Here is my first video test of the Dji Osmo Pocket audio adaptor.


It may raise more questions than answers and I will apologise for the quality of the video in advance.
I am not a YouTuber and not particularly good in front of the camera but here goes.
It should answer some questions but at present raises others.
If you have any questions I'll do my best to answer them.

Ray.                       


https://www.dropbox.com/s/xn3lifogjal1844/Pocket%20Adapter%20Test.MP4?dl=0

Thank you for this video!  You definitely did a great job explaining your findings and should consider creating a YouTube channel.  

Thank you, Joe, but it was a rushed job quality wise.
Each clip was the first take I took and should probably have re-shot some of the footage and spent time aligning clips more precisely, but I didn't realise my screen grabber worked at a slightly different frame rate than my cameras until I tried to put the clips together. It made alignment real-time a nightmare trying to match frame rates and timeline position. Production-wise, it's a mess but as accurate information wise as I could make it.

Hi Ray.

Thanks for the effort, these things don't just take 5 minutes to do.

Going by the take that had the adaptor's audio on it and knowing the mic used, I'd have to say that it's not very good, I can clearly hear encoding issues. This sounds like the encoder, although this could be the encoder that you've used for the MP4 upload. In any event, I'd say that it is not sounded at all well and if the only way to get a good level into the Pocket is to pregain by +20db, then that's going to preclude nearly everyone who's bought the Pocket and the adaptor as most nearly all of them won't have a mic with a 20db boost on it.

The first tests you made were a bit difficult to understand as there was no audio being played from the Pocket recordings. Would you be able to upload that footage unedited just straight file uploads, MediaFire.com give free storage space if you're close to your Dropbox limit.

From what I could see visually the first tests weren't very good either for the adaptor. I'd say that it isn't stereo and the levels are extremely low for its gain structure.
Again, thank you for all your effort. As has been suggested, you may want to do some YouTube stuff with this.

Cheers,
Dave.

Hi Ray.

I've just had another listen and like a complete tool I missed the thing that would have told me what was causing the audio compression artefacts. Whatever your camera was using to pick up audio, it doesn't suffer the same problems that are in the recording of the adaptor on the end take.

This without a shadow of a doubt is proof that the Pocket is causing serious issues with external mics. If you listen to the camera audio in the first half it doesn't have any aliasing or encoding problems but the Pocket audio at the end does. What's even worse about that is that I'd not heard that particular problem when using the Pocket's internal mics.

Just to be clear to anyone who doesn't understand what I've just said. Both audio takes, Ray's shooting camera/audio and the Pocket's adaptor audio take are both subject to the same final audio encode for Ray's upload file. If the artefacts werer introduced by Ray's final encoding it would effect both audio sources but it doesn't, clearly showing that the Pocket audio already had the issue. You can hear the problem more when Ray is saying anything with 'S' in it, this isn't siblance either.
This issue is usually a result of the A/D, audio bitrate or any number of factors during the encoding, or combinations thereof.

I now feel lucky that my Pocket had an issue with it that meant I could return it for a refund during its first 14 days. If I'd have kept hold of it and bearing in mind that I'd only bought it knowing the mic adaptor was being released. I'd have been beside myself if these were the issues that I'd have faced with the system.

I'd say that for close position stuff, vlogging etc. Users will probably have more success using the internal mics and just use a bit of protection on them.

The Pocket is one of the best small cameras around as far as picture quality is concerned, there isn't an argument in that. But it completely fails to match its picture quality and flexibility with its ability to capture external audio.

Cheers,
Dave.

Hi David.
Just to be clear.
The video recorded in the kitchen is using audio from my Nikon DSLR onboard stereo mics. I have not used audio from the Osmo Pocket in that clip. Only showing the waveforms produced by various mic output configurations showing the phase of the waveforms and amplitude levels from the Osmo Pocket.

