Hi.
I was writting this as a response to Ray CubeAce on another thread but thought it best to write a new thread so as to not side track and take the other thread off topic.
Hi Ray.
Yes there are certain issues with the processing and encoding of the audio. I've done a number of tests with the recordings and am still undecided on how the audio is processed.
It would appear that the main mic is the main mic and the bottom mic is used to detect certain noise or frequencies and then use a combination of both to try and cancel out unwanted noises/sounds. This would usually be wind, LF and such.
This cancelling can also be used to mask motor noise. If you try blocking either mic, in certain conditions you may hear the motors in the recorded/encoded audio.
In such an arrangement there can only be a mono encoding, unless processing is applied to manufacture spacial information.
The odd thing is that the encoded audio sometimes appears to be stereo. Or at least the undesirable elements are, such as wind etc. While dialogue is mainly fixed to the centre. I say mainly because if you centre both channels and reverse the phase of one of them, the centre information, dialogue, is only partially cancelled and other elements are partially cancelled as would be expected in a stereo recording.
This odd outcome could happen if the encoded audio was true stereo but I'm quite positive that the Pocket's internal mic system is mono. I'll also use the mic placements as an indication that it's not stereo because the positions would be wrong. Plus, and without an appreciably distance or baffle for seperation and time delay, the stereo field isn't so wide on mics that close together.
There are a number of potential processing side effects that can result in what I've noticed. The main one that springs to mind is the AAC encoding. The AAC encoding itself also works on a number of techniques, including psychoacoustic and phase related ones.
During the AAC encoding it's quite normal, depending on the encoder complexity and bitrate/bit depth, for miss-tracking of certain stereo sourced information within the two channels of a stereo encoding. The reason why I'm talking about stereo encoding while I think the mic system is mono, is because the final encode appears to be mono encoded across both channels of a stereo encode and not a duplicated mono source across a dual/split mono output.
If this were the case then it could be possible, again based on the quality of the encode system, that while the source may be mono the encoder could in theory apply artefacts as it may do with a stereo source if it is treating the source as stereo and encoding stereo. Thereby resulting in miss-tracking between the channels on the output encode. If this were the case then that could also account for certain frequency and phase differences/anomalies between the two channels of the final encode.
Add to that the moderate bitrate used and the potential compounding of issues is raised.
There's also things such as the A/D, pre-amp, gain, dynamics processing to also take into account.
Anyway, I could be totally wrong, maybe there's something similar to a M/S arrangement being employed within the mic system. I'm only a producer/engineer and not a technician, so there's a load of stuff that's way over my head that I may have completely missed the point on.
Maybe someone with more technical experience/information, maybe from DJI, could add some information to this thread. I'm quite sure there's a number of Pocket user's who would be interested in such things, I know I would. I did try sourcing extra info about the Pocket's audio system but couldn't find anything, maybe I wasn't looking in the right places but it does appear that this type of info is a bit thin on the ground.
I'll try and get the time to make a video showing the tests that I've done with my NLE and DAW.
BTW, Ray. Is the CubeAce in your forum name anything to do with Cubase?
Cheers,
Dave.
If anyone is interested in microphone audio stuff here's a few videos about some of the things I've botched together for my recording setup. You'll have to listen on headphones as some of this stuff is very wide stereo or something close to binaural. This is all mostly preliminary tests that I do before making a proper YouTube video basically testing the concept, so these are quite rough.
Have you ever had an early stereo LP from the States? If so you may have noticed you can't play such discs successfully on modern systems.
I ask this because I think Dji may have done something similar electronically with their mic encoding.
Modern styli work in a vinyl record by moving within the groove and the magnetic pickups are either side representing an X pattern. The earliest western electric method used a + pattern with the left being vertical and the right-hand sounds coming from the horizontal movement of the stylus. The result of playing such a disc is that it sounds very similar in a reverse sort of way to the Osmo Pocket. In as much as such an album played on modern equipment loses all the central portion of the image as it phase cancels and the resulting side channels have both the same content; almost. It is not the normal type of phase cancelation I'm used to hearing but similar to those old LP examples but with the 'side' information being canceled and the center sounds coming through. It does sound as if it has space around it but I've run it through several stereo expanders and whatever it is, it isn't even in frequency response.
So I am curious if for no other reason as to try to find out what is going on and whether this adapter is passive or not. The pocket base connector is certainly capable of supplying power.
