Audio bracketing

[daniel]

Are there any applications/plug-ins assisting post production with the use of recordings which deployed some form of audio bracketing in their acquisition?

Kind Regards,

Dan.

[Wandering Ear]

What help are you looking for?  Use one track, and if there are issues in a spot pull it from the other track?  Are you looking for a plugin that automatically marries two tracks recorded at different levels?

[daniel]

1 hour ago, Wandering Ear said:

What help are you looking for?  Use one track, and if there are issues in a spot pull it from the other track?  Are you looking for a plugin that automatically marries two tracks recorded at different levels?

On the 1 hand I'm thinking of something like a 'bag drop' situation with 1 of the recorders with an audio bracketing option or similar. On the other hand we have 2 companies making portable recorders with extended dynamic range with dual AD converters and a clever algorithm so I'm wondering has anyone developed a piece of PP software with a similar type of algorithm. We see this in the world of photography, Ie. In-camera HDR and post production HDR.

[John Blankenship]

If the person operating sound gear is knowledgeable and experienced enough to fully understand gain staging as a competent professional should, even a 16-bit recorder is sufficient for capturing high quality dialog tracks, and 24-bit is a luxury.  While dynamic-range-expanding technology may be nice to have in certain situations, its biggest benefit is to those less skilled.

Even as cameras may have HDR, and many shooters now use ultra-hi-def cameras so they can compose the image in post, I would much rather have a cinematographer who knows how to light and compose a shot in camera.  If the goal is an artistic image that works with all other factors to elevate the story telling process, I have significantly more faith having someone with vision, skill, and experience at the helm.

[mikewest]

Used to be done if feeding an identical mix to a 2 channel video camera.

Set one channel at normal level and the other at say 10dB lower

 

mike

[Wandering Ear]

On 3/2/2016 at 2:35 PM, daniel said:

On the 1 hand I'm thinking of something like a 'bag drop' situation with 1 of the recorders with an audio bracketing option or similar. On the other hand we have 2 companies making portable recorders with extended dynamic range with dual AD converters and a clever algorithm so I'm wondering has anyone developed a piece of PP software with a similar type of algorithm. We see this in the world of photography, Ie. In-camera HDR and post production HDR.

Comparing technologies like Neverclip, or techniques like recording the same mic to two tracks at different gain stages, while similar, isn't the same as HDR photography, and I don't see how a merging plugin is necessary.  What problem in post is this trying to fix?

For extended dynamic range recorders, no processing is necessary (except possibly a little gain), and there is no need to reverse engineer what the recorder is doing in post.  It's simply a piece of technology to capture a larger dynamic range.  Once it's captured, it's used as is.

As for recording on two tracks at different gain levels, it seems this is often done with one track as a safety, so if the hotter track doesn't clip, I don't see any advantage to trying to merge the two tracks, while I do see the possibility of affecting the audio quality with the merge.  I suppose it's possible for someone to write an algorithm that treats these two tracks the way an extended dynamic range recorder switches between converters, but a dialog editor can effectively make these choices quite quickly, and as John points out, if the tracks are recorded properly, there likely won't be much need to do so.

 

[daniel]

On this occasion there is currently no problem post is trying try to fix :-)

I agree with all the comments posted so far regarding properly recorded tracks.

The (hypothetical) context for this enquiry was:

~ A 'bag drop' situation, Eg. a car with no room for a sound operator but enough room for a multitrack recorder with an audio bracketing function (or similar).

~ Long takes with lots of dynamic range.

~ The type of production unlikely to have a dialogue editor.

 

If something like 'merging' bracketed channels is useful enough to put into recorders (Zaxcom and Sonosax) why not have something similar (like 'merging' bracketed tracks) in PP?

[John Blankenship]

6 hours ago, daniel said:

On this occasion there is currently no problem post is trying try to fix :-)

I agree with all the comments posted so far regarding properly recorded tracks.

The (hypothetical) context for this enquiry was:

~ A 'bag drop' situation, Eg. a car with no room for a sound operator but enough room for a multitrack recorder with an audio bracketing function (or similar).

~ Long takes with lots of dynamic range.

~ The type of production unlikely to have a dialogue editor.

 

If something like 'merging' bracketed channels is useful enough to put into recorders (Zaxcom and Sonosax) why not have something similar (like 'merging' bracketed tracks) in PP?

So you're saying you want the manufacturers that serve this industry to develop methods for productions to not hire professional sound mixers? 

That actually helps no one in the long run -- manufacturers lose their market -- sound mixers lose their jobs -- the production loses quality.

 

[daniel]

1 hour ago, John Blankenship said:

So you're saying you want the manufacturers that serve this industry to develop methods for productions to not hire professional sound mixers? 

That actually helps no one in the long run -- manufacturers lose their market -- sound mixers lose their jobs -- the production loses quality.

