GH4 4k 8bit vs external 4k downconverted to 1080p 10bit?

As I read it, the Adam Wilt article is talking about the difference between INTERNAL HD recording on the GH4 versus converting from QHD to HD in either the Odyssey or in FCP X.

But the question was about recording the output of the 10-bit 1080p down sampled HDMI output of the GH4 while shooting in QHD. I have not personal experience with this, but I have read multiple times that the HD output over HDMI is far superior to internal HD recorded material, which makes purchasing an HD 10-bit recorder for the GH4 and interesting and inexpensive option.

David Rokeby
 
Mitch Gross said:
If you look at the Adam Wilt article I referenced earlier in this thread there are examples.
Click to expand...
Like David already mentioned ...Adam compares 8 bit internal 1080 to 10 bit 4K external footage. Not what you where suggesting nor what I was asking you about.
Mitch Gross said:
While I do not have samples of that, I can state that there are differences to be seen.
Click to expand...
Then what are these differences? You are talking about a world of differences, while all I can do is post examples I hardly can keep apart myself.
 
I have been shooting on a Ninja 2 for a while now. Started with my EX1R and now with the GH4. The GH4 downconvert from 4K to 1080 works very well. I prefer it to the internal codec. I shoot both ProRes and DNxHD, depending on what I am doing.
 
I se color sampling differences as well a edge smoothing, particularly in high frequency detail with movement. If you do not see these differences, then I guess you can say there is no difference for your work.
 
Mitch Gross said:
I se color sampling differences as well a edge smoothing, particularly in high frequency detail with movement. If you do not see these differences, then I guess you can say there is no difference for your work.
Click to expand...
If you see differences it means that your are actually looking at test results, no? Why don’t you share them with us.
In the article of Adam I come across this sentence: ‘Covergent Design’s Mitch Gross sent me several sample clips’… It means you have samples!

Wouldn’t it be cool if you did a 10 bit test between ‘4K down convert’ from the GH4 to the 7Q+ versus 4K from the GH4 downconverter in the 7Q+ to 1080. I know it is some extra work but as you wrote ‘Easy to test, just try it both ways’.
 
I don't have that material any more, and it was not designed for public distribution. I don't have a GH4 camera at my disposal.

If you do not see a difference then for you there is no difference to be had.

When I tested (and this was nearly a year ago), I could see differences.
 
Mitch Gross said:
and it was not designed for public distribution.
Click to expand...
Strange, you want to advertise how much better the 7Q+ is than the internal down converter of the GH4 ....and then you decide not to publish the test but instead use 1080 material that came straight out of the camera ...which makes even internal Down converted footage look good. What am I missing?

For those who don't know the difference between 10 bit 1080 to a recorder and 10 bit UHD down converted to a recorder ...here is a picture.
10bit1080.jpg
 
I'm not trying to start a fight here. We published our own tests back when we released this function in February. Adam Wilt also shot his tests. I shot some quick versions of my own material when I was experimenting with the process before the release. There's nothing nefarious here.

In certain situations I can see a clear difference between UHD->HD out of the camera v. UHD->HD out of the Odyssey. I will also note that when in that mode on the Odyssey, the Focua Assist and Pixel Zoom are both utilizing the full UHD signal, making accurate focus far easier and more precise. Separately and what were published as tests, the HD internal recording of the GH4 v. the UHD->HD path to the Odyssey for recording is a huge and obvious difference in image quality.
 
I do not think anyone is suggesting that an Odyssey does not offer a variety of advantages over recording UHD->HD straight out off the GH4 into something like a ninja star. And I think everyone is aware that internally recorded HD in the GH4 is very significantly inferior to UHD recorded in 10-bit to Odyssey and then down sampled to HD.

But I think that we are unconvinced that there will be substantive difference between HD down sampled from UHD after capture by the Odyssey, and UHD output as HD to a 10-bit capable HD recorder. The scaling might be better in the Odyssey versus the UHD->HD conversion at the HDMI output of the GH4, but scaling down an image by 2 in the x and y is a pretty straightforward process algorithmically and I am not convinced that fancy scaling approaches yield any substantive gain. I have coded my own scaling algorithms, and so speak from some not insignificant experience. But I also don't pretend to know everything.

