Sharpness increase pressing record on P4K

I noticed that when I press record on the Pocket 4K that the image sharpness increases. When the recording stops sharpness goes back again. It's even visible with focus assist: I have colored lines on and they also get more, the moment I press record and again get less, when recording stops. Is this an expected behavior? I never noticed before.

Detail Sharpening is off.

Thanks in advance.

No one?

Could someone with the P4K maybe check? That would be great!

Thanks in advance!

If you are referring to the lines of color that appear when an object is in focus, I have tested it with my P4K camera and I have noticed that the intensity of the lines of focus looks different when we are recording and when we are in standby

If you are referring to the lines of color that appear when an object is in focus, I have tested it with my P4K camera and I have noticed that the intensity of the lines of focus looks different when we are recording and when we are in standby

Thanks for confirming, Francisco! When you have a test chart in screen filling size you will notice that the sharpness increase is not only in the colored focus lines but in the whole image (even when you turn focus assist off).

Indeed the image does sharpen when recording, I wonder why, perhaps a lower quality algorithm is in play during standby to save battery???

Indeed the image does sharpen when recording, I wonder why, perhaps a lower quality algorithm is in play during standby to save battery???

I also wonder. Also if this actually changes the recorded image or is just a "screen thing".

I sold my pocket4k, I cannot check directly with it.
Have you tested also with hdmi output?
Setup camera in fhd out windowed, hdmi connected and test on external monitor if change quality during recording or is a simple monitor optimisation.
I told in fhd windowed to avoid that fhd hdmi out will somehow optimised and reduce this effects.


Inviato dal mio iPhone utilizzando Tapatalk

I can confirm the ‘sharper’ screen appearance when actually recording. This occurs for both 4K and HD ProRes HQ and BRAW Q0 on my camera. I don’t know why, but I’m sure the improvement is also in the footage.

I sold my pocket4k, I cannot check directly with it.
Have you tested also with hdmi output?
Setup camera in fhd out windowed, hdmi connected and test on external monitor if change quality during recording or is a simple monitor optimisation.
I told in fhd windowed to avoid that fhd hdmi out will somehow optimised and reduce this effects.

Yes, happens on my SmallHD 501 via 1080p HDMI from 1920x1080p windowed sensor resolution on P4K.

I can confirm the ‘sharper’ screen appearance when actually recording. This occurs for both 4K and HD ProRes HQ and BRAW Q0 on my camera. I don’t know why, but I’m sure the improvement is also in the footage.

Could this be a bug where "Detail Sharpening" is always applied?

EDIT: Just tested it. I recorded two clips in UHD BRAW CB 12:1.
#1 no Detail Dharpening.
#2 high Detail Sharpening.

I can't see a difference in Resolve. Even at 400% zoom.

Thanks for sharing your test results! I’m surprised that there’s no difference. In the past I believe we were told that the Detail Sharpening was designed for live broadcast work and not for our usual purposes where sharpening could be judiciously applied in post.

I’ll take another look at this as well including for ProRes.

Edit
As previously reported, hitting Record appears to sharpen the image with Detail Sharpening Off. It seems to sharpen the image about the same amount as if Detail Sharpening On High while viewing the monitor.

But with Detail Sharpening On High, when you hit Record, the screen image and presumably the recorded output are even at a higher strength! That’s probably too sharp.

At this time I plan to set Detail Sharpening Off and see what BMD has to say. I understand this will actually apply sharpening under 7.9 in post. I’m not concluding this is a good behaviour or not until I have an opportunity to test with models or actors.

Method
I’m making these claims by looking at the images in the monitor with Coloured Lines Green 50%.

With Detail Sharpening Off, it’s difficult to see the Focus Assist but I can see my test fine text lines reasonably well and they are a little blurry.

When I hit Record OR set Sharpening On High, I see strong Green lines with better text definition.

When I hit Record AND Sharpening On High, I see even stronger Green lines and more readable fine text.

