Welcome to our community

Be a part of something great, join today!

Register Log in
  • Hey all, just changed over the backend after 15 years I figured time to give it a bit of an update, its probably gonna be a bit weird for most of you and i am sure there is a few bugs to work out but it should kinda work the same as before... hopefully :)

High ISO - image looks clean, playback looks noisy

V

Viktor Sloth

Well-known member
Joined
Apr 12, 2012
Messages
69
Reaction score
0
Points
0
Location
Denmark
Website
duende.works
Hi guys,

Just a quick question.
I was doing some high ISO shots as a test - 6400, and I was so surprised over the image I saw on the monitor. It looked clean, crisp and wonderful.
I pushed record and went into the playback. In the playback the footage had loads of noise and didn't look anything like the noise I saw while shooting.
So, my question is: The applied noise, is that coming from the compression or something completely else? I assume the compression get's more and more important, the less light there is, but I have no clue frankly.

Thanks in advance

:)
 
That is irrelevant, because I see an acceptable image on the monitor, press record, and noise is applied. I just want to find the reason for that.

The ISO is off the charts for a perfume commercial, I agree - but works fine for this gritty project i'm working on these days. Thanks for the input though.
 
Are you talking about playback in camera, or on your computer? The playback in camera is not full quality either. Noise should be judged at fully quality and at 100 percent on your desktop IMHO.
 
I am actually not sure how much of a difference it makes but, on what were you monitoring the playback?
 
Viktor Sloth said:
Hi guys,

Just a quick question.
I was doing some high ISO shots as a test - 6400, and I was so surprised over the image I saw on the monitor. It looked clean, crisp and wonderful.
I pushed record and went into the playback. In the playback the footage had loads of noise and didn't look anything like the noise I saw while shooting.
So, my question is: The applied noise, is that coming from the compression or something completely else? I assume the compression get's more and more important, the less light there is, but I have no clue frankly.

Thanks in advance

:)
Click to expand...

If you are monitoring playback from the camera then you are looking at a half debayer of the image.

If you are judging ISO test footage for noise, I would recommend judging from delivery monitors rather then production monitors. If it looks good on a grading monitor, and it looks good watching it on YouTube on your iPhone, then it's probably good.

Feel free to share your ISO tests with the community. I think everyone would be interested to see a useable 6400 ISO!
 
Viktor Sloth said:
Hi guys,

Just a quick question.
I was doing some high ISO shots as a test - 6400, and I was so surprised over the image I saw on the monitor. It looked clean, crisp and wonderful.
I pushed record and went into the playback. In the playback the footage had loads of noise and didn't look anything like the noise I saw while shooting.
So, my question is: The applied noise, is that coming from the compression or something completely else? I assume the compression get's more and more important, the less light there is, but I have no clue frankly.

Thanks in advance

:)
Click to expand...

The answer is YES, IMHO compression plays a huge factor.

What you are looking at ON SCREEN is, as far as I know, a clean feed of what sensor is SEEING...not what it's recording.

So if it looks good at 6400 on SCREEN...but then you record 10:1....what will be captured will be a very compressed version of what you saw on screen.

And in lowlight, even a bit of noise presumably can get very exacerbated by the affect of high compression (ie the high compression then "feeds" on the small amount of noise/gain in the image, making that noise look much worse).

Does this make sense?

This is why the Carbon is such a big deal. People will be able to shoot 3:1 at 6K (ie LOSSLESS) and 2:1 at 5K (ie basically uncompressed). Which are INSANE compresions for ever day use, but for occasional moments...like a quick LOWLIGHT SCENE at 6400....are very handy to have.

This is also why internal PRORES may become so important - as this (from what I'v heard) apparently SKIPS the compression all together. It won't be RAW, but it my be VERY CLEAN.

Does this help?
best
 
As an aside regarding noise, in another thread recently, Graeme Nattress (the wizard of REDcolor), noted the idea of shooting with frame averaging to reduce noise. He suggested for motion, using a setting of averaging 4 frames. I haven't tried it yet but it sounds promising.
 
