Sonny Posted June 29, 2010 Report Posted June 29, 2010 As I sit here with a potent drink, I thought I'd share the events of today and see if this is the norm...That is, if my still trembling hands will allow me to type it. About three weeks ago I wrapped on a movie. It's the third I've done for this particular director. This is by far the biggest production we've done, the first that didn't feel like a student film, and the first for either of us on the RED. I'm also doing the post sound, so today I went to his house to take a look at the first 20 minutes of rough cut. Now, on location I recorded to my 744T, but I also sent a feed to the camera for him to edit with. I had heard that RED's audio was...buggy...but for the sake of ease in post, I secretly hoped that my "master" recordings could sit on a hard drive somewhere...unheard and unloved while the "scratch track" took all of the glory. Although I didn't monitor camera audio on set (obviously favoring my recorder), I knew the levels were good, and it sounded fine going out the door to the camera. I had also heard nothing but good things being reported back after each night's session with dailies. Cut to today. The big fight scene happens about 30 seconds into the film...and the first thing I notice is huge distortion any time a character raises her voice. My initial reaction is to look at the director and say "What the hell was that?" To which he replied, "Huh...I was hoping you could tell me." It happened consistently throughout the viewing, during which my color went from a healthy pink-ish to...well...the word ashen got thrown around a lot. During my two hour drive home, the director and I spent much of it on the phone trying to figure out what had happened. Neither one of us noticed any problems on set, and missing distortion of this magnitude would have been severely out of character for both of us. Still, we couldn't figure out why the camera tracks would be so terrible now...especially when they had passed mustard at dailies. I could feel an ulcer form and start growing as I sped toward my master recordings...hoping that my professional integrity would be spared. Upon arrival, I booted up and tracked down the most egregiously offensive shots. Tentatively, I loaded each one and pressed play....SUCCESS! Each track played back as smoothly and beautifully as I remembered them being on set! I called the director with the good news and he had some more intrigue for me. Thinking that perhaps the culprit was the low-res transcoding process that he's using for his off-line edit, he went back to listen to the original RED files...Distortion. He unplugged his speakers and used the ones on the computer...Distortion. He plugged in his headphones...Even the most quiet moments had some weird mechanical distortion. Fortunately, we have the masters, so everything is groovy. We're still unsure of the underlying problem. Again, I had heard bad things about recording audio on the RED, but really? This is terrible! Is this normal? Has anyone else encountered this problem on set, or am I an idiot (well...a bigger idiot than I've ever acknowledged)? I feel pretty stupid right now for doubting myself that much, but it's tough when you're facing such awful sound and have no way of knowing how bad the problem is. Sorry for the rambling. Thanks in advance for any wisdom. Quote
RPSharman Posted June 29, 2010 Report Posted June 29, 2010 Can you give us any more information about what you used to set up the RED? Line level, wireless, tone alignment, padded cables, and that sort of thing. Can you also tell us who was watching and listening to dailies (perhaps they simply didn't notice at first view) and how the dailies' audio was produced? Perhaps the issue your director is having is in the import and/or playback of the files on his system. Has he tried to play them on something else? If not properly set up, the RED (like any other piece of recording gear) can produce poor results. Robert Quote
Sonny Posted June 29, 2010 Author Report Posted June 29, 2010 I used padded cables to feed it. Sent one channel (the mix) at line level into two inputs. I set tone at -20 db on one channel and -30 on the other. The director and the data wrangler watched dailies, and I can't imagine either one would have missed these problems. The import is likely the problem here. I don't know for sure where he was watching dailies, I suspect he watched them pre-import. I honestly have no idea what software he used for that, other than it's a RED specific data management program. On set, drives and cards were imported onto a laptop with that software. Today we were watching the session he has set up in FCP. If he was watching the dailies on that laptop...or on his computer through that software, we may have found the issue. Quote
Eric Toline Posted June 29, 2010 Report Posted June 29, 2010 What build version was the RED? Versions earlier than 16 or 17 were notorious for bad audio. Eric Quote