What this does to beauty shots will be interesting to determine. I might be able to test this more thoroughly next week.

EDIT: Just tested it. I recorded two clips in UHD BRAW CB 12:1.
#1 no Detail Dharpening.
#2 high Detail Sharpening.

I can't see a difference in Resolve. Even at 400% zoom.

That's no wonder as Detail Sharpening is not applied to BRAW, only to ProRes and to the HDMI output signal.

The monitor image appears to sharpen with both ProRes and BRAW. I’ll test next week with a model but good reminder about the HDMI as I can add the BMVA12G7 to the test that records it’s own ProRes HQ.

I might be able to test this more thoroughly next week.

Thanks for your testing and sharing the results.

That's no wonder as Detail Sharpening is not applied to BRAW, only to ProRes and to the HDMI output signal.

Thanks, that's super useful to know (should be in the manual, right?).

I assume the sharpening of the screen image (internal as well as via HDMI out) when hitting record has nothing to do with the Detail Sharpening option, because according to Robert the Detail Sharpening option works as expected.

Can anyone confirm the screen image sharpening in a firmware version prior to 7.9?
Or on other BM cameras than the P4K?

Thanks for your testing and sharing the results…

I have scheduled a test on Wednesday so should have results by Thursday or Friday.

Results of investigating Detail
Sharpening

Method
Recorded BRAW Q0 4K on BMPCC4K and
ProRes HQ HD on BMVA12G7

Varied Recording/ Detail Sharpening /from Off through Maximum settings.

Synced Timeline 2K with stacked V1 BRAW and V2 ProRes. So by disabling and enabling V2, I could quickly compare the clips.

Observations
With no detail sharpening, ProRes did not appear as sharp as BRAW. Perhaps that’s comparing apples and oranges but on the timeline both are 1080p30.

The sharpening options remind me of comparing the various flavours of ProRes as the differences between adjacent options are difficult to see. But by looking at the Maximum sharpening versus no sharpening, the ProRes easily reveals significant improvement. One might feel safer with the Medium sharpening but I didn’t notice anything ’bad’ with the sharpening in camera as I had expected. My pseudo documentary interview was shot with Detail Sharpening Maximum.

It appears Robert is correct that BRAW is not sharpened in camera, but regardless it’s almost as good as Detail Sharpening Maximum used ProRes. As Robert mentioned, even if the BMPCC4K is recording BRAW, sharpening is applied to the HDMI out feeding the BMVA12G7 for recording ProRes.

From now on I’ll shoot with Detail Sharpening Maximum until it causes me a problem. It seems to me to be cautiously implemented by BMD.

Very fascinating and useful, I'll up my in-cam sharpening for when I shoot ProRes now.

Sent from my SM-S908U using Tapatalk

Might I suggest a method to test for sharpening. It also helps to understand the math behind sharpening.

Make a target of a sharp delineation of uniform light gray and uniform dark gray and photograph.
Enlarge and examine across the edge with a waveform monitor.

Without sharpening the waveform should show a horizontal line then step to another level horizontal line at the edge.

If sharpening, the waveform will show an increase of the light gray as it nears the edge and a decrease in the line of the dark gray as it approaches the edge.

If blurring, the waveform will show a decrease in the light gray as it nears the edge and an increase in the line of the dark gray as it approached the edge.

In effect, sharpening increases the difference in tone at the edge (size of step) and blur decreases the difference. A flat normal step indicates none of either.

Thanks for sharing your results, Rick.

As Robert mentioned, even if the BMPCC4K is recording BRAW, sharpening is applied to the HDMI out feeding the BMVA12G7 for recording ProRes.

I don't think Detail Sharpening is the same sharpening that is applied to the monitor (internal as well as external via HDMI out). Because they stack / add up. Even when I have Detail Sharpening at High there is still more sharpening applied when I hit record. Also this increase in sharpness is not visible in the footage.

My assumption: Detail Sharpening ≠ sharpness increase on monitors when start recording

Yes, I was just interested to know what the button for Detail Sharpening was actually doing and if it was useful. It appears to be useful to me based on the documentary-style interview I recorded.

