Jump to content

Soundfield/Ambisonics delivery workflow


Mobilemike

Recommended Posts

Hey all,

I was curious if any of those members who are experienced with soundfield/Ambisonic recording could share a bit about their delivery workflow. 

Obviously I know each client is different, but in general how have you been asked to do things? 

Some of the questions I had were: As a sound mixer do you just deliver the A-format files and be done with it? Or have you been tasked with converting the files to B-format after recording and delivering those? If so, do you bill for that time? Have you ever been asked to deliver 5.1 or 7.1 (etc, etc) files?

Any other thoughts you have are appreciated! 

Thanks!

-Mike

Link to comment
Share on other sites

Another point is to have the camera array and the mic aligned in a consistent orientation. And I think this needs to have a knowledgable person tracking this through the whole post process as well.

Best to have this conversation in pre. Along with delivery format etc.

Once when I've tried to designate a 'front' to the camera so I can align up the front of the mic, the camera people smiled and said, 'but there is no front.' They just didn't get that I had to align the mic up correctly. And when setting up for the first shot is not the time to try and get people to think things through.

On another shoot I heard they rotated the image without a thought for the sound, so they had to rotate the sound later on in the edit and guess the correct orientation. It was a very busy city scene with a lot going on and I think they spent a lot of time trying to get it right.

Link to comment
Share on other sites

Those camera people are wrong, or just being dicks.  The front is where the image sequence starts--like what does is the viewer looking at when the playback begins.  The goggles/mask will decide that whatever position the viewer is in at that moment is "front" and apply the HRTF in relationship to that.  But there has to be a starting point.  The example of a busy scene with a lot of cross-axis movement is a good one--getting that back into "sync" with the image so that the audio HRTF makes sense without a correctly oriented start point would be tough.

Link to comment
Share on other sites

Hi Mike, I would suggest B format 'WXYZ' to be the easiest delivery for post to use - when I started with soundfields it was just the Mk5 that would have any A format output option, and all the post tools dealt with B format. Things might have changed with the cheaper newer mic now available but it's almost certain I would imagine that post will be working with B format tracks.

DO talk to post however (particularly if you have a problem, or time issues, providing B format). Find out what they're using - you never know, conversion for them might be easy quick and automatic?

Everything soundpod and Phil mention is worth taking into account: I don't have the same personal experience yet in ambisonics for VR and the like, but talking to post (in particular a senior editor or mixer) and then designing a shooting pattern (with the dicks!) may well be quite a challenging job in the end!

Jez

Link to comment
Share on other sites

Awesome, thanks so much for the insight everybody. I figured B-format would be the way to go but I wanted some confirmation so it sounded like I knew what I'm talking about :)

And soundpod and phillip - thanks also for the note about the front of the camera - that definitely seems important and it hadn't occurred to me!

-Mike

Link to comment
Share on other sites

Thanks Mr T

I have an SPS-200 and I believe it only outputs A format

It's not a microphone that I use every day but I have had no complaints

It connect straight into my 644 so ideal for location work, no added boxes

Cheer

mike

Link to comment
Share on other sites

4 hours ago, mikewest said:

Thanks Mr T

I have an SPS-200 and I believe it only outputs A format

It's not a microphone that I use every day but I have had no complaints

It connect straight into my 644 so ideal for location work, no added boxes

Cheer

mike

Hi Mike, you're spared my orig reply with a page freeze, but in short I guess costs were able to be brought down by bypassing both B format and the matrixing into 2 channel that the boxes did - so I meant 'cheaper than a ST450MK2' rather than cheap. And the whole ambisonics thing is now cheaper and more widespread now that expensive hardware, later supplemented by expensive software, is now perfectly replaceable with cheap or even free software. So A format tracks can be easily converted to B and used.

Can't really see, beyond purist archiving, much use for an A format feed, except perhaps recording an instrument going to a single on-axis 'nice' mono track if the ambisonics wasn't working for any reason (phase, studio ambience level) but you wanted the 'calrec velvet'? But I'm sure it was there on the Mk5 box.

I'm more 20th C than 21st with ambisonics, but VR or other installation stuff might just get me back into the phaze craze ... I haven't done any editing or mixing of ambisonics for a while and don't know what recent changes have been brought to surroundzone or Harpex but I can't see why post would be dealing with anything apart from the B format.

I'm sure it's now cheap/free and fast to get from A to B hence offerings of 'direct out' A format mics which a decade or so ago would not have been a cost-saving workflow as soon as post (or even just matrixing down to 2 channel) was factored in.

Damn, you missed all my jokes what with the 'pute freeze and had to suffer the 'considered' reply!