Philip Perkins Posted June 29, 2010 Report Posted June 29, 2010 You haven't mentioned if you were checking or monitoring the Red's audio all the time or even occasionally as you went. All video cameras have far less audio headroom than a recorder like your 744--what might seem like good levels going from your mixer to the 744 could easily be overloading the inputs of a video camera. The Red audio meter encourages you to set your ref level at a place I find is MUCH too hot even for normal production sound, let alone an action scene--I recommend that your ref level into the camera be no more than half it's audio scale. From the sounds of it you hit the camera too hard. That's not great, but you have the master audio so they are covered, and hopefully were planning to replace the camera audio w/ your SD tracks. I did the audio post on an indie feature lately that laid to rest any lingering doubts I might have had about the difference between video camera audio and audio from a pro audio recorder (in this case a 744). The difference was very evident even in normal-level dialog, with both versions recorded at good levels. The problems we used to have w/ RED audio were mostly having it disappear in transfer...not become distorted. I don't see how that can happen unless they did a bad analog transfer or the editor rendered the audio in some really mangelous way. Philip Perkins Quote
Sonny Posted June 29, 2010 Author Report Posted June 29, 2010 Didn't monitor on set. Because I'd heard bad things about RED audio, we were planning on using the 744 tracks. I was only feeding the camera for scratch track. It was only my laziness (love of efficiency if that sounds better?) that was allowing me to hope that we wouldn't have to go through the re-syncing process later. The confusion stems from the fact that the camera audio checked out during dailies. It wasn't until he started editing that the distortion appeared. I have to think that it's a file transfer or FCP issue. Besides, I hit the camera with tone at -20 and -30. I'll defer to your better judgment, but I'm having a hard time imagining that being too hot for the RED to handle. Sounds like I'll have to set it to peak out around there in the future. Quote
studiomprd Posted June 29, 2010 Report Posted June 29, 2010 " it sounded fine going out the door to the camera. " so ?? sorry, but anything and everything can happens once it leaves that door... problems with audio on RED have been tirelessly discussed, here and elsewhere; numerous authoritive folks have firmly advised trearing RED like a film camera (hey, your production used it because it is so much like film)... that means double system, which you did; you fed them a track (or 2) for dialies, too... also good. now finish it like film, that is double system. the audio capabilities and never ending, constantly changing foibles of recording audio on various versions of RED have been endlessly discussed and documented here, and elsewhere. Knowledgeable, experienced professionals have consistently advised double system sound with RED's. all the 'blame finding' discussions about builds, and settlings, and connections is just wasting time now... if the Camera operator got your good audio, but did not make usable recordings, that is not your problem... if the DIT was not able to obtain usable audio files from the media,, that is not your problem... if you recorded good audio, you did your job well... " Sounds like I'll have to set it to peak out around there in the future. " or not... Quote
Philip Perkins Posted June 29, 2010 Report Posted June 29, 2010 Didn't monitor on set. Because I'd heard bad things about RED audio, we were planning on using the 744 tracks. I was only feeding the camera for scratch track. It was only my laziness (love of efficiency if that sounds better?) that was allowing me to hope that we wouldn't have to go through the re-syncing process later. The confusion stems from the fact that the camera audio checked out during dailies. It wasn't until he started editing that the distortion appeared. I have to think that it's a file transfer or FCP issue. Besides, I hit the camera with tone at -20 and -30. I'll defer to your better judgment, but I'm having a hard time imagining that being too hot for the RED to handle. Sounds like I'll have to set it to peak out around there in the future. If you set your ref level to the mark where the RED seems to indicate it goes, the tracks will be, in my experience, too hot. There may be other factors at work, but it would be very unusual for audio that was recorded clean to acquire distortion in the process of the sort of digital file transfers that are done in filmmaking. There isn't anything mysterious about how files are ingested into Final Cut (or etc), it's just a file copy, basically. The editor might have inadvertently rendered the audio in some bizarre fashion, perhaps through an EQ that was doing a huge boost or something like that. Do you have a feel for this guy's skill level? There has been enough of this kind of problem with the Red's audio for it to have acquired a bad reputation and you did the right thing in any case by double recording. Philip Perkins Quote
Sonny Posted June 30, 2010 Author Report Posted June 30, 2010 I'm pretty confident in his skill. That being said, this was our first RED experience and I know there are some workflow considerations that come with that. Based on what I'm hearing though, I'm guessing that I just hit the camera too hard. I'm still perplexed as to how it could be missed during dailies, but I don't think I'm going to find the answer to that one unless I can catch it happening again sometime. At the end of the day, my recordings are good, so this is all just curiosity. Sorry if I've wasted anyone's time here. I understand that this was a relatively minor problem given the fact that I recorded dual system. If it's related to sound, though, I'd like to understand what's going on whether it's directly related to me or not. I searched the forums and didn't find anything directly related to the issue I was having. Thanks to those willing to indulge my curiosity. Quote
