Hi everyone,
I own an EX1 since a couple of weeks and I've been testing it for a few days. It is a great camera, offering a large range of possibilities for sure.

But as a matter of fact I've never been able to produce clean images. There's always some
kind of noise in the frame, sometimes even in well lightened situations on clear (in mean no shadows) surfaces.

I'm in manual everything, AGC is off, no extra gain (I use to shoot -3 or 0db, maximum I use is 3db). Whatever picture profile I set, it's always looks like something's going wrong.
I've been testing without PP, Bill raven's PP, Alistair Chapman's PP, and a lot of other ones, I've read topics about gammas and noise in the shadows on this site, but so far nothing helped me to get rid of this strange noisy picture. When i look at some footage I've made it just doesn't look right compare to others EX1 footage on the internet.

Unfortunately, I have'nt a second EX1 to compare side by side the footage they produce.
But before thinking it may be a deficient sensor unit I would like to test it properly so that's the point of this thread : Have you any ideas of tests I could do, then post for you to see, in order to make sure that the issue came from my use of the camcorder and not from the camcorder itself.

Have you tried turning off the Detail yet?

how are you watching the footage to see the noise?
on PC? on HD TV? etc

and what connection method are you using?

Detail is usually turned off (or -8 when I use Alistair Chapman's PP).

Unfortunately I can only view my footage on my PC monitor Samsung Syncmaster T220 (22" wide) wich has a resolution of 1650x1280 at max, so for now I only shot in 720p for the screen resolution is not enough to watch 1920x1080 full screen without rescaling.

I transfer my footage via Clip browser software 2.0. I create one instance of the file by drag and drop (so it turn to be a mp4 file) and i export an other one as an mxf (for editing purpose).
I cannot see a difference between the two file types when played on VLC.

Is there anyone out here who can think to a simple kind of test to actually verify if the lens is OK or if it does have a sensor problem?

I will post result footage.

I thought to shot a few seconds with iris fully open, no PP, and lens hood closed to see the amount of noise on a total black screen?

There is always some noise in the EX's pictures. HD cameras do tend to be noisier than SD cameras. Computer monitors are not the best way to view footage and can make noise look worse than it perhaps really is. I would try to find a way of looking at your pictures using a proper HD monitor. Have you looked at the SD output on a decent monitor, does this look just as noisy?

I've actually read on this forum -3db decreases the range on the camera... In fact, shooting with standard settings - not cine - at 0db gives you the fullest range much to my surprise.
I'd recommend purifier by innobits to remove most of your noise as it has saved my @#$ shooting high contrast beaches in hawaii...

Sorry Greg,
but what do you mean by "purifier by innobits"?

Google is your friend:

innobits.com - It is all about visions (http://www.innobits.com/purifier.html)

Why would you get noise shooting beaches in Hawaii? I would think there would be plenty of light in that situation. But then I've shot very little bright outdoor scenes with this camera (it's been pretty gloomy since January when we bought it) :)

Why would you get noise shooting beaches in Hawaii? I would think there would be plenty of light in that situation.

The EX cameras noise is peculiar in that it's not strictly low-light related. Even in full sunlight, shadow areas tend to show some noise. Increasing exposure could be the cure (but of course the highlights would then be blown away); another is crushing the blacks (but then the noise is suppressed along with details in those areas).

All in all, I guess we have to live with that, and learn to use optimum settings for each individual scene / lighting conditions.

I've used quite a few cameras in the EX price range, the EX1 fairs among the best in noise levels.

Also, of course we all know the settings and picture profiles can help to find the best image under a given situation.

Thanks Greg and Mitchell,
Google is my friend you're absolutely right!
I guess we were not talking about Tolkien's mythology "in(h)obbits" but about an existing plugin to denoise footage... Ok I get it!

Anyway, I've uploaded a piece of my EX footage to get some more feedback about this "issue".
Clouds timelapse - EX1 Noise test on Vimeo (http://www.vimeo.com/3897188)
Note that the online version is highly compresssed, so download the original file which is straight from the SxS card via clip browser software, no correction or edit.

