Using Focus Peaking to Assess Your Camera Noise

Testing Your Camera

When I enabled the external HDMI monitor’s focus peaking (Figure 5, below), it showed every little bit of contrast in an out-of-focus image. The only thing that could have sharp contrast in an out-of-focus image is the noise in the image itself. The focus peaking in the external monitor is so sensitive that it actually creates problems when trying to use it for critical focus in a shot. The focus peaking in my camera is much better.

Figure 5. Focus peaking is enabled manually on this monitor.

Either way, it wasn’t until I connected an external monitor to my camera that I was able to see even the camera’s own focus peaking highlighting the camera noise. The monitor I’m using here is only 1280x800 and not a FullHD monitor, but it is the focus peaking that makes the difference, so even lower-resolution monitors with focus peaking may work for this process.

In the monitor I enabled focus peaking and then I de-focused the camera by pointing at a darkish area (Figure 6, below). If the area you’re showing is bright, you may not be able to see the focus peaking as well. Also, if your focus peaking is colorized (red, blue, whatever), pick a different color background.

Figure 6. Focus peaking dots indicate noise in the video.

To make sure the camera image would not get too bright, I set the exposure to Program and the auto-compensation to −3 stops. This would help keep the image darker than the focus peaking highlights. The camera would adjust shutter and aperture to compensate as I adjusted ISO manually. The camera is in manual movie mode, not any of the still photography modes.

Starting at ISO 200, I cycled up through 400, 800, 1600, 3200, and 6400. Then, in a sub-setting in the camera, I enabled 1/3-stop increments. So my ISOs became 200, 250, 320, 400, 500, 640, 800, 1000, 1250, 1600, 2000, 2500, 3200, 4000, 5000, and 6400. While my camera offers ISOs higher and lower than that range for still photographs, they are not available in movie mode.

When I started testing I found that ISO 400 was better than 200. This was surprising to me. The lower ISO was not better. Then as I increased the ISO, I found that ISO 1600 was considerably cleaner than ISO 800. This said to me that ISO 1600 was undergoing some additional in-camera noise reduction. Testing higher ISOs revealed that 3200 and above were completely unusable because of the noise. 6400 is a complete mess (Figure 7, below).

Figure 7. ISO 6400 is a mess.

Adjusting the in-camera picture profile noise reduction settings showed that bringing the noise reduction up from −5 (so I could more easily see the noise) to 0 (the standard setting) dramatically lowered the noise at almost all ISOs--specifically the lower ones (Figure 8, below). However, ISO 800 was still markedly worse than 400, and 1600 is actually cleaner than ISO 800. So I learned more about what to stay away from, in addition to what works best.

Figure 8. Predictably, increasing noise reduction lowers noise at almost all ISOs. 

Then, when I set the camera to use 1/3-stops so I could more finely tune the in-camera ISO settings, I found that it wasn't ISO-400 that was the sweet spot. It was the slightly lower ISO 320 that was cleaner than anything else. Moreover, it wasn’t ISO 1600 that was the “high” ISO sweet spot, but ISO 1250 that is the cleanest high ISO setting. So by keeping with those two ISOs, 320, and 1250, I can ensure the cleanest image and use my aperture and shutter or an external ND filter to control the light hitting the sensor.

Topics for Further Testing

Of course, I am by no means done testing. This camera, like many cameras made today, is chock-full of image parameters (Figure 9, below). I can vary in-camera curves (such as a5 for highlight and shadow separately). Master pedestal a 15, luminance levels have three different settings, and there are eight different picture profiles that have five parameters each, and each of those parameters have 10 a settings and zero. That’s a lot of adjustability.

Figure 9. Other image parameters on the GH4.

Is it worth it to test every possible parameter of your camera? Probably not. But start with the settings you normally use, and begin to vary those settings around that starting point. You might be surprised to find that a group of settings, or small adjustments, to what you normally use can deliver markedly improved results.

For me, leaving the 1/3-stop setting enabled, and keeping to ISO 320 and 1250 is not that difficult at all. I no longer have to try and get more light and try to try to stay at ISO 200 because I think that’s where my lowest noise is. I now know where my camera performs best.

So knowing your camera, and testing its performance, is key to a better understanding of how you can deliver the highest quality images possible with the gear you have. Often, the solution isn’t another gadget or gizmo, no matter how tempting that may be. The only cost is some time taken to really dive in to the capabilities, and assess the problem areas, of the gear you already have. Then you know best what to do, and what not to do.