Thanks for the explanation Jeffrey. I judged the results by reading small text on some of the boxes for my familiar ‘test bed’ shots (in the office/edit suite above my fireplace mantle).

I just spoke to a friendly BM representative on the Blackmagic Summer Tour and he assumes that the camera uses different sensor read modes when standby vs when recording, with recording being the higher quality one obviously. Therefore the image appearance could slightly differ, when recording. They do it to save energy and prevent heat production in standby mode.

Smart.

Is this really desired behavior? I'm thinking it will be undesirable for people that are using their cameras with an ATEM mini.

Roger, I’m assuming this limitation only applies to the camera monitor. I don’t have ATEM. Is this something you can test with sharpness on the ATEM (recording) and from recording on the camera.

Edit
If it’s only a video display using ATEM, you have asked a very good question.

Is this really desired behavior? I'm thinking it will be undesirable for people that are using their cameras with an ATEM mini.

I totally agree, that would be undesirable.
If this change in sharpness when you hit REC appears on external monitors then it will appear on ATEM inputs I would think.

Reshot my Detail Sharpening tests with firmware 7.9.1. I’ve never shot anything with Sharpening in camera before 7.9 so even though it requires pixel peeping at 100% image size, I thought it would be interesting. The subjects are not in motion and include trees and part of an old home including a textured wall with pebbles in the stucco. With subjects in motion, as has been mentioned, it’s difficult to judge results.

Method
Shot BRAW Q0 and ProRes HQ with the BMPCC4K using Detail Sharpening Off, Low, Medium, High.

Although the different settings only display very small incremental differences to the image, it’s easier to see a difference between Off versus High.

My Conclusions
1) BRAW is unaffected by in-camera Detail Sharpening as mentioned by others.
2) BRAW appears sharper (due to apparently higher contrast) than ProRes for tree trunks but for small tree branches against a bright sky, ProRes appears sharper. Which could be more susceptible to false detail?
3) ProRes does not render the stucco wall and fir tree truck with as much detail as BRAW until you use Detail Sharpening Medium.
4) Initially I thought I would standardize on High but that does seem to be extreme for ProRes so I’m going to settle on shooting with Detail Sharpening Medium with BRAW and ProRes.
5) Detail Sharpening is passed to an external monitor so for example the BMPCC4K BRAW is unaffected but the ProRes recorded as backup on the BMVA12G7 does inherit the camera’s Detail Sharpening setting.
6) With more experience using these settings, I think its reasonable to expect some scenes, especially those in which the subjects are people, there’s likely no need for Detail Sharpening and ProRes could produce a better image. Where the scene is dominated by natural objects with detailed texture, BRAW might be the better choice (or ProRes with Detail Sharpening Medium). I’m concluding flexibility has some merit.

Quantization
I haven’t recently looked at comparing the BRAW Constant Quality quantization values, but as Jamie points out, if your media needs dictate longer recording times, Q5 produces very good results. When I run firmware 7.x tests, being economical isn’t as important as seeing the theoretically best possible results.

BRAW versus ProRes
John Brawley has also endorsed shooting ProRes since it’s generally easier to manage when workload is heavy such as shooting a few terabytes per day. And easier for editors using Resolve in terms of potentially simplifying workflow. And ProRes has other advantages in the image it creates. Still I like the 12bit Colour in BRAW compared to the 10bit Colour in ProRes HQ on the BMVA12G7 and BMPCC4K. When I shoot on the UM4.6K, I use 12bit ProRes 444.

Is this really desired behavior? I'm thinking it will be undesirable for people that are using their cameras with an ATEM mini.

I totally agree, that would be undesirable.
If this change in sharpness when you hit REC appears on external monitors then it will appear on ATEM inputs I would think.

Can confirm: sharpness increase appears on external monitors. Which is logical, because the sensor readout happens before sending the signal to other monitors.