Best, Jez x

Link to comment
Share on other sites

On 6 July 2017 at 4:03 AM, soundpod said:

Another point is to have the camera array and the mic aligned in a consistent orientation. And I think this needs to have a knowledgable person tracking this through the whole post process as well.

Best to have this conversation in pre. Along with delivery format etc.

Once when I've tried to designate a 'front' to the camera so I can align up the front of the mic, the camera people smiled and said, 'but there is no front.' They just didn't get that I had to align the mic up correctly. And when setting up for the first shot is not the time to try and get people to think things through.

On another shoot I heard they rotated the image without a thought for the sound, so they had to rotate the sound later on in the edit and guess the correct orientation. It was a very busy city scene with a lot going on and I think they spent a lot of time trying to get it right.

 

Sorry, site signed out on me again - I hoped the post would miraculously reappear (as it has done many times) but no. I should stop trying to post using iPads which I hate and cannot use!

And that was merged - so makes no sense: post is lost I reckon. Never mind, not useful, just a rant.

About stereo fixed to perspective and vector systems. Another time.

Link to comment
Share on other sites

Good to know Philip. I don't even know what FuMa or Ambix are ... and always thought the A to B process was the celebrated Magic Beans of our Oxford chums (where my ST250 happens to have lived for the last couple of years since my pal used it to record his choir)! Curious actually, since the 'genius' lies just there, the capsules of different mic types are different so so must be the conversions. If XY(+Z) and W were truly coincident 8s and a diffuse omni the maths (of B ) is simple, like MS. Best off to bed now though, up in 4 !

Never retrieved the previous post either.

Jez

Link to comment
Share on other sites

FuMa and Ambix are the 2 most common flavors of B format Ambisonic 4 track audio, with Ambix being the newer and is some cases preferred one.  B format is what the apps like Facebook's Spatializer need to make a 360 mix with  audio that will respond the the viewer's head turns within the VR space.  Most mics I know of make A-format, and so far I've been expected to convert it to B Format as part of the delivery process.

Link to comment
Share on other sites

I will offer to add time to do the B Format conversion, but mostly I end up delivering A Format.  When I do I make notes as much as I can, and send an email stating that the files are A Format and need to be converted.  It doesn't always get through though.

I also audio slate my takes with the mic orientation as a permanent record of how it was recorded cause notes get lost, and metadata can easily get erased somewhere along the process.

I try to align the mic front to the camera default, but I've found it a very hard sell to convince people to keep that orientation static throughout the process.  I usually end up hearing, we'll just put everything where we want it when we get to audio post, and there is some reason to that.  It's helpful if a clap or some sort of slate is used to sync the cameras, as that can be used to orient the mic in post when the image default has been shifted.

Link to comment
Share on other sites

On 6/27/2017 at 11:07 PM, Jan McL said:

Concur. Began my last project by presenting the craft services king with a small jar of homemade raspberry jam. He happened by just as I was contemplating who'd been especially good to us recently. The rest of the shoot I got a cuppa every morning as soon as power was run to his tent. Huge win.

 

On 7/8/2017 at 1:56 AM, The Immoral Mr Teas said:

Hi Mike, you're spared my orig reply with a page freeze, but in short I guess costs were able to be brought down by bypassing both B format and the matrixing into 2 channel that the boxes did - so I meant 'cheaper than a ST450MK2' rather than cheap. And the whole ambisonics thing is now cheaper and more widespread now that expensive hardware, later supplemented by expensive software, is now perfectly replaceable with cheap or even free software. So A format tracks can be easily converted to B and used.

Can't really see, beyond purist archiving, much use for an A format feed, except perhaps recording an instrument going to a single on-axis 'nice' mono track if the ambisonics wasn't working for any reason (phase, studio ambience level) but you wanted the 'calrec velvet'? But I'm sure it was there on the Mk5 box.

I'm more 20th C than 21st with ambisonics, but VR or other installation stuff might just get me back into the phaze craze ... I haven't done any editing or mixing of ambisonics for a while and don't know what recent changes have been brought to surroundzone or Harpex but I can't see why post would be dealing with anything apart from the B format.

I'm sure it's now cheap/free and fast to get from A to B hence offerings of 'direct out' A format mics which a decade or so ago would not have been a cost-saving workflow as soon as post (or even just matrixing down to 2 channel) was factored in.

Damn, you missed all my jokes what with the 'pute freeze and had to suffer the 'considered' reply!

Best, Jez x

Yes but

I have recorded concerts with my Soundfield for mixing (in Peter Jackson's Park Road Post)

with great effect for 5.1 release with stereo options so why B format?

mike

Link to comment
Share on other sites

  • 2 weeks later...

OK, maybe time for a little Ambisonic refresher here: apologies for those who already know, but I sense some real confusion here.

