Thursday, 5 June 2014

GH4 shutters

Recently, an interesting Panasonic sensor data sheet popped up. It is believed to describe the sensor sitting in, e.g., the Lumix GX7 and GM1.

The data sheet says that there are two options for electronic shutter readout: 12 bits in 1/15s and 10 bits in 1/22.5s. It has been demonstrated that the Lumix GX7 has a readout speed of 1/15s, and I have measured the speed of the electronic shutter in the GM1 to around 1/25s.

I also saw that the GM1 had slightly less details in the shadows at ISO 200 when using the electronic shutter. This is because it only uses a 10 bit readout with the electronic shutter.

Worse image quality with the E-shutter?

So, does the same apply to the Lumix GH4?

To test this, I have taken the same image with the Lumix GH4 and GH3 at ISO 200, with and without the electronic shutter. I underexposed the images two stops, to exaggerate the shadow noise problems. The exposure parameters were: f/3.5, 1/4s, ISO 200:

Mechanical shutter
Electronic shutter
Lumix GH3

Lumix GH4

You can download the GH4 RAW files. Mechanical shutter Electronic shutter

Looking at the JPEG output images above, there is little difference to find. However, when I process the RAW files in Silkypix, increasing the brightness by three stops, I find some difference in the shadows:

What we see here is that the Lumix GH3 image quality is pretty much comparable when using the mechanical and electronic shutter. However, when using the electronic shutter with the Lumix GH4, you lose some effective dynamic range. There is more noise in the shadow areas.

This is because the GH4 is designed to use a 10 bit output in electronic shutter mode, to speed up the sensor readout, rather than 12 bits with the normal shutter. Less bit depth effectively means less dynamic range, and more noise in the shadow areas.

Speed of E-shutter readout

The GH3 electronic shutter had a readout speed of 1/10s, which is very slow. This leads to significant rolling shutter artifacts, that you can read about here. How does the GH4 compare?

One way to test the speed of the electronic shutter is to take a photo at a fast shutter speed in artificial light. For about a century or so, people have been using incandescent light bulbs for electronic indoor lightning. Even when used on alternating current (AC), the light is stable. Since the filament is heated, it emits light also when the alternating current is at zero.

However, traditional incandescent light bulbs are now being replaced with the energy saving fluorescent light bulbs. They tend to flicker at 100Hz (in Europe) or at 120Hz (in the US). The lights don't flicker at 50Hz and 60Hz, as you might expect. This is since during each period, the electrical current reaches two peaks, see the illustration below:

Here are images taken at ISO 1600, 1/400s with both cameras:

Lumix GH3

These results are quite easy to explain. With the Lumix GH4, I get about 3.3 stripes horizontally. Each stripe corresponds to 1/100s, hence, the total exposure takes about 1/30s.

The Lumix GH3 has a much slower electronic shutter. There are ten stripes in the image, which means that the exposure takes 1/10s. This causes a number of problems, making the electronic shutter mode fairly useless.

Here is a summary of recent cameras:

CameraElectronic shutter readout speedBit depth
Lumix GH31/10s12 bits
Lumix GX71/15s12 bits
Lumix GM11/22.5s
10 bits
Lumix GH41/30s10 bits

Panasonic has made the electronic shutters faster, but at the expense of the bit depth. This could cause somewhat worse dynamic range when using the electronic shutter.

Even if the Lumix GH4 has a faster electronic shutter readout than previous Panasonic cameras, you still risk rolling shutter artefacts. I took these five images in a row by holding the camera with an outstretched hand, using the Lumix G 14-140mm f/3.5-5.6 at 140mm, f/5.6, 1/320s, and using the E-shutter:

As you can see, even if the shutter speed is fast enough to avoid blurring images, camera shake still creates geometric distortions, due to the rolling shutter effect. So don't use the E-shutter with a long lens, unless you use a tripod. There could be cases where you might not mind these distortions, e.g., if you photograph non-straight shapes, for example nature.

If you intend to overlap the images later for HDR or panorama stitching, then you just cannot use the electronic shutter.

Mechanical shutter

The Lumix GH4 also has a new shutter unit. This allows for a faster flash sync speed than the GH3. Here are the flash sync speeds of recent cameras:

CameraFlash sync speed
Lumix GH31/180s
Lumix GX71/320s
Lumix GM11/50s
Lumix GH41/250s

Here is how the mechanical shutter works. In the illustration below, the shutter curtains move across the sensor in 1/100s (Lumix GM1), but most cameras have faster moving shutter curtains. In this example, the shutter speed is 1/500s. The maximum flash sync speed is 1/100s, but the camera implements a slower flash sync speed (1/50s) for some margin.

