Please select Into the mobile phone version | Continue to access the computer ver.
Pocket internal mic test & fixed shutter/iso in low light
746 13 2019-3-21
Uploading and Loding Picture ...(0/1)
o(^-^)o
David_Harry
Captain
United Kingdom
Offline

Hi.
Having returned my first Pocket and then buying a second one, I'm now trying some shooting scenarios to work out just what is it that the Pocket is best suited to.

I've now compared its footage to the best action camera and best smartphone to see how well its footage stacks up for pure image quality and resolution and can say with certainty that the Pocket is the best of the lot for pure recorded image quality when exposed and focused correctly.

So with that in mind I decided to try a very specific test for dialogue quality using the internal mic system and setting the picture manually.

This test was done in fairly low light so I was able to fix the exposure without the need of a ND that'd usually be required in brighter scenarios. In this test I've set the shutter to a fiftieth as I'm shooting 25 and set the ISO to 100. This is so I can check for things such as jittering within movements. I've also only used one channel of audio, centred it in post and applied a limiter and raised it to -.3db to get a healthy output level. I recorded the audio in high mode.

I've used D-CineLike for my gamma. Although I don't usually use such things, it helps to raise mid tones and help exposure in low light scenarios. It also works out that in certain scenarios and exposures, the D-CineLike gamma works OK within a standard rec.709 gamma space without the usual necessary adjustments that’d be required.

I'll post more examples as I do them and add to my downloadable files if I have any takes that may be useful to others.

Cheers,
Dave.








2019-3-21
Use props
DJI Gamora
Administrator

Offline

Hi there @David_Harry, thank you for sharing this video showing how you test and tweak your Osmo Pocket settings. Well done, Cheers!
2019-3-21
Use props
David_Harry
Captain
United Kingdom
Offline

DJI Gamora Posted at 3-21 14:59
Hi there @David_Harry, thank you for sharing this video showing how you test and tweak your Osmo Pocket settings. Well done, Cheers!

Hi Gamora.

You are very welcome.

Like I say in the video, I will be concentrating on what the Pocket does best and not what it can't do.

The image quality of the Pocket is excellent for such a small camera and will be perfect for certain shooting scenarios. It's just a simple case of putting some effort into testing for those scenarios and using it for those.

Cheers,
Dave.
2019-3-21
Use props
DJI Gamora
Administrator

Offline

David_Harry Posted at 3-21 15:07
Hi Gamora.

You are very welcome.

It's my pleasure and we really appreciate your test with the Osmo Pocket. Keep on sharing and expecting more. Cheers!
2019-3-21
Use props
Ray-CubeAce
Captain

United Kingdom
Offline

Hi Dave.
Interesting video.
First I would say the audio had less fidelity to it in some ways with the high-frequency end suffering most.
Birds in the background and similar noises. Also, voice still sounds a bit weird and not natural with an almost burbling quality to it, but clear and understandable nonetheless.
I can't really comment on the video too much as you are using it as intended, ie. for a piece to camera. Not my normal way to film. I am detecting some jitter on lower speed pans but the higher speed twirl was surprisingly smoother than I would have expected.
I'm surprised at your frame rate choice though. 25 fps is great for a TV set but most of us will be watching on monitors with much higher frame rates. If I were you I would try again at 60fps and compare results. I've been using 30fps and compared to looking at my sons' output from his pocket at 1080p (neither of us has used superfine as yet) his video output is much sharper and in general, doesn't suffer any perceivable jitter at all that I've noticed. I can even make out the veins in his eyes when talking to the camera.

I'll be posting one of mine later this week using the Pockets own colour profile on full automatic to show how I'm getting on with my X axis and ministry of silly walks to try to reduce bobbing. Then you can tear my effort to bits
Good trial considering recent weather. Keep them coming.
What were your own conclusions?
2019-3-21
Use props
David_Harry
Captain
United Kingdom
Offline

Ray-CubeAce Posted at 3-21 15:22
Hi Dave.
Interesting video.
First I would say the audio had less fidelity to it in some ways with the high-frequency end suffering most.

Hi Ray.

My own conclusions are that after dialling back my expectations, I'm appreciating it more for what it can do and it will nail it down to specific tasks.

I was surprised at how well it handled focus with such low light. I'm perfectly fine with dialogue quality as there's not many things that'll be very good with internal mics.

I was very surprised by how well it held the blacks when I went into the dark section, this may not have been translated so well after the YouTube compression. I know it was black, but some cameras don't show it properly.

As far as I'm concerned, motion artefacts were effectively none existent. Even some higher priced cameras sometimes won't look as good. I'd say that that sensor readout of the Pocket is faster than average.

As for frame rates etc. I shoot 25 because of traditional PAL frame rates and it will match my other cameras which are video cameras. I'll only ever shoot 50 if I want proper optical/temporal half speed, but I don't do much slomo anyway.

On that point. I have my monitor synced for PAL frequencies, so I don't see certain issues that others may when viewing 25/50 on a monitor.