Google is your friend:

innobits.com - It is all about visions (http://www.innobits.com/purifier.html)

Why would you get noise shooting beaches in Hawaii? I would think there would be plenty of light in that situation. But then I've shot very little bright outdoor scenes with this camera (it's been pretty gloomy since January when we bought it) :)
So I clicked on the link and I'm really surprised that they would use the Beijing example to show off the noise reduction. In that picture, it looks like any other noise reduction in that the detail is softened tremendously and the black levels are so clipped on the processed version that even more detail is missing. Notice the bushes in the background to the left of the girl and the petals by her elbow. In this example I would prefer to have the grainier picture to the one that has that too "glassy" kind of look. I'm sure the innobits purifier is a fine noise reducer, but if you use it, definitely dial it down a bit.

Going by the specs you listed on the Vimeo site, I'd turn off SLS immediately. This will enable you to quit using f16 and you'll regain your image sharpness. I'm curious why you were using SLS in the first place? This additional exposure time might be causing some of the noise buildup, but I'm just guessing on that one.

At an earlier point in the thread you say that you are shooting 720, may I suggest shooting at the cameras native 1920 setting, and avoid in camera scaling - for me I find 720 images horribly muddy and soft as compared to the full raster output of the EX1.

Viewing on a computer monitor is also no way to investigate problems with footage from this camera, at least look at the image on a broadcast specification monitor via HDSDI or at least, component analogue.

You say that you are trying to avoid scaling on your Samsung, give the camera a chance and let it shoot what it designed for - 1920x1080.

Oh, yes, never use anything higher that F8, your image will get more soggy the higher the F stop, try for the optimum F5.6.

just my 2€

Paul

SLS will increase visible noise. As the sensor is "on" for longer any noise it is producing will be allowed to accumulate for longer and thus become more visible.

That's an important point to keep in mind,
Thanks Alisatair!

By the way thanks Paul and Dave,
You're right and I know 1080p is cleary more detailed, so I guess i have to wait until I find a true HDTV monitor to test and watch footage.
Thank you guys for all our answers.

SLS will increase visible noise. As the sensor is "on" for longer any noise it is producing will be allowed to accumulate for longer and thus become more visible.

I am not sure this is right - noise is more visible when the ratio to clean signal is low- if the sensor is exposed for longer then the clean signal will be larger - the noise is pretty much a constant on the sensor - hence GAIN increases small clean signal (low light) AND the constant noise which then becomes visible.

So noise is a function of gain more than anything else

I am not sure this is right - noise is more visible when the ratio to clean signal is low- if the sensor is exposed for longer then the clean signal will be larger - the noise is pretty much a constant on the sensor - hence GAIN increases small clean signal (low light) AND the constant noise which then becomes visible.

So noise is a function of gain more than anything else

I'd have to disagree with you on this. Anybody who has ever shot a time exposure, for example, with a digital camera will start seeing noise and glowing "stuck" pixels as the length of time the sensor is exposed to light. So, my contention that SLS was one of the sources of his noise problems seems to make sense.....to me, anyway.

I am not sure this is right - noise is more visible when the ratio to clean signal is low- if the sensor is exposed for longer then the clean signal will be larger - the noise is pretty much a constant on the sensor - hence GAIN increases small clean signal (low light) AND the constant noise which then becomes visible.

So noise is a function of gain more than anything else

I think Alister is right. Assuming gain is unchanged, the same amount of light is needed to give a correctly exposed image whether normal or slow shutter is used. The difference is that with the longer accumulation time, the 'rate' at which that light needs to arrive is lower. Hence the greater sensitivity and, all other things being equal, the need to stop down as the SLS frames are increased. So the signal part of S/N is constant.

Assuming that noise accumulates over time, the longer the accumulation period, the more noise will be present. Constant signal and, with SLS, increasing noise suggests that the S/N will decrease.