 

No, I'm not saying that, you are. I wasn't trying to be contentious with this topic but as an aggregate I'm reading 'merging' software wouldn't work very well but will do sound people out of jobs, which seems a bit contradictory. I was just trying to explore the possibility such a thing might make certain recording techniques a little easier and effective. And if anyone is 'misguided' enough to develop such a piece of software I imagine it'll be a software company like Apple, Adobe or Avid (Ie. companies unbothered by the existence location sound people) rather than our hardware manufacturers - although the possibility of licensing the use of an already developed algorithm to a software house may offer a hard to resist revenue stream for said IP holder.

[Tom Visser]

I was editing a video that I shot myself and though I'm not regularly an editor or post sound professional, I dabble occasionally - even with a few mics, I found that it was cumbersome to deal with even more than a single track, I had recorded ISOs and Mx tracks in the field, meaning I had 4 tracks for 2 mics - basically a safe and a split broadcast target level simultaneously.  At the end of the day, it would have been simpler to deliver to the picture editor (me - a guy who would rather have a god sounding single mono sound file) that sounded good.  In this respect, I think delivering more tracks is not the ideal answer, but more advanced mix tools (the machine) and smart decisions by the mixer (person) in the field to streamline the process for the picture cutters.

What I do personally is utilize my Zaxcom wireless and take advantage of the built in limited / dynamics to give me -9dBfs of fast limiting (peaks above will occasionally pass) and feed via AES to my recorder, and with +6dB of "gain in hand" to push the levels on the mix buss.  If I'm working TV (split mono) or fast web delivery type of projects (mono), the mix will be aggressive and the to hit the -12 to -8dBfs range so that the deliverable sits right in the broadcast comfort zone.  I'd say that this setup sounds good to me and even allows me to be relatively "careless" with my gain staging.  If it's hotter than ideal, the dynamics of the signal chain still sort of keeps things sorted out but will still sound "good" for 90% of the picture / sound editors out there.  When doing doc or narrative, something that I'm more assured will have a more competent post workflow, I'll ease up on the levels a bit and keep the ISO closer to the -20 to -18dBfs area and have a mix that is a bit looser and a few dB below the target broadcast levels.  Even with bag drops, haven't had an issue with this setup.  When I first got my Zaxcom wireless I only used a few dB of limiting and obviously it would be detrimental to over limit a signal - but after awhile, once I realized the context in which the wireless was being used - I would never ideally be recording full scale audio, that is - stopped thinking of the Zaxcom as a recorder and more as a piece of the system, then I started making use of the limiter more and enjoying the convenience that it provided.

With respect to audio bracketing, I could see something like this being implemented, but in such a way that doesn't impact the way that data is handed over - meaning more confusing audio tracks.  What might be smarter is to have a 32 bit internal mix and record buss, but then the deliverable media can have some sort of smart finishing dynamics process - whether it is simple 100% transparent lookahead limiting or some sort of finalizing style of compression / leveling / limiting like you would see with a typical post tool that for all intents and purposes, looks like the same 24 bit 48KHz files we have been delivering all along.  I think some of this tech is already being implemented in one way or another by some of our manufacturers in the community.

With 24 bit or higher recording, digital connections, and high performance analog circuits, the need to really push levels is somewhat of an antiquated idea from a technical perspective, but still important as a convenience factor and making it "look like we are doing our job" (and don't underestimate this need).  I even have one editor friend who actually rates mixers' competence by how hot the levels are on his NLE.  This makes me smile, but ultimately I completely understand that I can't really argue with his logic, because at the end it is his preference and idiosyncrasies get to make the final call.

[daniel]

2 hours ago, Tom Visser said:

I was editing a video that I shot myself and though I'm not regularly an editor or post sound professional, I dabble occasionally - even with a few mics, I found that it was cumbersome to deal with even more than a single track, I had recorded ISOs and Mx tracks in the field, meaning I had 4 tracks for 2 mics - basically a safe and a split broadcast target level simultaneously.  At the end of the day, it would have been simpler to deliver to the picture editor (me - a guy who would rather have a god sounding single mono sound file) that sounded good.  In this respect, I think delivering more tracks is not the ideal answer, but more advanced mix tools (the machine) and smart decisions by the mixer (person) in the field to streamline the process for the picture cutters.

