1536807344_thumb.jpg

Link to comment
Share on other sites

Yeah, it's chromatic aberration. It's a lens issue, not the sensor. Most lenses, except the very expensive ones, exhibit it to some some degree. As with all lens aberrations, it will show up worst when wide open.

 

You could:

 

A. Use better lenses

B. Stop down

C. Fix it n post

  • Upvote 1
Link to comment
Share on other sites

  • Premium Member

I am not an expert by any means, but have you systematically investigated this issue?

 

Try resetting the camera to default settings with NO changes on image processing enhancement and at a ISO/ASA that is well within the native range of the sensor (probably around 100).

 

Choose a scene where this issue tends to be exacerbated and, keeping good notes and with slates, shoot short segments as you change settings. Examine the shots and see if you can tell where the noise comes into the image.

 

Also, try it with a HDCAM tape and remove the converter and recorder.

 

Check to see if the outboard converter/recorder could be acting as an antenna for ingress of stray electromagnetic impulses from nearby powerlines and such.

 

Put your cell phone in the car; get it away from the camera. That could be a source of electronic noise.

 

Edit: "this camera was well over 100 grand at one point, it should not be doing this". Just because something cost a lot in the past, doesn't mean it didn't suffer from the problem all along. Check with former owners/shooters; it might not be an unusual artifact, regardless of past price...

Link to comment
Share on other sites

Thanks for the info Frank. I was meaning to post on here last night as I found a solution to the problem. It seems if I shoot on HDCAM tape as intended with the camera and digitize the tape via HD SDI into my KiPro, it not only gets rid of the line problem, but the image looks far less noisy and much cleaner! I guess the 900s just were not meant to have their live component out and HD SDI out be used as a recording feed. So this will be my process from now on. Very happy!

Edited by Mike Krumlauf
Link to comment
Share on other sites

Hi, the image above is a classic example of chromatic aberration, a lens issue that manifest as green-magenta "fringing"on high contrast transitions. Use an appropriate HD lens to correct this issue.

CA has nothing to do with the lens being "HD" (which's a meaningless marketing term since it isn't that only "HD lenses" resolve 110 lp/mm or, say, 40 line pairs at decent contrast) or not. High-end EFP, box lenses and especially D-cinema lenses like Fujinon E series or Canon FJ are better corrected for CA at wide apertures.

Link to comment
Share on other sites

Thanks for the info Frank. I was meaning to post on here last night as I found a solution to the problem. It seems if I shoot on HDCAM tape as intended with the camera and digitize the tape via HD SDI into my KiPro, it not only gets rid of the line problem, but the image looks far less noisy and much cleaner! I guess the 900s just were not meant to have their live component out and HD SDI out be used as a recording feed. So this will be my process from now on. Very happy!

Interesting... Because, if I remember correctly, the parallel port at the back of non-R f900 output a 10-bit signal straight from the DSP, just upstream the downsampling/3:1:1 circuitry. I had the service manual vol.2 somewhere though, should look up.

Link to comment
Share on other sites

Mike,

 

Glad you got it sorted out!

 

Frank

As am i believe me. I say that maybe the F900s are not meant to shoot directly live from an HD SDI via the HDCA 901 or Miranda Box because this issue traveled between 3 cameras and it was showing up in the monitor out BNCs on the side of the camera. More than happy to film with the camera via HDCAM as it was designed to do and digitize later.. the little bit of info in bit information i might be loosing is well worth fixing the problem. 8bit HDCAM 3:1:1 is just fine.

Link to comment
Share on other sites

CA has nothing to do with the lens being "HD" (which's a meaningless marketing term since it isn't that only "HD lenses" resolve 110 lp/mm or, say, 40 line pairs at decent contrast) or not. High-end EFP, box lenses and especially D-cinema lenses like Fujinon E series or Canon FJ are better corrected for CA at wide apertures.

I was going to say as well and i started a new post in the lens section that i know that my lens is "HD Spec". I've heard that the fujinon lenses are the best and sharpest for these cameras.. the glass inside my lens was made by canon and the housing is optex.

Link to comment
Share on other sites

It seems if I shoot on HDCAM tape as intended with the camera and digitize the tape via HD SDI into my KiPro, it not only gets rid of the line problem, but the image looks far less noisy and much cleaner!

That's strange. As far as I remember, the HD-SDI output on the 900's was 10 bit, so it should technically be better than the 8 bit HD-CAM recording

Link to comment
Share on other sites

  • Premium Member

Maybe if you threw that HD-SDI output up on a scope, you might see a problem in signal output. It might have some (now) non standard signal component or show some electronics fault that is giving you the "heebie-jeebies" via external capture.