All first-order Ambisonic microphones currently available use what's known as a tetrahedral array of four cardioid-ish capsules. Viewed from the front, those from Soundfield, Sennheiser and Core Sound have capsules arranged to point as follows:

Front Left Up, Front Right Down, Back Left Down, Back Right Up. (Note that there is no actual front capsule - probably best to choose  FLU which is nomally channel 1 in the output designation if you're going to scratch and verbally ident.)

The DPA-4 mic reverses the up and down for each direction and as it's rare as hen's teeth, we'll ignore it for the moment.

The SPS200, the Sennheiser Ambeo and the CoreSound TetraMic all deliver just the four capsule outputs and use software to carry out further processing. WIthout the further processing, these four mic outputs comprise what is known as an A-Format signal and if that's what you deliver to post, without any further details of mic orientation, you're just delivering four mic signals pointing in random directions.

Soundfield, Sennheiser and CoreSound all have software-based tools to convert the A-Format signal to what's known as B-Format, which allows the magic to happen: Effectively, what you get in a B-Format signal is four mathematically defined microphones: an omni reference and three figure-of-eights, effectively pointing front-back, left-right and up-down. Further processing introduces correction filters and, in the case of the CoreSound, capsule corrections as well, specific to each particular microphone. Only CoreSound does this, Soundfield and Sennheiser rely on a generic corrections, reckoning that their capsule tolerances are pretty tight, although some have found this not to be the case.  Soundfield's SurroundZone 2 (http://www.soundfield.com/products/surroundzone2) is free and deals with signals both in A & B-Format for its own microphones: CoreSound provides two free VST programs to deal with the A-B format conversion and capusle correction and then another for B-Format to various other output formats. (http://www.core-sound.com/TetraMic/1.php) Sennheiser's Ambeo effort currently only does the A-B conversion for its own microphone, but see below for an alternative.(http://en-uk.sennheiser.com/ambeo-blueprints-downloads)

So, if you're delivering A-Format, you need to tell post which manufacturer the mic is from and the orientation at recording, i.e. front-fire, end-fire, and if inverted.

The higher-priced Soundfields have the A to B-Format conversion carried out in hardware and the ability to correct for microphone orientation at the time of recording, plus a few other niceties, like a high-pass filter, gain control and, in the case of the ST450, the ability to output B-Format and a derived stereo output at the same time, with pattern and width control for the stereo output on the front panel, without affecting the B-Format output, which can be a time-saver if full surround is not required.

if you're delivering B-Format and have corrected for mic orientation at recording time, this should probably be noted for information.

B-Format signals can then be transcoded to any channel count, from mono to full surround with height, with the right software. Some of this software is free, as in the Soundfield and CoreSound offerings, along with programs from the likes of Daniel Courville, (http://www.radio.uqam.ca/ambisonic/) Angelo Farino (http://pcfarina.eng.unipr.it/Ambisonics.htm) and Bruce WIggins, (https://www.brucewiggins.co.uk/?page_id=78); some is not, like Dave McGriffy's premium VVEncode software for the TetraMic (https://www.vvaudio.com/products) and, the most comprehensive and powerful program, Svein Berge's Harpex-X, which can handle both flavours of Soundfield and the Sennhesier Ambeo and gives you all very many possibilities for transcoding. It's expensive, but it's worth it, in my opinion. Harpex-X also deals with FuMa (Furse-Malham) and AmbiX (or ACN or SND3, or whatever you want to call it) channel designation which, instead of the FuMa sequence of WXYZ, uses a numbered designation, the first four of which, 0,1,2 & 3 correspond to FuMa's WYZX, although there's some level jiggery pokery as well. The reasons for the change are to do with the ease of defining higher-order signals (more channels, better localisation.)  http://www.harpex.net/

If you want to do really astonishing stuff in post with 1st order Ambisonics, the you should invest in Richard Furse's Blue Ripple software packages. http://www.blueripplesound.com/product-listings/pro-audio

I hope this clarifies things a bit... (or maybe not.)

Oh, and if you want Ambisonics on the cheap, only without the height element, the Zoom H2n with the V2 Firmware does a pretty good job.

Regards,

John

Edit: I left out Matthias Kronlachner's excellent ambiX software - it's here:

http://www.matthiaskronlachner.com/?p=2015

 

 

 

 

Link to comment
Share on other sites

  • 2 months later...
On 7/23/2017 at 12:50 PM, soundmanjohn said:

OK, maybe time for a little Ambisonic refresher here: apologies for those who already know, but I sense some real confusion here.

All first-order Ambisonic microphones currently available use what's known as a tetrahedral array of four cardioid-ish capsules. Viewed from the front, those from Soundfield, Sennheiser and Core Sound have capsules arranged to point as follows:

Front Left Up, Front Right Down, Back Left Down, Back Right Up. (Note that there is no actual front capsule - probably best to choose  FLU which is nomally channel 1 in the output designation if you're going to scratch and verbally ident.)

