Does anyone ever use Audio recorded in camera?

[Stephen Pruitt]

And, for those wondering if less-than-sterling location audio can in some way be saved or, at the very least, improved, there is iZotope's RX-3, which is undoubtedly the most amazing $1000 solution for audio capture problems on the planet. I can't tell you how many times that has saved our tails on our latest film. This thing is GOLDEN.

In cases where things are so bad that not even RX-3 can help, you can always do what they do in Hollywood. That's right: Foley and ADR.



Stephen

 

[Patrick Tresch]

I think Marc Wielage will love iZotope's RX-3!!! LOL!!!

I use the Epic onboard audio for 2 purpose :

- another way to check synch & Pluraleyes
- give the director a way of checking the take with sound instatly after recording.

For the rest, I love my sound guy who ever he is.

Pat

 

[Stephen Pruitt]

Say what you will, Mr. Tresch, but Emmy and Academy Award winning sound teams use RX-3 (and similar tools) every single day. . . when they aren't looping every single bit of dialog, that is.



Stephen

 

[Patrick Tresch]

Shurely a great tool! But I know this will please Marc!

Have a nice day Mr. Pruitt!

Greatings from Lausanne.

Patrick

PS: Marc probably allready used RX-3... horses for courses

 

[Mike P.]

Is RX-3 better than Cedar? Or are they for completely different purposes (if so, what)?

 

[Stephen Pruitt]

I have never used Cedar, but they are essentially the same sort of tool. Cedar is much more expensive, and may be worth the additional investment for badly damaged tracks, but RX-3 is just golden for most of our typical audio issues.

I just used it this morning, as a matter of fact.



Stephen

 

[Will Keir]

What's this setup look like with the audio mounted to the camera?

We did comprehensive tests, just to see if using audio directly into the camera was as good or some where close to external recording.

Using the same mics for both options.

Played the camera recordings and external captured recordings on the day side by side in a cinema studio , The Audio engineer literally couldn't tell the difference.

Hence why we now record to camera everything. (VIA A Mixer of course)

At the end of the day, it all comes down to the MIC used, and the person that's feeding correct signals to the camera.

PS,
That said... we used a small cheap Sennheiser G2 lapel mic the other day, Amazing , booming, crisp results. ?.... go figure.

 

[Chris Workman]

Off topic, but I couldn't find a thread dealing with this. I was pulled onto a multi-cam project (2 Scarlets and an Epic) and asked to pump out proxies. They didn't have access to the timecode jambox so I told them audio would be the next best way to sync multicam footage. All 3 cameras were supposedly recording scratch audio, but when I pulled the footage there was no audio to be found. This is my first time handling on-camera audio so I'm in the dark here. First question: Does on-camera audio get baked into the R3D file or is it a separate file within the mag? Also, I was not told that the entire project would be shot in either 48fps of 60fps. At this frame rate, is the camera still able to capture audio? I would think that these higher frame rates would throw things off. Thanks for helping out a newb!

 

[Elsie N]

Off topic, but I couldn't find a thread dealing with this. I was pulled onto a multi-cam project (2 Scarlets and an Epic) and asked to pump out proxies. They didn't have access to the timecode jambox so I told them audio would be the next best way to sync multicam footage. All 3 cameras were supposedly recording scratch audio, but when I pulled the footage there was no audio to be found. This is my first time handling on-camera audio so I'm in the dark here. First question: Does on-camera audio get baked into the R3D file or is it a separate file within the mag? Also, I was not told that the entire project would be shot in either 48fps of 60fps. At this frame rate, is the camera still able to capture audio? I would think that these higher frame rates would throw things off. Thanks for helping out a newb!

Higher (or lower) framerates than the timebase are considered varispeed by the camera and thus, no sound recording.

 

[Will Keir]

I believe Audio is baked into the R3D. I have seen lots of RDM folders, RDC folders, never saw any separate audio folders.

Off topic, but I couldn't find a thread dealing with this. I was pulled onto a multi-cam project (2 Scarlets and an Epic) and asked to pump out proxies. They didn't have access to the timecode jambox so I told them audio would be the next best way to sync multicam footage. All 3 cameras were supposedly recording scratch audio, but when I pulled the footage there was no audio to be found. This is my first time handling on-camera audio so I'm in the dark here. First question: Does on-camera audio get baked into the R3D file or is it a separate file within the mag? Also, I was not told that the entire project would be shot in either 48fps of 60fps. At this frame rate, is the camera still able to capture audio? I would think that these higher frame rates would throw things off. Thanks for helping out a newb!

 

[Ryan Sauve]

Higher (or lower) framerates than the timebase are considered varispeed by the camera and thus, no sound recording.