The maximum sound levels I could get from the stereo mic in its hyper-cardioid configuration would be fine for loud concerts but probably little else. It was a low impedance electret condenser mic using a 3v internal battery. The mic works fine with anything else I have but is a bit hissy by today's standards. It normally sounds quite natural when using it with my minidisc recorder. (Yes I still have minidisc)

I agree that a 20dBm boost is a massive amount of signal gain to have to find but that has made me think is it the Pocket protecting itself from clipping? I didn't check the internal mic's level in the same conditions. I will check that this evening. I also agree the bitrate of 192kbps may be too low and contributing to the poor high-frequency reproduction, particularly at low recording levels.

The second clip of me in my editing room is audio directly from the Osmo Pocket but as you say, part of what you hear could be from the video encoding using an HEVC codec to an MP4 compressed format. The clip sounded very sibilant to me in the video as well but that could be because my room absorbs most frequencies as the walls have two foot of Rockwool behind them as well as the acoustic tiles and three feet of Rockwool in the ceiling. As high frequencies are absorbed most and the mic is pointing straight at me it may have picked up more high frequencies than it would in normal use. My room is designed for mixing down not recording. I have a link here to the audio in 16-bit wav format which did sound much better in Cubase.

https://www.dropbox.com/s/usblnx ... est%20Osmo.wav?dl=0 I think that sounds slightly better.

I will go out and about with the adapter next week and see what I can get in normal use. For now, I'm making a molded holder between the Pocket and the mic adapter out of epoxy so it can't get bent or damaged in use.

Hello and good day Ray. Thank you for sharing this link with us today. Great work and thank you for your support.

Did you notice that the left and right channel are not exactly the same!
I used a lavalier microphone directly into the adapter > OP

Small example where I highlighted some of the differences.

NoëI Posted at 3-17 07:46
Did you notice that the left and right channel are not exactly the same!
I used a lavalier microphone directly into the adapter > OP

Hi Noel.
Did you read any of the subtexts on my video?
Yes. I noted the variations which disappeared once audio levels could be increased to below -6dBm. Then variables between channels stopped and effectively became a mono output across both channels, but I could only use a mono mic with enough gain to test that out. Hence it raises more questions than answers. I didn't expand the waveforms as much as your example in the stills shot you provide as that would have been too fast moving to have been picked up accurately in the video.
I don't have a stereo mic with that amount of gain to do further testing. I may rig up a reduced line output for the Pocket later if I can.

No I did not read your text :-/

When I opened your post it was just yours. I downloaded the dropbox file, looked at it and left the house.
After I got back home I did some testing and posted my findings .. I should have refreshed the page 1st I guess :-)

NoëI Posted at 3-17 08:58
No I did not read your text :-/

When I opened your post it was just yours. I downloaded the dropbox file, looked at it and left the house.

For anyone who has this adaptor and wants to work out if it is indeed stereo. The easiest way of checking would be to simply connect a stereo audio source into the adaptor, say a computer, CD, phone output and play some stereo music into it. Just connect via a two channel TRS to TRS 3.5mm cable, the ones usually supplied with most microphones that use a 3.5mm output or any 3.5mm stereo TRS cable or extender.

This will answer the question immediately.

Or better still. Why don't DJI just chime in and give the answer. There's been many posts and people asking this question, it's a real easy one to answer. If DJI have already answered this question in another post, then I'll apologise as I've not seen anything.

Cheers,
Dave.  

David_Harry Posted at 3-17 09:18
For anyone who has this adaptor and wants to work out if it is indeed stereo. The easiest way of checking would be to simply connect a stereo audio source into the adaptor, say a computer, CD, phone output and play some stereo music into it. Just connect via a two channel TRS to TRS 3.5mm cable, the ones usually supplied with most microphones that use a 3.5mm output or any 3.5mm stereo TRS cable or extender.

This will answer the question immediately.

Hi Dave.
I finally gave in and used a line input to the Osmo Pocket.
I only recorded a few seconds of a music track for fear of damaging the pre-amp.
I then tagged the same clip from the original source to the end of the file for comparison.
The clip from the Osmo Pocket comes first and can be downloaded from
https://www.dropbox.com/s/g3ke86 ... eo%20input.wav?dl=0
The Osmo soundtrack is a lot brighter for some reason and gave a reading of +4dBm on the Cubase input meters.