1. As the exposure commences, the first shutter curtain starts moving downwards, exposing the sensor.

2. After the duration of the shutter speed has expired, the second shutter starts moving downwards. In my example, the shutter curtains travel across in 1/100s, so the shutter speed of 1/500s corresponds to 1/5 total opening.

3-6. Both shutter curtains continue downwards, exposing each row of the sensor for a total of 1/500s.

7. As the second curtain shutter closes, the exposure is finalised.

If the scene is stable during this operation, there are no issues. However, if the subject moves horizontally, it will be skewed. Or if the light flickers, you will get horizontal stripes.

To test the mechanical shutter speed of the cameras, I made a Lego Technic contraption which spins a propeller at 33 rounds per second (2000 RPM):

These pictures were taken with the Olympus 50mm f/2 1:2 macro lens for Four Thirds:

Lumix GH3
f/2, ISO 3200, 1/4000s
f/2, ISO 3200, 1/8000s

We see clearly that the GH4 shutter is the fastest: The propeller rotates less during the exposure.

Mechanical shutter noise

Using an Iphone, I have measured the noise from the mechanical shutter of the GH3 and GH4, to compare them:

It turns out that the Lumix GH4 has slightly less noise, even if the shutter curtains travel faster. Good news!

Mechanical Shutter Noise
Lumix GH3


The Lumix GH4 improves upon the GH3 in terms of the shutter readout speeds. Both the electronic shutter and the mechanical shutter have become faster.

The electronic shutter suffers from the same problem as the GM1, though: The readout is limited to only 10 bits, which gives you more noise in the shadows. Now, is this a real problem? Probably not, unless you often raise the shadows a lot. Most of the time, you will not notice any real difference.

The mechanical shutter curtains move faster, and do so without causing significantly more noise. This is good, as the Lumix GX7 mechanical shutter is known for being rather noisy.

All in all, I think these features make both shutters more useful on the Lumix GH4, compared with the predecessor.


  1. I hope you mean for the mechanical shutters section, the gh3 shot was taken at 1/4000 and the gh4 at 1/8000. Not only does the gh3 not have 1/8000 shutter but I don't see how one mechanical 1/8000 shutter can be faster than another mechanical 1/8000 shutter lol.

    1. I was a bit sloppy in the description. The GH3 has a max shutter speed of 1/4000s, and that is what was used. That is why the GH3 image is a bit brighter.

      As for comparing the speed of the shutter curtains: That is a different topic than the speed of the shutter. So two shutter units can operate at different curtain speeds, even if the shutter speed is the same.

  2. Fredrik, Thank you for posting. That is very interesting.

  3. In order for the GM1 to have a flash sync speed of 1/50s, wouldn't the electronic shutter on the GM1 need an "electronic shutter readout speed" of at least 1/50s and probably 1/100s in practice for safety? John

    1. The electronic shutter cannot be combined with the flash at all. The flash sync speed relates to the mechanical shutter only.

    2. Since there is no first curtain mechanical shutter on the GM1, the flash sync depends on the speed of the electronic first curtain, otherwise the mechanical second curtain operating would overtake the electronic first curtain if it were travelling slower than the mechanical second curtain. Yes, I realize the shutter speed is determined by the time gap between the first and second curtains, but for a flash to evenly illuminate the sensor, the first curtain must be fully open before the second curtain begins its travel. The reason the GX7, for example, can have a faster flash sync is because it has a first curtain mechanical shutter. One reason the GM1 mechanical shutter top speed is only 1/500s is because it had to be slowed down to accomodate the relatively slow first curtain. It is too bad Panasonic doesn't use the technology available to Sony, since Sony provides an electronic first curtain option with a much faster electronic shutter readout speed largely avoiding the "rolling curtain" shutter effects. John

  4. Great comparisons! I would say that the GH4 is a nice step up from the GH3.

  5. Hi, I use my GH4 with some YungNuo 560 IV speedlights.
    If the flash is mounted on the camera the real sync Speed (to avoid banding) with mechanical shutter is 1/500s in 3:2 aspect ratio.
    If used with a radio trigger, the flash/camera combo is capable of 1/400s which is pretty fast for sport and action photography.
    I don't know why panasonic rate the official sync speed to 1/250s because the GH4 is capable of 2x faster.

    1. I think this is quite common. Normally, the manufacturer will put a fairly large margin in their flash sync speed, perhaps to avoid problems if there is a delay somewhere, e.g., in the flash controller.

      I have also tested the speed of the shutter curtains for some cameras, and I often find that the camera is capable of almost twice the flash sync speed.

  6. This comment has been removed by the author.