Panasonic GH4 Review

[Perplex]

Relax about what? Alot of people I know have been waiting and waiting and waiting for something like this come along. Its a hard product to make and we understand that but times about up and the need to move on to new things is here. I have chosen to just buy native M43 lens for our GH4 system. It seems like the smartest choice at the moment. 

which lens did you go with? I was thinking about the olympus 17mm 1.8 in the interim.  Also, getting a "el cheapo" ef to mft adapter to use on my canon sigma 18-35 wide open

[jonpais]

I generally don't wait around for nonexistent products. I work with what's available.

[Guest Ebrahim Saadawi]

4x smaller bands = finer gradient.

Simple enough yeah?

I've shot a plain sky in the 4K mode on the GH4, and graded heavily to induce the banding affect. What I had was an image with a few color band between the skies blue gradients, I count them and they were 5 in 4K.

I then took this image, rendering it via Vegas Pro 13 to 1080p, using the Avid DNxHD 444 10bit 185mbps codec.

The resulting 1080p video had the exact same amount of banding.

-I thought maybe the downscaler in Vegas is off, so I took the 4K file, put it in Mpeg Streamclip, and downscaled to the same Avid DNxHD and ProsRes HQ 10bit.

Also the resulting files have the exact same amount of banding.

Andrew, help me if I got something wrong but I just can't for the life me reducing the banding or increase gradient when downscaling.

(To anyone, if you don't have access to a GH4 and want to try yourself, use any camera at full resolution JPEG, which is 8 bit, it should band, and then try downscalimg that to 1080p and see if it reduced banding, lets get to the bottom of this!)

[Andrew Reid]

Well, breaking the image before you do the downscale and convert to DNxHD is kind of a 'shit in, shit out' situation isn't it?

 

Try grading it after processing instead!

[Darktide Media]

We are stuck, thinking of going the Lumix way since they have DFD and OIS but the Olyumpus look so much better made and I really like the idea of the 40-150 2.8 they are doing and the 300 F4 will be amazing for surf work. 

[amanieux]

I've shot a plain sky in the 4K mode on the GH4, and graded heavily to induce the banding affect. What I had was an image with a few color band between the skies blue gradients, I count them and they were 5 in 4K.

I then took this image, rendering it via Vegas Pro 13 to 1080p, using the Avid DNxHD 444 10bit 185mbps codec.

The resulting 1080p video had the exact same amount of banding.

-I thought maybe the downscaler in Vegas is off, so I took the 4K file, put it in Mpeg Streamclip, and downscaled to the same Avid DNxHD and ProsRes HQ 10bit.

Also the resulting files have the exact same amount of banding.

Andrew, help me if I got something wrong but I just can't for the life me reducing the banding or increase gradient when downscaling.

(To anyone, if you don't have access to a GH4 and want to try yourself, use any camera at full resolution JPEG, which is 8 bit, it should band, and then try downscalimg that to 1080p and see if it reduced banding, lets get to the bottom of this!)

it is because when you downsample 4k 4:2:0 8 bits to 1080p you get 4:4:4 but with only 10 bit for y and 8 bits for u and v , as your sky does not change its luma but just the shade of blue your are still in 8 bits for the shades of blue so the banding is still there, if you had shot in black and white then your banding thickness would have been /4 (going from 8 bit to 10 bit in the Y channel). or record the 4k externally in 4:2:2 when the new atomos recorder will be available, you will get 11 bit for the chroma (blue banding size /8) if you downscale to 1080p (4:2:2 10 bit 4k, downsampled to 1080p will give you 4:4:4 in 12:11:11 bits. and by the way, with your present test, even if you still get your 5 bands of blue, the border of the bands should be less pixelated by a factor of 2 as chroma precision is still 8 bit but chroma resolution is doubled.

[Butch]

I am really enjoying this camera.  Here's a test video I made.  Sorry it's so short.

 

[antoniopantoja]

I'm curious as to what the big June 3rd announcement was? I figured I'd see 50 people asking you even before today

[Michael1]

I am really enjoying this camera.  Here's a test video I made.  Sorry it's so short.

 

Cool test video!

 

Michael

[filmandy]

I am really enjoying this camera.  Here's a test video I made.  Sorry it's so short.

 

One of the better bits of GH4 footage I've seen. Well done. 