Also, and I'm not saying that this applies to your understanding. The pocket does not shoot at either 30 or 60 FPS. This is why you sometimes see problems with Pocket footage and other video cameras when watching on a computer monitor. Many computer monitors will be refreshing at 60 and the footage being played is not 60 or 30. This will cause the perception of jitter due to lack of sync between the source and monitor depending on how the GPU/monitor respond to a video's native time base this will cause interpolation. I've also seen many examples of this in peoples' edits, which results in permanent interpolation in the output file that can never be reversed. Again, this depends on how projects are setup in an NLE and how a NLE handles the frame rate of the source compared to the timeline/project.

I've still got some ways to go before I'll be happy that I'm using the Pocket to the best of its abilities, so will be dialling back my expectations and observations/comments, unless I notice proper technical issues and proper mechanical & software problems.

In fact, I'll be posting a clip soon with a very odd audio problem to see if anyone else can identify it and give me some advice.

I'm also going to hold back on any critique of the audio until I've had time to properly test it. I hope that there are ways of using the Pocket and adaptor to good effect and if so, I'll be doing some examples and tutorials.

I look forward to your future tests and videos.

Cheers,
Dave.
2019-3-21
Use props
BobWinNV
Captain
Flight distance : 284209 ft
  • >>>
United States
Offline

Very informative video Dave.  Thanks for sharing.
2019-3-22
Use props
NoëI
Second Officer

Netherlands
Offline

Nice one David. But wow... Can you talk!!!
2019-3-22
Use props
David_Harry
Captain
United Kingdom
Offline

BobWinNV Posted at 3-22 10:44
Very informative video Dave.  Thanks for sharing.

You're very welcome Bob. Cheers, Dave.
2019-3-22
Use props
David_Harry
Captain
United Kingdom
Offline

NoëI Posted at 3-22 11:00
Nice one David. But wow... Can you talk!!!

Hi NoëI. If you thank that's a lot of talk, you don't want to hear me after a drink

Seriously though NoëI. The problem with many people who do examples or guides is that they quite often don't give all the information for the test conditions and/or settings with the equipment being used. This is why my videos tend to be long, I try to be as thorough as possible. Plus I also try to give extra information pertinent to what I'm doing, for instance extra info about video or audio, which may be helpful to others using the equipment.

I've also worked in audio/music production for over 30 years and video post for about 20. So it becomes a bit of an occupational hazard when I get going   And it would appear that I do like the sound of my own voice

Cheers,
Dave.
2019-3-22
Use props
Ray-CubeAce
Captain

United Kingdom
Offline

David_Harry Posted at 3-21 16:07
Hi Ray.

My own conclusions are that after dialling back my expectations, I'm appreciating it more for what it can do and it will nail it down to specific tasks.

Hi Ray.

My own conclusions are that after dialling back my expectations, I'm appreciating it more for what it can do and it will nail it down to specific tasks.

I was surprised at how well it handled focus with such low light. I'm perfectly fine with dialogue quality as there's not many things that'll be very good with internal mics.

I've found the focus and noise handling at night to be very good as well.

I was very surprised by how well it held the blacks when I went into the dark section, this may not have been translated so well after the YouTube compression. I know it was black, but some cameras don't show it properly.

If you did any editing, what video codec did you use? I find the HEVC codec to be good for MPEG-4 encoding.

As far as I'm concerned, motion artefacts were effectively none existent. Even some higher priced cameras sometimes won't look as good. I'd say that that sensor readout of the Pocket is faster than average.

I find that to be true on original footage but when translated across the internet so others can view a lot will depend on their connection speed, buffer capability etc. For instance, while watching your video a second time, I recorded the frame rate my browser was giving me real-time and it bounced around by +/- several frames per second and can vary each time I connect to it.

As for frame rates etc. I shoot 25 because of traditional PAL frame rates and it will match my other cameras which are video cameras. I'll only ever shoot 50 if I want proper optical/temporal half speed, but I don't do much slomo anyway.

That makes sense. I wondered.

On that point. I have my monitor synced for PAL frequencies, so I don't see certain issues that others may when viewing 25/50 on a monitor.

Also, and I'm not saying that this applies to your understanding. The pocket does not shoot at either 30 or 60 FPS. This is why you sometimes see problems with Pocket footage and other video cameras when watching on a computer monitor. Many computer monitors will be refreshing at 60 and the footage being played is not 60 or 30. This will cause the perception of jitter due to lack of sync between the source and monitor depending on how the GPU/monitor respond to a video's native time base this will cause interpolation. I've also seen many examples of this in peoples' edits, which results in permanent interpolation in the output file that can never be reversed. Again, this depends on how projects are setup in an NLE and how a NLE handles the frame rate of the source compared to the timeline/project.

Most editors will pick up frame rates automatically but different editors will read the file information differently. I see this time and again on editing forums when people try to mix old footage with new. It can and does cause all types of problems that not even building a new frame table can resolve. If you are using two sources that can give the same output this is less of a problem.