You've got to try it Terry. I did a test here with my daughter dancing around a dark room (so specifically looking for naughty motion artifacts) and it looked pretty darn good. I've been using 16 frame averaging for stills shooting and it produces the best looking digital images I've ever seen. (But I would say that, wouldn't I??) :-)

Graeme
 
Howdy Graeme,

Can you give me a little more info on frame averaging? I tried it out the other day on a test and I couldn't figure out what was going on. I set frame averaging to 4 frames, and rolled. Looked weird. Play back looked sped up, and I think my frame rate jumped to 96. Is this somthing that needs to be treated in post? I'll mess with it more, but a quick explanation of how to set it up would be helpful.

Cheers,

Nick
 
Graeme Nattress said:
You've got to try it Terry. I did a test here with my daughter dancing around a dark room (so specifically looking for naughty motion artifacts) and it looked pretty darn good. I've been using 16 frame averaging for stills shooting and it produces the best looking digital images I've ever seen. (But I would say that, wouldn't I??) :-)

Graeme
Click to expand...

You're right, Graeme, I really need to give this a go. I LOVE learning little tidbits like this! Thanks.
 
Graeme Nattress said:
You've got to try it Terry. I did a test here with my daughter dancing around a dark room (so specifically looking for naughty motion artifacts) and it looked pretty darn good. I've been using 16 frame averaging for stills shooting and it produces the best looking digital images I've ever seen. (But I would say that, wouldn't I??) :-)

Graeme
Click to expand...

Yea Graeme...any additional info on it would be greatly appreciated! When to use, what happens in post, etc.
 
Nick Gardner said:
Howdy Graeme,

Can you give me a little more info on frame averaging? I tried it out the other day on a test and I couldn't figure out what was going on. I set frame averaging to 4 frames, and rolled. Looked weird. Play back looked sped up, and I think my frame rate jumped to 96. Is this somthing that needs to be treated in post? I'll mess with it more, but a quick explanation of how to set it up would be helpful.

Cheers,

Nick
Click to expand...

http://www.reduser.net/forum/showth...ght-RED-mini&p=1568585&viewfull=1#post1568585
 
Keep in mind you'll want a 360 degree shutter, which at 4 frames average and 96fps will make 24fps with a 360 shutter. if you have less than 360 degree you can get a kinda strobbing in your motion blur. You can put the iso pretty high doing this though.
 
I tested at 180 shutter and didn't see any strobing. Indeed, what I did see was a smoother motion than I'd normally expect for 180 shutter. Try it and see. There's nothing special to do in post.

Graeme
 
Thanks everyone for the response.
Everything make perfect sense. Nice to get it clarified how much noise the compression makes in a low light situation with high ISO.


Frame averaging sounds very interesting!
Nice.

Not entirely sure I understand the technical aspects of it?
So let's say I shoot with 2 averaging frames and want to make a 25p clip. I then have to get the framerate up to 50, and then some magic happens and it's less noisy and better quality?

For stills it indeed sounds interesting as well.
 
When you turn on Frame Averaging, the fps indicator will now be showing you how many recorded frames per second you'll be making. Set that to 25 and you're good to go. Internally the sensor will be running at 50fps, averaging 2 frames into one and recording 25 every second. For stills you can go to much higher averaging and produce some of the lower noise images you'll have ever seen on a digital camera.

Graeme
 
It works a bit like HDR in terms of processing power. You have to be capturing 2x the frames (as Graeme is describing) in order to create a motion strea. Ie 48 frames per sec to get 24 or 50 frames per second to get 25. You can not record sound in this mode which is no big deal, you really see the benefits when you timelpase or for stills extraction as Graeme has stated.

This is is tempting me to do some more tests.
 
Graeme, sorry if I was mistaken about the strobbing.

I was imagining that at 50 fps, 180 shutter and two frame averaging, you'd have .01 seconds exposed, then .01 seconds with the 'shutter' closed, then .01 seconds exposed, then closed again. Wouldn't the average have gaps temporally that would affect the motion blur?


Edit:
Also, and I should test, but does 50fps averaged into 25fps at 800 iso have less noise than 25fps at 400 iso? I imagine that 50fps 180 shutter is one stop darker than 25fps 180 shutter so you'd need to double the ISO.

On the other hand, it's a really neat tool for changing exposure in the camera besides just ISO and shutter.
 
Last edited:
You must log in or register to reply here.
Back
Top