This

 

[Alan Maxwell]

Hence why we now record to camera everything. (VIA A Mixer of course)
.

Mark can you mention which model sound mixer you use ?

 

[Mike P.]

Off topic, but I couldn't find a thread dealing with this. I was pulled onto a multi-cam project (2 Scarlets and an Epic) and asked to pump out proxies. They didn't have access to the timecode jambox so I told them audio would be the next best way to sync multicam footage. All 3 cameras were supposedly recording scratch audio, but when I pulled the footage there was no audio to be found. This is my first time handling on-camera audio so I'm in the dark here. First question: Does on-camera audio get baked into the R3D file or is it a separate file within the mag? Also, I was not told that the entire project would be shot in either 48fps of 60fps. At this frame rate, is the camera still able to capture audio? I would think that these higher frame rates would throw things off. Thanks for helping out a newb!

Sound is baked into the .r3ds.

If you shot at 48 or 60fps (likely 48 at 3k on the Scarlets), the audio would only be recorded if your timebase was also set to 48 or 60... Sometimes RCX defaults to a channel that the mic wasn't plugged into though, so it could be a matter of setting the audio to the active channel in RCXp.

 

[Christoffer Glans]

I mostly spend a day just syncing and organizing material when doing an edit for something. I started a thread about why we still record sound separately with an always buggy timelock. There are very few productions where everything went so smooth that there were no problems in post. As an editor I'm sick and tired of being handed material that wasn't capture entirely correctly. It's not that the errors aren't easily fixed, it's that it demands a detective job to find the errors so nothing gets lost when auto-syncing.

I'm all for recording directly into the camera. If you can backup recording externally, but still have a high fidelity sound going into the camera which can be used as the main audio, then that is a MUCH better choice and reduce editing time A LOT.
Then you just need to transcode and you're set to go.

I'm still a strong believer of trying to figure out a working wireless transfer of sound between the sound operator and the camera so that he/she will still work as always, but instead of only getting sound on the external recorder, we transfer high quality sound to the camera. Maybe this is one for the redlink developers?

 

[Tim Morten]

Just buy a Zaxcom or Lectro. WiFi is problematic for latency-critical data like audio. Zaxcom and Lectro have this solved by using their own digital protocols.

 

[Nick Fury]

I endorse the Zaxcom audio input, via their proprietary Zaxnet technology. Great for scratch track and the ERX units, are very light-on board batteries go all day and will synch Timecode, if using a non Dragon camera (at the moment), just fine, throughout the same ERX box. Can attach the ERX, with velcro, wherever or with an aftermarket metal enclosure.

Thank you very much

Fury

 

[Marc Wielage]

I mostly spend a day just syncing and organizing material when doing an edit for something. I started a thread about why we still record sound separately with an always buggy timelock. There are very few productions where everything went so smooth that there were no problems in post. As an editor I'm sick and tired of being handed material that wasn't capture entirely correctly. It's not that the errors aren't easily fixed, it's that it demands a detective job to find the errors so nothing gets lost when auto-syncing.

You can solve the problem of "material that wasn't captured correctly" just by hiring the right people. As I often say, this is a people problem, not a technology problem.

I'm all for recording directly into the camera. If you can backup recording externally, but still have a high fidelity sound going into the camera which can be used as the main audio, then that is a MUCH better choice and reduce editing time A LOT. Then you just need to transcode and you're set to go.

Hire an assistant editor at a modest fee, and have them sync everything up and organize the footage for you. Good assistants can literally do this in dailies in 1 minute per shot. If you have 200 shots per day, that's only a little over 3 hours -- not the end of the world. Great assistant editors pay for themselves in terms of making backups, organizing scenes, checking sound sync, filling in all the metadata, and otherwise doing all the essential but non-creative work that the editor need not worry about.

I'm still a strong believer of trying to figure out a working wireless transfer of sound between the sound operator and the camera so that he/she will still work as always, but instead of only getting sound on the external recorder, we transfer high quality sound to the camera. Maybe this is one for the redlink developers?

As I've said before: what if you have many different sound channels? What if you need more sound channels than the Red (or any other camera) can handle? What if every actor on the set has a wireless mic? What if the level of every actor is different? What if the sound levels have to constantly be monitored and changed as the scene progresses? What if you need to shoot varispeed, but still record sound? What if you need to record wild tracks without picture? All of these are standard procedures for many projects I've worked on over the years.

I worked on a sound-heavy feature last year where we had upwards of 10 channels of wireless mics on at least half the scenes in the picture. The rest of it rarely had fewer than 5 or 6. Where are those going to go in a 4-channel Red workflow?