I've still got some ways to go before I'll be happy that I'm using the Pocket to the best of its abilities, so will be dialling back my expectations and observations/comments, unless I notice proper technical issues and proper mechanical & software problems.

Same here. Early days yet

In fact, I'll be posting a clip soon with a very odd audio problem to see if anyone else can identify it and give me some advice.

Look forward to it.

I'm also going to hold back on any critique of the audio until I've had time to properly test it. I hope that there are ways of using the Pocket and adaptor to good effect and if so, I'll be doing some examples and tutorials.

I look forward to your future tests and videos.

That won't be long   
2019-3-22
Use props
NoëI
Second Officer

Netherlands
Offline

David_Harry Posted at 3-22 13:08
Hi NoëI. If you thank that's a lot of talk, you don't want to hear me after a drink

Seriously though NoëI. The problem with many people who do examples or guides is that they quite often don't give all the information for the test conditions and/or settings with the equipment being used. This is why my videos tend to be long, I try to be as thorough as possible. Plus I also try to give extra information pertinent to what I'm doing, for instance extra info about video or audio, which may be helpful to others using the equipment.

Dave. To bad you live on the other side of the North sea. I would have loved to get some camera lessons from you :-)

I've already learned a lot from your "DJI Osmo Pocket internal mic test, fixed shutter & ISO in low light" video.
2019-3-23
Use props
David_Harry
Captain
United Kingdom
Offline

Ray-CubeAce Posted at 3-22 15:31
Hi Ray.

My own conclusions are that after dialling back my expectations, I'm appreciating it more for what it can do and it will nail it down to specific tasks.

Hi Ray.

The codec I use for my masters is an intraframe coded not and interframe one. It's actually for 10bit 4:2:2 material which may sound like overkill for 8 bit 4:2:0 material but will ensure a visually lossless transcode and for the best part tachnically lossless.

As for the NLE. I'd be very careful using an NLE that auto detects a clip's frame rate and sets its project by it. You should always setup a project manually and tell any NLE what to do. If you let the NLE set certain parameters, be it global for the project or for individual sequences, depending on the NLE, in certain instances you'll run into a load of trouble.

The main problem when settings are taken out of your hands is that you may make a mistake and only notice way down the line in an edit, at which point you've probably wasted all that time.

There's also the issue of how an NLE deals with the temporal difference of mixed frame rate time lines and material, nearest neighbour, frame blending etc. which will have a huge impact on the look of the interpolation.

If I get time I'll try and do a quick tutorial explaining this stuff in context to Pocket footage.

The biggest problem is when users and manufacturers refer to footage as being 30 FPS and 60 FPS which is incorrect with video cameras, as the Pocket doesn't shoot 30 or 60. If pocket footage were used in either a 30 or 60 FPS protect/timeline and it hadn't been adjusted for that time base, it will interpolate and can look quite nasty.


Cheers,
Dave.
2019-3-23
Use props
Ray-CubeAce
Captain

United Kingdom
Offline

David_Harry Posted at 3-23 01:40
Hi Ray.

The codec I use for my masters is an intraframe coded not and interframe one. It's actually for 10bit 4:2:2 material which may sound like overkill for 8 bit 4:2:0 material but will ensure a visually lossless transcode and for the best part tachnically lossless.

Hi Dave.

The main reason I considered the Osmo Pocket is that it has identical frame rates to my Nikon DSLR.
The main differences are audio quality with the Nikon using PCM 16 bit 48kHz at 1536kbs and obviously photosite size differences. Both are 59.940 and 29.970 respectively as well as being matched across the other frame rates which I obviously wasn't aware of until the updates came out.
The Nikon audio amps are actually quite good as is the auto levels although I still tend to set them manually.  The only area I'm finding difficulty in is setting Kelvin temperature for white balance with the Pocket which so far I've left on auto.

While my editor does detect frame rates It also has a splash screen at startup of a project allowing the operator to set the parameters. Where it is useful is if I ever wanted to mix clips from earlier material which would warm me of incompatibilities.

There may be a difference in the GOP structures of the two cameras I'm not aware of and I haven't as yet had time to set up a project using the two together. This will be my next round of tests.

I'm only using HEVC for uploading to Flickr because of the file upload limitations and can decrease file sizes by two thirds over using MPEG-4 with no real visual degradation.  If I have a video that keeps within the file size limit Flickr allows then I stick to MPEG-4.

I don't have 10bit or 4K monitors yet and while the contrast ratios are quite good I personally can't see a difference between that and an 8-bit output beyond some videos become blocky in the shadows and have banding in their skies, possibly because of a lack of bit depth but more probably because the videos lack a smooth contrast themselves either through the re-encoding after editing or tweaking of parameters of the producer or the re-encoding that goes on at the various hosting sites. Difficult to tell with any certainty.
All can say for certain is whatever I produce and see on my own screen is not often what I get on playback once uploaded to the net for whatever reason.
Now, if I could figure that one out, I'd be a happy bunny.
2019-3-23
Use props
Advanced
You need to log in before you can reply Login | Register now

Credit Rules