[UPDATED]

I should add I only used the line level at half volume if anyone else wanted to risk sticking a line output into a mic input.

Thanks for sharing

Ray-CubeAce Posted at 3-17 14:11
Hi Dave.
I finally gave in and used a line input to the Osmo Pocket.
I only recorded a few seconds of a music track for fear of damaging the pre-amp.

Hi Ray.

Thanks for the sample.

That's really odd because it's defo stereo, yet your first mic test with stereo didn't appear stereo on the waveform??????  More questions than answers again.

The tone change is massive, although that could be down to driving the input hard. I'd hope that the tone of mics wouldn't be as effected.

I'd still say that the adaptor isn't very good and the whole audio system of the Pocket is seriously lacking. It's as if the audio was a bit of an afterthought, or that's how it appears to me.

Interestingly. I've yet to see and hear any example of the Pocket with an external mic where the sound quality matches its video quality, or anything at all with good external mic quality.

Maybe there was a reason why the adaptor wasn't released with the camera. I wonder how many people would have bought the Pocket with the intention of good external mic recording if they knew what it was going to be like.

Thanks again for the sample.

Cheers,
Dave.

David_Harry Posted at 3-18 10:25
Hi Ray.

Thanks for the sample.

Hi Dave. I can't get my head around the tonal change either. It's like trying to record a vinyl album without an RIAA tonal curved pre-amp.  Nor the fact the two channels are more closely matched in power output.
At least I know it wasn't the mic that caused the sibilance in my room. Maybe it's to counteract the onboard mics, in which case there may be the possibility of a firmware fix.

Ray-CubeAce Posted at 3-18 12:09
Hi Dave. I can't get my head around the tonal change either. It's like trying to record a vinyl album without an RIAA tonal curved pre-amp.  Nor the fact the two channels are more closely matched in power output.
At least I know it wasn't the mic that caused the sibilance in my room. Maybe it's to counteract the onboard mics, in which case there may be the possibility of a firmware fix.

Hi Ray.

Maybe it's all down to specifically matching the Pocket's gain to better match each mic that's connected?

It's a real pain as besides the obvious niggles with the Pocket, of which some may be sortable via FW, it defo has the edge over every other small/action camera as far as pure picture quality is concerned. I've recently got a GP7 and although they're both suited to slightly different things, having compared footage from both the Pocket obviously has the adavantage with regard pure picture quality. It better matches my Sony camcorder footage compared to the GP7.

This is why the whole audio thing is so frustrating. If the whole mic system worked as expected with no fuss, it'd be easier to get past some of the picture/focus stuff. But even with the best picture in the world, bad audio ruins the day more.

Anyway, maybe there's some fixes on their way and maybe local stocking of the adaptor that'd help to bring the price down.


Cheers,
Dave.

David_Harry Posted at 3-18 16:54
Hi Ray.

Maybe it's all down to specifically matching the Pocket's gain to better match each mic that's connected?

Hi Dave.
I'm not sure if this will have an impact on me as much as I haven't had that much need for the Pockets' audio yet as a main source of audio for a project.
As I've always worked with a minimum of two cameras this was supposed to work as an alternative angled view source of material.
I'm not even sure that will work now after trying to put the video together and have seen how difficult it was to sync the two outputs together.  Still, I'm giving it a year to get used to working with as I find it takes me that long to get to grips with workarounds and the foibles of working with any new gear.
The Pocket has some obvious advantages for me such as people, in general, tend to ignore its presence and that's a big bonus for me.
I can set it up easily and quickly, and it has potential due to it's tracking, to get above crowds for clearer unobstructed views where I can't get closer or it's not desirable to get closer for whatever reason.
I can think of things visually I can do with this that just isn't possible with my other camera. The gimble, although not infallible, gets rid of the 'jello' effect I suffer from trying to move around with my DSLR and much lighter and easier to handle than trying to use a glide-cam. Plus I find the unique 3x3 stills and motion-lapse two added extras I don't currently possess.
From my personal, probably unique perspective, the Pocket is still representing good value for money and some organisations have already used some of the stills I've taken with it for their publicity requirements not realising I didn't use my DSLR to take them.  They are also keen to find out what I can do with it using video which I'll find out on my first video assignments with them which will be at the end of April and then again in Mid May.
I will see how it goes but will say this. The adapter is expensive for what it is audio quality wise. To my mind is badly thought out from a design point of view. The way it sticks out at the bottom of the camera where you would hold it represents a possible weak point where damage in use to the Pockets base port seems likely at some point.  An inbuilt socket would even be a bad idea. To my mind, an inboard wireless system would have been a safer bet for something that is supposed to be small and portable, but as a first effort on a new concept, I still have to give it eight out of ten. Just for the new possibilities it opens up and I'm comparing it to much more expensive gear.