 

- Andy

[Perplex]

I'm curious as to what the big June 3rd announcement was? I figured I'd see 50 people asking you even before today

seems like it's delayed, and andrew is keeping hush about it until they "fix" it.  

[antoniopantoja]

seems like it's delayed, and andrew is keeping hush about it until they "fix" it.  

 

 

thank you brother!

[Michael1]

There is so much hyperbole in this post I logged in just to reply.  No, the stills from the 1/4 sized GH4 sensor will never be as good as the full frame 5d3, nor will the skin tones and colors.  It might "af" as fast as a Nikon D4 for single shot AF, but please understand that people that shoot D4's/1Dx's also need to TRACK moving subjects, and the GH4 cannot come close.  Watch the latest "The camera store" video where he takes out all the latest mirrorless to a motorcycle track - the full frame "dinosaur" DSLR smokes all of them in AF tracking and speed.  "better than most cases actually".  Really?  I lol'd.  Still as good as the 5d3?  I realize this is your "camera du jour" Andrew, but do you even read what you type?

 

I just looked at that video, and I have no idea what you are talking about.  The reviewer said, "If I really had to compare it (the D4S) to the GH4, I would have to say they are so close, as to be basically equal." 

 

Michael

[Tim Naylor]

Also depends on the exposure. James Miller has had some nice shadows out of it

 

Thanks for this reference. Very helpful. Anyone in NYC with a GH4 care to run some tests. I'm now thinking about using it as a B cam for some hard to mount  situations (for the Alexa). Contact me asap.

[Stefan Antonescu]

I am really enjoying this camera.  Here's a test video I made.  Sorry it's so short. 

[Tobler51]

Unfortunately I'm still not allowed to spill the beans.

 

The announcement is delayed!

 

Will be worth waiting for though.

 

Im using NIKON Lenses...still worth the wait? or is it just something for the Canon Boys in da house :)?

[arthrodoc]

Hardware glitch:

 

This happened to me after 2 weeks of daily use in Europe while traveling, after a few thousand pics...

 

Pressing the review button to check images, or going into any item in the review menus, will lock up the camera. It's on, with an image in the display, but frozen. Have to take the battery out to get out of the frozen state.  Still takes pics, but can't review.

 

Reset to base settings doesn't fix it.

 

There is a thread on the dpreview m43 forum (crash crash crash) where it appears two others have had the same. One user mentioned a fix ... and it works temporarily. If I use it over wifi with the Image App, I can review images, and go into the menu, both tethered or on camera. Unfortunately, as soon as I disconnect, or turn off the camera, it reverts to bad behaviour.

 

Note: this is only for image review (video or still.) It still records everything. I just can't review anything on camera. 

 

The Canadian service centre  for Panasonic (Nortown Photo Service; Panasonic farms it out to them for initial assessment) said they hadn't seen this (yet), and their tech contact at Panasonic said the same.  I figured I'd wait till the upcoming initial firmware upgrade to see if it fixes this, and if not, sent it in then.

 

From the behaviour, seems it is likely some obscure combination of events has set some software switch which won't reverse.... can't be common since I can only find the other single post about this...

 

..Mike

[hmcindie]

Well, breaking the image before you do the downscale and convert to DNxHD is kind of a 'shit in, shit out' situation isn't it?

 

DNxHD does not cause banding. Ever.

[arthrodoc]

Unfortunately, the 1.1 firmware does not fix this.

 

This is what I've sent to the service center:

____

 

Had trouble even upgrading ...  and it is still frozen after the upgrade.

 

With the bin file on a card, you put the card in, and turn on the camera, then press play ... guess what, it displays the screen asking if you want to upgrade (see attached pic) but then freezes, presumably because it is in the playback menu!

 

I tried the WiFi hack I described before... started it up, connected to WiFi, and after doing this a few different ways (card in initially, card out..) I got to a state once Wifi connected, where I got the same screen as in the pic, and when I tapped the <yes> onscreen, it worked!

 

I would note that the first time, I tried using the control wheels to change No to Yes, and it froze.

 

Unfortunately, after upgrading it is still frozen in playback.

 

I did check the version after upgrading, and it does say version 1.1 now.

 

As long as I don't press the playback button, and stay out of the playback menu, camera works.