Link to comment
Share on other sites

Maybe if you threw that HD-SDI output up on a scope, you might see a problem in signal output. It might have some (now) non standard signal component or show some electronics fault that is giving you the "heebie-jeebies" via external capture.

I can connect the Miranda box to my Mac via FireWire and I have a program called scopebox. What chart would I be looking at and how could I tell if there were any abnormally?

Link to comment
Share on other sites

  • Premium Member

Does the camera have an internal color bar generator? You could throw up bars and look at the waveform and compare it to a proper waveform, however I am unsure if Scopebox has the needed variable gain adjustments that allow the waveform to be expanded to see the trace at 10X. The trace should be fairly smooth; any spikes or excessive jittering, especially in the timing signals, could be something out of spec enough to interfere with image processing. Noise spikes big enough to be interpreted as a "1" can throw off sync on sequencing clocks and throw your pixel order out of phase on a line read-out. A raised noise floor can bias the signal enough to trigger false positives on super bright pixels and screw up your image processing.

 

You could check with the local TV station's Chief Engineer and ask him to have a look at the output on a good Tektronix HD scope in exchange for a trip to Burger King! He/She should be able to point out any noise or ringing in the signal, but be sure to try it straight from the camera and THEN through the Miranda Box. You might have a bad component, or even cable, in the imaging chain.

  • Upvote 1
Link to comment
Share on other sites

Does the camera have an internal color bar generator? You could throw up bars and look at the waveform and compare it to a proper waveform, however I am unsure if Scopebox has the needed variable gain adjustments that allow the waveform to be expanded to see the trace at 10X. The trace should be fairly smooth; any spikes or excessive jittering, especially in the timing signals, could be something out of spec enough to interfere with image processing. Noise spikes big enough to be interpreted as a "1" can throw off sync on sequencing clocks and throw your pixel order out of phase on a line read-out. A raised noise floor can bias the signal enough to trigger false positives on super bright pixels and screw up your image processing.

 

You could check with the local TV station's Chief Engineer and ask him to have a look at the output on a good Tektronix HD scope in exchange for a trip to Burger King! He/She should be able to point out any noise or ringing in the signal, but be sure to try it straight from the camera and THEN through the Miranda Box. You might have a bad component, or even cable, in the imaging chain.

I was thinking that maybe the CCD Gain amplifers might be a little high but I dont want to mess with them without knowing what to look for on the waveform. Would I just be looking at the standard waveform or an RGB parade? If i dont have a choma chart, is there something pratical i can point the camera at in my apartment to properly set the noise level in the imager? Would the pre amp board have to be messed with too or leave that alone? Are their specific user controls to readjust the clocks on the CCD or does that require specific tools only a tech's office would have? Again, im not at all getting the vertical line feedback when shooting HDCAM tape, but, i am still a little interested as to if the noise in the image is to high. I know the 900s are older cameras but at -3db and 0db, there should not be any noise in the blacks right?

Link to comment
Share on other sites

  • Premium Member

I wouldn't try to adjust it without a calibrated reference signal.

 

I was suggesting you just look for noise in the basic waveform. You might see it in the timing trace outside of the RGB in parade.

 

There should be minimal noise in the blacks at "normal" gain levels.

 

Frankly, I would suspect your outboard converter before I suspected your camera.

Link to comment
Share on other sites

I wouldn't try to adjust it without a calibrated reference signal.

 

I was suggesting you just look for noise in the basic waveform. You might see it in the timing trace outside of the RGB in parade.

 

There should be minimal noise in the blacks at "normal" gain levels.

 

Frankly, I would suspect your outboard converter before I suspected your camera

Thanks Frank for all your help on this. I'm sure using the panavision version without panavision lenses might be part of my problem, but small at best. I know that the pre optical filter is different from a normal 900 and the CCD is calibrated a little different to accommodate their superior primo lenses that they made for it back in the day. I know of one place here in Denver called CinemaShot which could take a look at the camera for a small price... I recently moved here from Chicago so my options compared to my hometown are limited in camera support.

Link to comment
Share on other sites

There seem to be a lot of dots for a test color bars signal. Can you post an image of the bars themselves?

 

It's also possible that it's a cable issue. Or it's your Ki-Pro. Try another recorder, or record directly to a capture card in your computer.

Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
×
×
  • Create New...