What's the purpose of a digital recorder?

Also, higher sample rates ARE supported on Blue ray, so that is an argument for recording higher sample rates. But as another poster pointed out, any sample rate above 44.1khz is out of human hearing range, and really only effects the harmonic content of the audio below 20 kHz, the highest frequency humans can hear.

Sample rate conversion does create some noise, but not anything you can generally hear. If your on a multi million $$$ production, ya, your probably going to record 192khz sound, and be a purist.

Actually no, BIG budget films shoot the dialog at 48/24. Some people can hear above 20K, especially young girls, most adult males don't make it over 15K. If you spent much time with an iPod drop that down to 12K.

And the ONLY place 44.1 comes in to film work is in the DolbyDigital 35mm optical digital stream.

BlueRay is not really in the mix since low budget folks can't afford the license rates.
 
Last edited:
That's a typo Scott? Did you mean 44.1/24?

Actually what I meant was "Actually no, Big budget films shoot dialog at 48/24".

A comma really changes things. Yours is also correct. Sometimes in your mind the meaning is clear and the fingers just don't understand ;~)
 
underrepresentedunderrepresented

underrepresentedunderrepresented

Not sure why picture shooting at a high frame rate would make a difference, unless your thinking of slowing down production to match a slow-mo shot. That generally doesn't work anyway. Slowed down sounds don't generally sound the way most people think the should sound so you usually build fake slow-mo, well really nearly always.

Actually, yes, there have been quite a few instances where the recording of a production track at 96/24 was beneficial during the sound editorial process that required a direct match to a 60fps/120fps/300fps slow-mo sequence. 3 months ago I was at a boxing gym with the crew shooting our main actor jabbing the speed bags at ultra speeds. Well, since there were no lines of dialogue and I was recording at 48/24 on several takes already, I decided to up the sample to 96/24 so post could have the extra room to play with. 2 months later, turned out that I received an e-mail indirectly from the post editor letting me know that despite the rather large and resource-hungry files I provided, the difference between processing the 48 and 96 tracks were greatly underrepresented from the usual workflow. Pretty high up with an Emmy or two.

Generally speaking, it's almost always from the ground up in post and not taken from production. It's a give and take on what we're recording in the first place.

And while I'm kind of with you on the high feq sounds, there is the catch that to be recorded you need a mic and front end electronics that can "hear" those high freq sounds. Most mics only barely have any pickup above 20k. If you have wind protection on the mic your loosing even more. So in practical terms I don't think it's worth it 99% of the time. IF you are going to do a bunch of slow down on it you do get a "smoother" sound, even if most of the "high freq" you recorded is just electronic noise from the mic. We did tests on a very quiet Foley stage with high end mics with nothing in front of them and those recordings did pitch smoother and you could here a bit more "shimmer". But not much.
Out side with less premium mics and wind protection... I know it's the new thing to do and folks with big reps are recording at high sample rates but I haven't seen/ heard anything to convince me it's worth it except in a few specialty situations.

And the Pain in the Butt factor gets significant when you know your doing post. If I'm not doing post? It's a flip of a switch, easy and I can be the coolest high sample rate guy around. But I don't like being the picture on the post editors dart board. And I have had to sample rate convert a huge stack of production sound that some bozo did at 96 (with a crappy mics, bad technique, lousy levels, into some hand held pocket recorder). Nothing brings joy like having a drive full of high sample rate turds to try and polish.

This now lies in between technical specifications and real-life trial and error testing methodology. While several big reps are going the route of high sample rate capture >88.2/96 and others sticking with 44.1/48 because they couldn't hear a difference, the results are quite startling when a sequence of effects are processed more than two times within a clip. An A/B comparision between the same single unprocessed/processed clip is easy to deceive and once the equation becomes more complicated that involves a heavy-action sequence this amount of "smoothness" becomes very apparent to the ear like a stack of dominoes ready to fall.