You need to sync up all the sound -- original multi-tracks with picture -- and ask the on-set sound mixer to give you one "mix" track to use for editing and dailies. At least this will give you a rough mix that will work for the edit. Later on, you export an OMF or an XML for the sound editor to use in a simultaneous sound finish. Any experienced post supervisor will tell you this is very standard for scripted productions all over the world.

BTW, this was for a little $400K indie without a very large sound budget, but we still managed to work it all out in a reasonable compromise for the production.

 

[Antony Newman]

Interconnecting cables + processing stages add noise to the low level signal from a microphone when in a analogue form.

Worst Case (Maximum Noise + Distortion)
===================
+ A Lavalier microphone creates an analogue signal
+ A local preamplifier (on person) amplifies this + transmits it on an analogue radio frequency
+ A receiver converts the radio signal back to the original signal
+ Preamp before the camera boosts / mixes the signal
+ Camera converts the signal to digital using its 16bit ADC (analogue to digital converter)

Best case (Minimum Noise)
=================
+ A Lavalier microphone creates an analogue signal + A local preamplifier amplifies + ADC (on person) converts this to this to a 24bit signal
+ This can either be transmitted as a digital signal down a cable or
A miniscule module buffers 1GB transmits and
Optionally broadcasts this out on a 802.11ac (or other) and
Optionally broadcasts to a client (ie Red Dragon camera) a synchronisation stream.


THOUGHTS:

24 bit vs 16bit : The advantage of this is you would not need a soundcheck monitor to check that volume is correct,
it would ISO on an R3D.

Option of a miniscule 24bit ADC recorder could mean an 'invisible' recorder (and very low power)

Syncing digital signals into the EPIC : Would give the option (even in Varicam / hi speed mode), for the external devices to sync everything they have recorded into the Epic (thereby keeping All audio and video in one place). No need to pay a person to do what electronics & software can do better.

Less cables.

Mixing could be done in post - OR there could be a foolcontrol control, to record suggested volume levels for each channel into the Camera.

AJ

 

[Christoffer Glans]

You can solve the problem of "material that wasn't captured correctly" just by hiring the right people. As I often say, this is a people problem, not a technology problem.

This is not true. First off, the technology is sometimes bad, the Red One is notoriously bad with keeping timecodes locked on all productions I've worked on. Doesn't matter that the timecode is synced several times during the day, it always drift into unworkable material.
Second, I'm tired of hearing this "just hire the right people". All productions aren't able to hire the best of the best, some productions don't have those budgets and even if they do get the right people and the right technology, there are times during productions were there just aren't time to do everything right, where you just need to get that camera up and capture before the light is gone. I seriously doubt that all of your productions have been flawless because you have the "right people". Seriously.

Hire an assistant editor at a modest fee, and have them sync everything up and organize the footage for you. Good assistants can literally do this in dailies in 1 minute per shot. If you have 200 shots per day, that's only a little over 3 hours -- not the end of the world. Great assistant editors pay for themselves in terms of making backups, organizing scenes, checking sound sync, filling in all the metadata, and otherwise doing all the essential by non-creative work that the editor need not worry about.

Once again you think all productions have unlimited budgets? What about music videos? Not all music videos are made for high end artists with millions of dollars to spend on high end directors and productions. The world is not this simple and hiring an assistant editor at a "modest fee" is not going to generate what you describe because the "modest paid" assistant editor is probably not the most knowledgable. I've had assistant editors just like that and they don't know much of what they are doing because "modest paid" means they are new.

As I've said before: what if you have many different sound channels? What if you need more sound channels than the Red (or any other camera) can handle? What if every actor on the set has a wireless mic? What if the level of every actor is different? What if the sound levels have to constantly be monitored and changed as the scene progresses? What if you need to shoot varispeed, but still record sound? What if you need to record wild tracks without picture? All of these are standard procedures for many projects I've worked on over the years.

I worked on a sound-heavy feature last year where we had upwards of 10 channels of wireless mics on at least half the scenes in the picture. The rest of it rarely had fewer than 5 or 6. Where are those going to go in a 4-channel Red workflow?

You need to sync up all the sound -- original multi-tracks with picture -- and ask the on-set sound mixer to give you one "mix" track to use for editing and dailies. At least this will give you a rough mix that will work for the edit. Later on, you export an OMF or an XML for the sound editor to use in a simultaneous sound finish. Any experienced post supervisor will tell you this is very standard for scripted productions all over the world.

BTW, this was for a little $400K indie without a very large sound budget, but we still managed to work it all out in a reasonable compromise for the production.