I would welcome examples that prove me wrong. I like learning things that show me how to improve image quality. And I totally understand, Mitch, if you do not want to spend any time providing these examples.
 
I'm simply not in the position to be providing examples at this time. Lots of other stuff going on right now and I don't have a GH4 handy.

I will note that there are many ways to skin the cat when it comes to downscaling, and simple 2:1 Nearest Neighbor (aka Impulse) sampling is often not the most appropriate. There's also Parzan, Cubic, Lancos 6, sinc and other methods. They all have their own advantages/disadvantages, and we use a mix in our system. I question if that it true in the GH4, as the camera only has so much processing power to spare what is essentially intended as a preview monitoring output.
 
Mitch Gross said:
...as the camera only has so much processing power to spare what is essentially intended as a preview monitoring output.
Click to expand...
You're doing it again why? Only once the HDMI output is mentioned as a monitoring signal in the manual and that is when the bit mode is set to 8 bit 422, as you can still record internally when HDMI is set to 8 bit 422.
10 bit 422 is seen as ideal to save your footage to an external device.
The only sidenote 4KDowncovert gets is that the signal has a slight lag. Probably because they prevert a decend signal above a fast signal out.


If we start pixel peeping 4K downconvert vs the supersampled 7Q+ on high dynamic moving footage at 800%, we will no doubt see some differences in favor of the 7Q+ ...and it would be a disgrace if we didn't. But those differences take place in micro world. The world of professionals using Reds and alike, the world the 7Q+ was mentioned for.
 
My favorite DP and I made an interesting test yesterday using my rental GH4 and a new Video Devices PIX E5; we recorded 4K internally to a SD card at 8bit and externally on the PIX E5 in Pro Res HQ at 8bit and also at 10bit. To both of us everything looked great, if there was a difference and there may be, it would take younger eyes than ours to see it.

The “Big Plus” is when recording in 8bit with the PIX E5 we have a SD Card in camera recording as a back up.
 
bunk said:
You're doing it again why? Only once the HDMI output is mentioned as a monitoring signal in the manual and that is when the bit mode is set to 8 bit 422, as you can still record internally when HDMI is set to 8 bit 422.
Click to expand...
This is not my first rodeo.

If the internal recording is set to 4K and the video output set to HD, then the logical assumption is that the intended use of this output is for viewing the image on an HD monitor. It is for convenience. This is the same as the SD video outputs of HD recording cameras that existed for many years.

As a manufacturer of near-camera recorders, I can tell you that the camera manufacturers by and large do not think about the off-camera recording all that much. The priority in design goes to internal recording, to their own internal system. That is the mentality. So when I see a mode for internal recording in UHD with an HD video output, I feel fairly comfortable in assessing it as a monitoring output.
 
Vardalos said:
My favorite DP and I made an interesting test yesterday using my rental GH4 and a new Video Devices PIX E5; we recorded 4K internally to a SD card at 8bit and externally on the PIX E5 in Pro Res HQ at 8bit and also at 10bit. To both of us everything looked great, if there was a difference and there may be, it would take younger eyes than ours to see it.

The “Big Plus” is when recording in 8bit with the PIX E5 we have a SD Card in camera recording as a back up.
Click to expand...
Differences are not particularly seen when directly viewing the material. Differences are generally seen when manipulating the material for color correction, or the concatenation when repeatedly recompressed for broadcast, satellite uplink, web distribution, etc. This is when banding, motion artifacts, high frequency detail distortion and other issues become more apparent. This is not particular to Panasonic or the GH4. I am speaking now in general terms.
 
Mitch Gross said:
Differences are not particularly seen when directly viewing the material. Differences are generally seen when manipulating the material for color correction, or the concatenation when repeatedly recompressed for broadcast, satellite uplink, web distribution, etc. This is when banding, motion artifacts, high frequency detail distortion and other issues become more apparent. This is not particular to Panasonic or the GH4. I am speaking now in general terms.
Click to expand...