Even with microphones that are technically spec'ed at 20kHz top, the margin of error between capturing true 96 is a toss between a 0.01-100% depending on the quality of electronics within the machine or equipment. The better-designed equipments that one uses is linearly dependent on the detail of recording one gets only if the recording technique is the same. It's exactly the same argument in the camera industry of shooting at a higher framerate to achieve for a smoother visual display when played within the same duration of 1 second. When you look at frame by frame, there are bound to be several frames (2-3) that look similarly or exactly alike at the cost of double the storage and rendering resources needed. Same concept behind a jar half-full of water that starts out rampant with splashes progressively spinning smoother into a vortex. Just a matter of how precise one wants/needs at the end, within reason.

I had a photographer friend who used to get really pissed off at folks who would ask him what camera takes good pictures. He would always say "the camera does not take the pictures, the photographer does!".

Valid point.
 
Alan,

You should probably mention what type of shows those were. The reason is that films like a doc or ENG or even low budget with limited post is, A) far more likely to use production beyond dialog and B) doesn't have the work flow issues that a big picture does.

I your first instance, your really talking about recording FX not dialog, though it was in production. And maybe I would have heard those hits and used them. I would have listened to them certainly but unless they were just perfect for the scene as cut I would have built that sequence anyway so... Generally with things like that the "real" sound is good as a guide and maybe as a base layer but it doesn't hear the detail the camera sees so you still need to add a lot.

In the fast turn around world of course your time is very limited so you will end up using more production.

And this type of thing was one of my exceptions, you knew it was going to be slowed down. And it was FX, or maybe "Live Foley?" ;~).


Your second point is probably valid, I haven't heard that test, but it's also kind of moot in a real sound post. You are never going to hear that A/B in a mix. Those subtle bits are going to mostly get swallowed in the rest of the sound track.

I do wish folks would stop bringing up 44.1. There is really no place for 44.1 in a film and video work flow. I know what you are saying but some newcomer is going to read that and think it's OK to roll at 44.1 and it will "save space". Then some poor smuck is going to have to convert it all and re-sync.

And I think you are probably right. One place that shooting 96 probably makes sense is for high frame rate non slomo films(stop with the 44.1/ 88.2 stuff, this isn't CD land nobody in film wants to see those sample rates. They are only used in music recording and only because it's the CD sample rate).

I'm not sure how many of those are going to be coming out though. It's kind of the flip side to the 24P craze. High frame rate films apparently have a very 3D feel about them. To match that hyper clarity sound post may well have to use a different work flow.


But we are also talking really on a few different levels. What shakes out in the world where $30 is a "low budget" and what is practical in the world where a few $100,000 is a high budget are not really closely related. For most of the members of this list they are definitely in the 100K is high world and I will be pretty firm that for them dialog at anything other than 48 is a waste of resources. Every minute someone spends dealing with high sample rate sound (that just will not be heard in playback) is time that could have gone into working on the film. I have seen folks ask about recording at 96 (because it will sound better) for shorts they are going to post to YouTube. That is a pure waste of resources.

Shooting dialog on Cameron's next 1/2 billion dollar epic is an entirely different matter.
 
Hey Scott,

I brought up and mentioned about the higher sample rate because sometimes it does save post the extra time, and other times making it extremely difficult and frustrating for the entire editorial team. For folks who are somewhat familiar of this workflow, be advised that the post supervisor should always give you the go ahead before committing to such a task.

The project I was referring to was a short film and we had the freedom to record at what we deemed was appropriate for that particular scene or day. It wasn't exactly live Foley per se since we did marry it to picture on set.

The last thing I do want to bring back up is the 44.1 problem. I've gotten several projects where the session was already prep'd, BUT it started in 44.1! Not only did it come from one group of people but from different states and country. It's daunting to hear why 44.1 is even used in the film world, but it never hurts to mention time to time for reminders. We know both you and I and others on this list strictly use 48, but there are also many others who would simply stick with 44.1 because that's what they know.