 

If I do either of those, camera freezes, and I have to remove the battery to get it back functioning.

 

Hardware glitch:

 

This happened to me after 2 weeks of daily use in Europe while traveling, after a few thousand pics...

 

Pressing the review button to check images, or going into any item in the review menus, will lock up the camera. It's on, with an image in the display, but frozen. Have to take the battery out to get out of the frozen state.  Still takes pics, but can't review.

 

Reset to base settings doesn't fix it.

 

There is a thread on the dpreview m43 forum (crash crash crash) where it appears two others have had the same. One user mentioned a fix ... and it works temporarily. If I use it over wifi with the Image App, I can review images, and go into the menu, both tethered or on camera. Unfortunately, as soon as I disconnect, or turn off the camera, it reverts to bad behaviour.

 

Note: this is only for image review (video or still.) It still records everything. I just can't review anything on camera. 

 

The Canadian service centre  for Panasonic (Nortown Photo Service; Panasonic farms it out to them for initial assessment) said they hadn't seen this (yet), and their tech contact at Panasonic said the same.  I figured I'd wait till the upcoming initial firmware upgrade to see if it fixes this, and if not, sent it in then.

 

From the behaviour, seems it is likely some obscure combination of events has set some software switch which won't reverse.... can't be common since I can only find the other single post about this...

 

..Mike

[Jay Turberville]

It is only logical when you have 4 times the data you need 4 times the space. If there is an argument that 50mbit/s is "too few" for HD then something less than 200mbit/s is "too few" for 4K, by following the same argument. I am just pointing that out because only few people have written about this. 

 

Nobody seems to be talking about "compression efficiency" in regards to the A7S. But most bash the 50mbit/s, furthermore without having used it. How is that for conceptual pixel peeping?

 

That might be logical if you are going to display the image at the same final per pixel magnification and view it from the same distance.  In other words, if you were going to display it at twice the width and height of the 1080p image.  But we know that's not how videos/movies are typically viewed.  So no, it really isn't logical. The practical reality is that an image with higher pixel count can tolerate more compression because each pixel and the resulting artifacts from noise and compression is visually smaller.

 

If we want to get into pixel peeping, we need to consider question of the information density coming from the sensor.  When shooting 4K on the GH4, we are getting one image pixel for each sensor pixel.  But the GH4, of course, uses a CFA sensor.  And in the real world of CFA image processing, such images are not 1:1 pixel sharp.  This is due to the nature of the Color Filter Array as well as the use of the low pass filter in front of the sensor that reduces color moire.  A good rule of thumb is that the luminance resolution is generally about 80% of what the pixel count implies.  The color resolution is even less.

 

Years ago I demonstrated online via a blind test that an 8MP image from a 4/3s camera (E-500 I think) only has about 5Mp of luminance info.  Dpreview forum participants could not reliably distinguish between the out of camera 8Mp image and the image that I downsampled to 5Mp and the upsampled back to 8Mp. They got it right about 50% of the time.

 

So what's the point?  The point is that before you can even (mis)apply your logic about scaling bandwidth with image resolution, you must first be sure that the two images are similarly information dense. What is the real information density of the A7S image before it is encoded to 1080p video?  I think of CFA images as being like cotton candy - not information dense at all.  About 35% of the storage space used is not necessary from an information storage standpoint. (Though it may be useful from a simplicity of engineering and marketing standpoint.)

 

And even with this, there's surely plenty I'm not considering.  For instance, if the A7S uses line skipping  (I have no idea what it really uses) that will introduce aliasing artifacts.  How do the aliasing artifacts affect the CODEC's ability to compress efficiently.  The bottom line is that all too often camera users apply simplistic math to photographic questions that are actually more complex.  There are often subtle issues that they fail to consider (like viewing distance, viewing angles, how a moving image is perceived differently than a still images and more.)
 

Personally, as a guy who sometimes leans toward the school of thought that "if a little bit of something is good, then too much is probably just right.",  I kinda wish that the GH4 did offer 4K recording at 200mbs.  Why not given that the camera can record at 200mbs?  But the reality is that I can't really see much in the way of CODEC artifacts in the 4K footage I've shot with the GH4 so far.  So 100Mbs is probably just fine.  50Mbs on the A7s is probably pretty good as well.