+1
 
Mitch Gross said:
So when I see a mode for internal recording in UHD with an HD video output, I feel fairly comfortable in assessing it as a monitoring output.
Click to expand...
Mitch, you must have hurt your head one time to much during a rodeo. There is no talk about internal UHD recording with monitoring to a monitor.
This thread is about 10 bit recording.
The OP wants to know if 10bit 4K Down convert to any recorder is on par with 10bit UHD to any recorder scaled in post to 1080. The answer to that question is ...yes! There are differences when you zoom in 400 to 800% ...micro differences.
The priority in design goes to internal recording, to their own internal system.
Click to expand...
.Aha! That explains why the YAGH sells so poorly. People knew they where conned by Panasonic.
Thanks Mitch for the insight.
 
bunk said:
Mitch, you must have hurt your head one time to much during a rodeo. There is no talk about internal UHD recording with monitoring to a monitor.
This thread is about 10 bit recording.
The OP wants to know if 10bit 4K Down convert to any recorder is on par with 10bit UHD to any recorder scaled in post to 1080. The answer to that question is ...yes! There are differences when you zoom in 400 to 800% ...micro differences.
.Aha! That explains why the YAGH sells so poorly. People knew they where conned by Panasonic.
Thanks Mitch for the insight.
Click to expand...
I feel like I'm talking in circles here. Forgive me as I have not had a GH4 in front of me in a while.

There are either five or six ways to end up with an HD signal after having shot with a GH4.

- Shoot with the camera in HD mode. Record internally.
- Shoot with the camera in HD mode. Record externally.
- Shoot with the camera in UHD mode but set to output an HD signal over its video output. Record externally. (Does this mode exist on the camera? I think this is where I may be getting confused.)
- Shoot with the camera in UHD mode. Record internally. Down convert in post.
- Shoot with the camera in UHD mode. Record externally. Down convert in post.
- Shoot with the camera in UHD mode. Record externally in Odyssey7Q+ using our 4K->HD mode to down convert at point of recording.

The first two options are pretty sucky, in my humble opinion. I just don't like what Panasonic does here.
The third option may not exist but if it does I have a recollection of it showing issues. Dave, my mind is going.
The fourth & fifth options will work and yes, there is not much difference in quality between them and the sixth option, other than the fourth option uses the internal codec which can show various forms of artifacting both in shadow noise and high frequency motion.

An advantage of the final option is that we use 1/4 of the SSD space to save the file compared to saving a full UHD signal for post conversion. This means 4x longer record time, 1/4 offload time, 1/4 post storage space for the original files, no downconversion stage in post and no second storage of the downconverted file. It adds great efficiency to the workflow for essentially the same result.

As to my previous comments about internal v external recording, this is not a dig at Panasonic. It is the mentality of all of the manufacturers as a general rule.
 
Mitch Gross said:
I feel like I'm talking in circles here.
Click to expand...
Mitch you're priceless. It’s not you who is talking in circles it’s everyone else. You come up with a new angle each and every time …now it’s memory loss.

Here is the way you can externally record with the GH4 in any format, but we choose Rec Format > MOV > UHD.

In the menu go the the HDMI Rec Output and…

- under Bit Mode choose 8 bit or 10 bit out. We choose 10 bit out
- under Info Display choose On or Off. We choose Off
- under 4K Down Convert choose Auto, 1080p or Off. We choose the first two times for Off and the third time for 1080p for our comparison test.

Hook a Q7+ to the GH4

1] Record with Off and down convert to HD in post.
2] Record with Off and let the Q7+ do its down covert magic to HD.
3] Record with 1080p to the Q7+

Open the recordings in the NLE off your choice and compare them.

Out of experience with my own recorder I can tell you there there will be not much difference between 1 and 3 and since there is not much difference between 1 and 2 (your words) there will be not much difference between 2 and 3.

And guess what 3 is also only 1/4 in disk space.

Now replace the Q7+ by a VideoDevices recorder or an Atomos recorder or a BlackMagic recorder …and you will get the same results if you set 4K down Convert to 1080p.

Over and out
 
You must log in or register to reply here.
Back
Top