The most important thing to remember is to use a separate recorder at all times if possible. It's not the end of the world without one.
 
Does any real sound professional want to trust their audio is being recorded properly to a camera crew's eyes all day? Until camera systems offer some sort of high quality integration that gives the audio professional confirmation and control of the audio being recorded to the camera, I don't see double system leaving us. Especially with a new digital camera coming out every 6 months, in my experience, sound guys would rather not be worried about the latest possible bug in camera x's firmware that makes you think you're recording audio when you're not, etc.

Having said that, in camera audio is great for quick dailies when shooting digital on a small show where you may not have someone with the time to sync dailies. Also, shooting double system is the way a lot of film crews are used to shooting, it's also the most reliable & high quality "wireless" sound you can get. It also makes the 2nd AC feel more important. :D
 
Alan,

Couple of things. One despite not wanting to talk like 44.1 is an OK choice (It isn't for anybody reading this list) it was used a bunch in film post when digital was just coming in. There used to be two big camps, the 44.1 folks and the 48 folks. The 44.1 folks reasoning was it took less space (harddrives were small and expensive so this was a BIG issue), you can't hear the difference (a big argument then), and DD is going to end up at 44.1 anyway.

The catch in the US was that only self imported or very pricey DAT's could record at 44.1. The RAIA had forced DAT manufacturers to go to 48 so that users couldn't digitally transfer CD's to DAT. Sample rate largely didn't make any difference because everything was going to be mixed analog. So some facilities were all 44.1 and some were 48. Since a lot of SFX libraries were on CD you could do a digital transfer if you worked at 44.1, but then you had to do a SR conversion if the production was at 48 (though usually DAT was only the backup and the "real" production was on 1/4" and was then digitized at whatever the dialog folks worked at. So you had shows where FX is at one sample rate and dialog is on another, Foley is on 2" analog tape and ADR is on?????
It was not uncommon to have multiple DAW's also. PT, Sonic Solutions and WaveFrame were often used on the same show because FX folks were heavy PT users, Foley and ADR tended to like WaveFrame and music folks liked Sonic because it did 24 bit before everyone else.

That is all history now and there is no good reason to not be at 48. I have seen sessions at 44.1 come out of FCP not sure why or how? And sessions that were started by music folks who love 44.1 because of CD's. Actually music usually comes in (on low budget projects) at 44.1.

So you and I know the ugly truth but for new filmmakers You want to stick to 48.

As to high sample rates. As I said I don't go there often, and I am not generally happy to be surprised in post with a bunch of high sample rate stuff to convert. BUT if the project is planned around it and post either knows it's coming or actually wants it for some reason that is a totally different thing.

I don't like talking about it too much because new folks tend to latch on to that kind of thing as a "why not" with no thought to the "why" part. Every once in awhile we get someone who doesn't know much about it but has decided they want to record analog production. Again there could be good reasons to do it but they haven't generally gone there. They just decided it would be the thing to do and have not got a legitimate reason for doing it.

So my worry is some new filmmaker reads this discussion and decides they will get "better quality" recording if they go 192/32bit float. Technically that should be "better" but what a god awful nightmare for post. And you just know they are doing it on some low end pocket recorder with a $100 shotgun and no shock mount. Or a Azden wireless set up!

I'm not worried about you going high sample rate. I know you would talk to me beforehand (assuming post is on then) and the recordings would actually be good in the first place.
 
I shouldn't stress to much about this all.

TV-shows/news/reality programs = sound directly on cam... nobody will complain, it pays good, you can pay your bills and everybody is happy.
It's much faster, easier, no syncing, nobody hears it...

For film and because of the capability of multitrack-recording and because cinema speakers should be better then a 5 dollar TV... it's better to record in higher quality on
a seperated recorder. (and also if it's on Alexa, Red, 5D,7D 60D)
 
Back
Top