What I do personally is utilize my Zaxcom wireless and take advantage of the built in limited / dynamics to give me -9dBfs of fast limiting (peaks above will occasionally pass) and feed via AES to my recorder, and with +6dB of "gain in hand" to push the levels on the mix buss.  If I'm working TV (split mono) or fast web delivery type of projects (mono), the mix will be aggressive and the to hit the -12 to -8dBfs range so that the deliverable sits right in the broadcast comfort zone.  I'd say that this setup sounds good to me and even allows me to be relatively "careless" with my gain staging.  If it's hotter than ideal, the dynamics of the signal chain still sort of keeps things sorted out but will still sound "good" for 90% of the picture / sound editors out there.  When doing doc or narrative, something that I'm more assured will have a more competent post workflow, I'll ease up on the levels a bit and keep the ISO closer to the -20 to -18dBfs area and have a mix that is a bit looser and a few dB below the target broadcast levels.  Even with bag drops, haven't had an issue with this setup.  When I first got my Zaxcom wireless I only used a few dB of limiting and obviously it would be detrimental to over limit a signal - but after awhile, once I realized the context in which the wireless was being used - I would never ideally be recording full scale audio, that is - stopped thinking of the Zaxcom as a recorder and more as a piece of the system, then I started making use of the limiter more and enjoying the convenience that it provided.

With respect to audio bracketing, I could see something like this being implemented, but in such a way that doesn't impact the way that data is handed over - meaning more confusing audio tracks.  What might be smarter is to have a 32 bit internal mix and record buss, but then the deliverable media can have some sort of smart finishing dynamics process - whether it is simple 100% transparent lookahead limiting or some sort of finalizing style of compression / leveling / limiting like you would see with a typical post tool that for all intents and purposes, looks like the same 24 bit 48KHz files we have been delivering all along.  I think some of this tech is already being implemented in one way or another by some of our manufacturers in the community.

With 24 bit or higher recording, digital connections, and high performance analog circuits, the need to really push levels is somewhat of an antiquated idea from a technical perspective, but still important as a convenience factor and making it "look like we are doing our job" (and don't underestimate this need).  I even have one editor friend who actually rates mixers' competence by how hot the levels are on his NLE.  This makes me smile, but ultimately I completely understand that I can't really argue with his logic, because at the end it is his preference and idiosyncrasies get to make the final call.

Thanks Tom, a very interesting read.

[Wandering Ear]

In HDR photography, the goal is to record a dynamic range that exceeds the capabilities of the recorder (camera).  Modern digital audio recorders can easily record a dynamic range greater than our source material, I.e. dialog, so I don't see where it's be desirable to merge tracks in post.  I do lots of the type of recordings you describe without issue using tools at hand.

If such a plugin existed as you described, it may do something for a very small and specific set of circumstances, but it would just be one way to approach it, and there are other effective ways available already.  I don't foresee this having any impact on wether sound mixers get hired or not.  The possible dynamic range of your recordings ranks pretty low in the list of things necessary to get good sound, and even lower on the list of what it takes to get hired.

There may be an application for what you describe, but I don't see it.

[syncsound]

On 3/5/2016 at 8:15 AM, John Blankenship said:

 

On 3/8/2016 at 7:24 PM, Wandering Ear said:

In HDR photography, the goal is to record a dynamic range that exceeds the capabilities of the recorder (camera).  Modern digital audio recorders can easily record a dynamic range greater than our source material, I.e. dialog, 

 

I would add that that same dynamic range exceeds the capabilities of most D-to-A converters and human hearing, as well. Even 20 bits is @ 120dB of range, which is enormous.

[daniel]

There are 2 recorder manufacturers who spent valuable resources developing systems which use dual A-D converters on a single channel and proprietary algorithms to give extended dynamic range to a recorded track and therefore a safety net for operators who may have to contend with audio elements other than dialogue (eg a door closing or some other high energy event during a take), but the consensus here is to do effectively the same thing (albeit split across 2 boxes and departments) is a waste of time. Reading through some of the c. 400 posts across 2 'topics' about 'never clip' it seems there are a lot of folks who believe extended dynamic range is a useful thing, although some state it isn't. Obviously for owners of Zaxcom's and Sonosax's latest recorders this idea is irrelevant but for owners of other recorders I think this technique could be nicely facilitated with a plug-in for DAWs and NLEs because amongst others, there are OMBers using 1 of the several modest recorders available with the audio bracketing function for whom a 'merge audio' or 'HDR audio' tool would be useful in an edit. Not least as these recorders don't necessarily have the high grade analogue audio components, including limiters, available on more expensive kit but they do have a lot of recording tracks. Saying this, I still have an SQN for some jobs because I prefer the limiters to that on my SD mixer - they are 'smoother' and more 'robust', which for me effectively allows a more intuitive way of managing compression because I don't need to dive into a menu.

[Wandering Ear]

It's not that I think dynamic range isn't important, it is.  I think the recorder is a better place to advance that dynamic range.  Possibly an auto combine plugin could save some time in post, but as it is post can already make full use of the dynamic range of two tracks recorded at different levels by making a few quick edits.  Even with high dynamic range tracks recorded on the 2 recorders you mention, post will have to process that track to get the dialog into a volume range suitable for playback.  In essence, it's important to record the dynamic range to maximize the quality, but at least in dialog recording, the dynamics have to be tamed quite a bit in order to fit in the final mix.