I've used CubeAce as part of my forum name for some time since the early Cubase forum days. My Flickr and other accounts have similar forum names. I worked in London recording studios during the 70s but eventually needed to actually earn some real money for a living
Have you ever had an early stereo LP from the States? If so you may have noticed you can't play such discs successfully on modern systems.
Hi Ray.
I thought I knew how records worked but obviously only how we done it in Europe.
That's really interesting what you said about that American system, I never knew about that thanks for mentioning it.
Yes, there's a whole bunch of questions about the Pocket's audio. That centring of dialogue and then the odd spacial thing, like wind distortion sounding stereo, can be caused by the encoder. I also thought there could be something similar to M/S going on. Whatever it is and as we've both said, it is some kind of phase or frequency cancellation thing.
Anyway, stuff that boring stuff. I've been using Cubase since before it was Cubase, I'm a proper FanBoy. I did a lot of music production in the early 90s with the first Cubase on Atari, had the Midex expander with the T/C and a bunch of other stuff including an external hard drive system, I think it connected via the DMA? port on the ST.
I used to run it with 5 fully loaded 32Meg Akai S1000/S1100s, basically using the samplers to run and sync audio for Cubase. When Cubase went to Audio and XT my head exploded. All my producer friends were using ProTools, oh I can still remember all the arguments
When Steinberg released Nuendo I had one of the first ones in Europe, my business partner at the time done a lot of film score. I've a box somewhere with the original dongle, serial or parallel I can't remeber, with a very low serial number. I actually posted a BBC documentary on it, the first UK transmission of anything that was posted on Nuendo. That said and although they were quite similar, I'm a Cubase person not Nuendo.
I got more into video/film about 20 years back so my use of Cubase as a pure music tool dropped significantly. I produced an album on it last year but nothing since, so I'm still only on version 8.
Anyway, this is a DJI forum not the Steinberg one, so I'll leave it there. But very glad to meet another Cubase user.
Cheers,
Dave.
BTW. Those arguments with my ProTools friends have never ended As a proof of concept I done this session about 3 or 4 years back and demo'd this cover on a really small Celeron based mini PC to prove how well Cubase and VST utilises the CPU compared to PT and that mess of a native plugin architecture that it uses
I thought I knew how records worked but obviously only how we done it in Europe.
Hi Dave. The Western Electric recording system didn't last long as the main inherent defect was the stylus was prone to groove jumping on heavy transients due to the way the groove was cut. I only have two album examples and one very old stylus cartridge assembly capable of playing them back. I won't even go into armed forces 16" 78s then
I have the original dongle as well for the PC version of Cubase VST, which is when I started to delve into digital recording with Win98. Prior to that, I was still using analog with 10" tape machines. I have over 400 reels sitting at home. Mainly BASF and AGFA, Ampex and 3M, with a revamped Revox A77 1/2 track 7.5-15 ips machine. I haven't been pro for some years leaving to get a job I could actually get a mortgage with.
Nice demo.
The Pro Track fandom simply came from it being the best new kid on the block at the time. It worked well with Neve desks if I remember rightly. Steinberg's clever gamble was making VST open source. Many a manufacturer could learn something there. I always felt the early versions of Cubase were tighter on their MIDI timings than later releases but maybe not as versatile.
Getting back to the Osmo Pocket.
Really this is my secondary camera for video use. If need be I have other means of getting the audio but I wanted something that can hardly be noticed for wandering and taking alternative angled shots to mix to and fro from the main footage. It's not just the portability for me but the way people at present at least, ignore it even when you get relatively close. Very useful when covering events. Our event season starts late April so I want to get some practice in well before the season kicks in.
My son had been using the Go-Pro Hero Black until he saw my unit. Now he has one because he does a lot of low light work and the Osmo's output is much cleaner and the stills side of it makes it much more of a useful tool for him as well. He also prefers the image stabilization over the Go-Pro but has held on to the Go-Pro for tougher situations.
The Osmo's audio bitrate has me puzzled but then again some video formats don't go much above that and still sound OK unless you get into surround sound. I'm more concerned I can't have any real control over volume or a way to monitor it. The auto sound control is the best I've heard in recent times with little to no audible pumping but so far I haven't really tried a very hard dynamic range test yet and the noise canceling may still be hiding something. This is why I have to find out before the event season is upon me.
I will post recordings as soon as I practically can.
lvl.4