The DPA-4 mic reverses the up and down for each direction and as it's rare as hen's teeth, we'll ignore it for the moment.

The SPS200, the Sennheiser Ambeo and the CoreSound TetraMic all deliver just the four capsule outputs and use software to carry out further processing. WIthout the further processing, these four mic outputs comprise what is known as an A-Format signal and if that's what you deliver to post, without any further details of mic orientation, you're just delivering four mic signals pointing in random directions.

Soundfield, Sennheiser and CoreSound all have software-based tools to convert the A-Format signal to what's known as B-Format, which allows the magic to happen: Effectively, what you get in a B-Format signal is four mathematically defined microphones: an omni reference and three figure-of-eights, effectively pointing front-back, left-right and up-down. Further processing introduces correction filters and, in the case of the CoreSound, capsule corrections as well, specific to each particular microphone. Only CoreSound does this, Soundfield and Sennheiser rely on a generic corrections, reckoning that their capsule tolerances are pretty tight, although some have found this not to be the case.  Soundfield's SurroundZone 2 (http://www.soundfield.com/products/surroundzone2) is free and deals with signals both in A & B-Format for its own microphones: CoreSound provides two free VST programs to deal with the A-B format conversion and capusle correction and then another for B-Format to various other output formats. (http://www.core-sound.com/TetraMic/1.php) Sennheiser's Ambeo effort currently only does the A-B conversion for its own microphone, but see below for an alternative.(http://en-uk.sennheiser.com/ambeo-blueprints-downloads)

So, if you're delivering A-Format, you need to tell post which manufacturer the mic is from and the orientation at recording, i.e. front-fire, end-fire, and if inverted.

The higher-priced Soundfields have the A to B-Format conversion carried out in hardware and the ability to correct for microphone orientation at the time of recording, plus a few other niceties, like a high-pass filter, gain control and, in the case of the ST450, the ability to output B-Format and a derived stereo output at the same time, with pattern and width control for the stereo output on the front panel, without affecting the B-Format output, which can be a time-saver if full surround is not required.

if you're delivering B-Format and have corrected for mic orientation at recording time, this should probably be noted for information.

B-Format signals can then be transcoded to any channel count, from mono to full surround with height, with the right software. Some of this software is free, as in the Soundfield and CoreSound offerings, along with programs from the likes of Daniel Courville, (http://www.radio.uqam.ca/ambisonic/) Angelo Farino (http://pcfarina.eng.unipr.it/Ambisonics.htm) and Bruce WIggins, (https://www.brucewiggins.co.uk/?page_id=78); some is not, like Dave McGriffy's premium VVEncode software for the TetraMic (https://www.vvaudio.com/products) and, the most comprehensive and powerful program, Svein Berge's Harpex-X, which can handle both flavours of Soundfield and the Sennhesier Ambeo and gives you all very many possibilities for transcoding. It's expensive, but it's worth it, in my opinion. Harpex-X also deals with FuMa (Furse-Malham) and AmbiX (or ACN or SND3, or whatever you want to call it) channel designation which, instead of the FuMa sequence of WXYZ, uses a numbered designation, the first four of which, 0,1,2 & 3 correspond to FuMa's WYZX, although there's some level jiggery pokery as well. The reasons for the change are to do with the ease of defining higher-order signals (more channels, better localisation.)  http://www.harpex.net/

If you want to do really astonishing stuff in post with 1st order Ambisonics, the you should invest in Richard Furse's Blue Ripple software packages. http://www.blueripplesound.com/product-listings/pro-audio

I hope this clarifies things a bit... (or maybe not.)

Oh, and if you want Ambisonics on the cheap, only without the height element, the Zoom H2n with the V2 Firmware does a pretty good job.

Regards,

John

Edit: I left out Matthias Kronlachner's excellent ambiX software - it's here:

http://www.matthiaskronlachner.com/?p=2015

 

 

 

 

 

Link to comment
Share on other sites

Be aware that on a Zoom F8, with the latest software, you can record A format on 4 tracks, and B format on the other 4. Likewise on the Sonosax SX-R4+ (it is coming very, very, soon). The B format conversion occurs within the machine at the time of recording. just saying.......

 

Laters..... sb

Link to comment
Share on other sites

  • 2 weeks later...

And, joy of joys, Len Moskowitz has announced the OctoMic - a second order Ambisonic mic with eight capsules (the ninth channel is derived

in software) which will make life even more complicated, but the audio will be better with a larger sweet-spot and better localisation. I'll be getting mine in the spring of next year, along with an SD MixPre-10.

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...