Why don't you read what I wrote? I described having the sound operator working just as they always do with an external recorder, but in addition, the main audio is recorded to the camera. By main audio I mean dialogue, the audio editors actually need. We never need those additional tracks, those are for the sound designers.

Have you ever tried to edit a feature film with 10 tracks of audio? Not gonna happen. So keeping the extra tracks on the external recorders while transferring the main tracks to the camera will give the editor what they need. If there are more then 4 tracks in my edit, I take the rest out, seriously, it's a mess to work with a feature edit with so many tracks, even if I bake them together the audio will sound like crap with so many channels at the same time. Hence, I don't need those channels, it's the sound designer who does and that means there is no reason for the main audio not going into the camera.


I just think that people are afraid of changing their workflow. Having audio on the footage directly is a huge advantage for post production and even for watching dailies. You can talk all you want about "hiring people", but the less you need to spend on unnecessary things the more you can spend on other parts of the production. And it's not all about the money or people. I've had assistant editors that worked their asses of trying to manage files shot on a Red One with drifting audio. Even though they are fast with the syncing, it might require detective work to secure that all material and footage is managed. A feature I'm editing now was postponed because the clean up and syncing was a mess... and it wasn't the people that made it be like that.

I really don't get why so many are against improving the workflow, it's not like we reduce the tracks, the quality or need for sound operators. It's about improving the workflow and after working for 8 years as an editor I've encountered too many faults and errors to think there's anything good about keep doing things the way we are doing them.

 

[Marc Wielage]

Have you ever tried to edit a feature film with 10 tracks of audio? Not gonna happen. So keeping the extra tracks on the external recorders while transferring the main tracks to the camera will give the editor what they need. If there are more then 4 tracks in my edit, I take the rest out, seriously, it's a mess to work with a feature edit with so many tracks, even if I bake them together the audio will sound like crap with so many channels at the same time. Hence, I don't need those channels, it's the sound designer who does and that means there is no reason for the main audio not going into the camera.

I will ignore a lot of your remarks because you clearly have no clue what you're talking about. There is a bare minimum of budget you need to pull off a successful shoot. If you can't hit that budget, then you have no choice but to cut back your expectations. Saying "we can't afford to hire good people because we no money" begs the answer, "well, get more money." Or make a less-complicated film. But if you have three people on set gaffing, two people running the camera, an electrician, a set, and a grip truck, you can find a way to hire a sound guy to run audio. If you don't, then you're going to wind up getting the sound you deserve.

As to number of tracks for the edit: we always, always advise that the on-set sound mixer provided a mono mix track for the editor to use for syncing and dailies. 90% of most major scripted films (small and large) just cut with the mono track. Period. Import all the tracks, then turn all the unused tracks off and leave the mono selected. That's all you have to do. The mono mix should be enough to use for cutting the entire picture in most cases. Later on, when you export an OMF or an AAF for Pro Tools or a similar higher-end sound editing/mixing program, the re-recording mixer and the sound editor can use all the other tracks to reassemble, recut, and remix the movie as required using alternate takes, wireless lavs, wild lines, ADR, foley, and whatever else is needed. It's not rocket science, but you do need good people and time.

Keeping the Red camera jammed to timecode can be problematic, but this is a Red problem that's currently being worked out. I know of few other cameras that have these issues. And in truth, I've worked on many dozens of Red projects where we had zero timecode problems, once the initial set up was done (strapping an external SB-T timecode source and running the camera in external jam). The last thing I ever want to do is harass the camera crew with more wires, more boxes, or more paraphernalia on the camera. I think the more sound remains an external function to what the camera does, the better it is for the DP and the camera operator. Both stay totally out of the other's way, and each is primed to do exactly what they're most expert at doing.

I just think that people are afraid of changing their workflow. Having audio on the footage directly is a huge advantage for post production and even for watching dailies. You can talk all you want about "hiring people", but the less you need to spend on unnecessary things the more you can spend on other parts of the production.

To me, this kind of thing is spoken by very ignorant people who have not worked in the business a long time. It's sad to me that many neophyte filmmakers wrongly assume that either there are problems there that only they can solve, or (even more foolishly) they believe that nobody has ever tried a faster or better way to do something before.

Workflows are many, varied, and complex, but sound is one of the easiest, simplest things about it. The techniques are known and have worked for many years. Again: none of this needs to be expensive or involved. There are affordable ways to do it, provided you're informed, prepared, have a rock-solid schedule in mind, and (most importantly) are very mindful of managing your expectations. If you can only afford to edit sound and mix for a week, so be it. It's still going to be better than some moron trying to do all the sound cutting in Final Cut Pro or Premiere, or an idiot who tried to record the entire movie with one mic going directly to camera.