Ray.

Ray-CubeAce Posted at 3-18 17:38
Hi Dave.
I'm not sure if this will have an impact on me as much as I haven't had that much need for the Pockets' audio yet as a main source of audio for a project.
As I've always worked with a minimum of two cameras this was supposed to work as an alternative angled view source of material.

Hi Ray.

I totally agree, the Pocket has its place. It's great as a b-cam where audio isn't an issue, even a synced second cam in a two can shoot where its internal audio is easily good enough for a sync mark or clapper. It's also great as an A cam, again where audio isn't an issue, or at least internal audio.

The issue is that many people have seen it as a one stop solution, I know I did. I was expecting to be able to use it for interviews as the smallest thing possible and other situations where dialogue was important.

Having owned and extensively used many small cameras for ultra light weight solutions, I can say with certainty that the Pocket has the ability to capture the best definition pictures of them all, along with the Yi 4K+. I suppose if I went into the whole Pocket thing with only video in mind, then I'd not have been so disappointed.

To be honest, none of these issues to date will stop me from getting another one, it's just that I'll always be quite reticent when picking one up thinking what could have been. DJI could have really cleaned up with the Pocket.

I've no doubt that many lessons have been learnt and that the Osmo Pocket 2 is going to be the killer.

Cheers,
Dave.   

OFF topic but I still use Minidisc too. I have two MZ-1’s.. the original first commercially released minidisc  bought them in 92 for 900 USD a piece!!!! They still work and I still use the original (although recelled) batteries.

On topic: is this indeed a stereo adapter? My anecdotal evidence suggest its not true stereo? Hard to understand from the banter in this thread.

marctronixx Posted at 3-19 07:51
OFF topic but I still use Minidisc too. I have two MZ-1’s.. the original first commercially released minidisc  bought them in 92 for 900 USD a piece!!!! They still work and I still use the original (although recelled) batteries.

On topic: is this indeed a stereo adapter? My anecdotal evidence suggest its not true stereo? Hard to understand from the banter in this thread.

In all honesty, I don't know the answer as to whether it is or is not stereo. It could be that for whatever reason Dji did something to the firmware or software of the pocket to do with the noise canceling and odd positioning of the inboard microphones that in an update could be rectified. It could be that the DAC is not capable or there isn't enough processing power. I just don't know.
What I can say is.
That the external mic adaptor does not seem capable of supplying a normal mic level output.
That whatever is fed into one channel migrates over to the other channel in a slightly altered form but does not seem to be phase canceling the waveform.
That MediaInfo reports it does have two separate AAC audio streams encoded into its output that is backed up by visual inspection in the three programs I can monitor it in.
That the AAC-LC codec used is a mp4a-40-2 which although is recorded at only 192kb/s should be able to produce a high-quality, full spectrum stereo, sound reproduction across both channels.
I am at a loss as to why the Pocket is behaving the way it seems to be behaving.

As for the mini-disc. I have one separates player, a Sony, and one portable recorder, which is Panasonic. Both good and still reliable machines. I was sad to see their general demise but on the plus side, I have more than a few unused discs to use up.
I also had a Zoom H4 digital recorder but lost it or had it stolen during a hectic scheduled filming weekend. Apart from handling noise (Not brilliant for a handheld device), it had a nice clean output I could multitrack on in the field.