Philip Perkins Posted June 30, 2010 Report Posted June 30, 2010 I'm pretty confident in his skill. That being said, this was our first RED experience and I know there are some workflow considerations that come with that. Based on what I'm hearing though, I'm guessing that I just hit the camera too hard. I'm still perplexed as to how it could be missed during dailies, but I don't think I'm going to find the answer to that one unless I can catch it happening again sometime. At the end of the day, my recordings are good, so this is all just curiosity. Sorry if I've wasted anyone's time here. I understand that this was a relatively minor problem given the fact that I recorded dual system. If it's related to sound, though, I'd like to understand what's going on whether it's directly related to me or not. I searched the forums and didn't find anything directly related to the issue I was having. Thanks to those willing to indulge my curiosity. No worries, and it is NOT a minor problem. We need to be concerned with any recorded audio that we have anything to do with that a client hears. They often don't make fine distinctions between "scratch track" and master track" (anymore), so one's rep is always on the line. I work closely w/ DITs, video assist people, dailies viewing rooms and editors in this regard--I want it all to sound good (and be in sync). Philip Perkins Quote
Philter Posted June 30, 2010 Report Posted June 30, 2010 Oh! So this is where everyone went... Tuned out years ago from RAMPS, a Buddy pointed me here. Though this has been run a bunch of times here I'm sure, it's all new to me. Yeah, the RED. I'm doing an el cheapo MOW starting next week on one. No budget for double system.(will record as well on a little non-TC recorder because...) I did a check last week on it to be sure. This is the first time this particular cam has been used for real. I was taking the adapters out of their plastic wrap to connect. I can't tell you the exact build number on it though I'm told it is current... and my check's in the mail... I used my SD442 for the test. Tone (0dBu) on it lined up exactly at -20dBfs on the RED.(proving a recent build?) Too easy right? I watched levels react on the RED's meters as if nothing was amiss. Peaks seemed identical to the SD. My first clue something may be up was the headphone o/p. I was warned that's not a good judge so I didn't over react when indeed it was horrible. (not an exaggeration - digital clipping, etc.) But playback in the camera and in the editing suite, though looking very healthy (-4dBfs max) were badly distorted. Not digital clipping. Electronics being over driven. We checked two places(edit suite and a post house) because I couldn't believe the levels were bad. After playing with some pads in my kit, a 12dB pad in line with the RED i/p cleaned it right up and seemed to give a safe amount of headroom. A 7dB pad didn't give enough save headroom. Now those two pads were tested for their rating between the SD and an Audio Developments mixer. Not into the RedCam, so the exact rating could very well be somewhat different. Also, from what I read in their manual, apparently the new adapter cables still have a 4dB pad within. I believe the earlier builds used a higher pad within. But I don't have direct experience with that. Also from the manual, they advise that -2dBfs is the ceiling. Not zero. And go on to clarify that there is 18dB over reference(-20dBfs) before going over. Odd. As an aside, with the pads in line the Red's meters reacted nearly as high as the SD's meters. Meaning to me that the electronics must have been compressing the signal as it distorted. I guess that's not surprising. Phil Brouwer Quote
studiomprd Posted June 30, 2010 Report Posted June 30, 2010 " the electronics must have been compressing the signal as it distorted. " BINGO! we have another winner here! the audio of the RED is not up to snuff, at least for real professionals! I hear this distortion frequently on Monday's from students who shot their projects over the weekends, and then find out their audio is NG on Monday morning, when they play it back. They promise me that the meter never ever went into the red, and was usually only in the green. In their case they make a common mistake (well, two, actually!); they have sent a line level signal into a mic level input; by adjusting the input levels really low, the meter never goes into the red! the distortion is massive (unreparable) clipping of the input signal! (the second error was not listening to the headphone out of the camcorder or they would have clearly noticed this problem! The non- industry standard spec's audio circuitry on the RED camcorders is not professional quality and this the repeated advice to treat the RED as part of a double system configuration. " No budget for double system. " they will get what they pay for! guess that defines "semi-Pro" cash your checks promptly! Glad you are doing thorough workflow tests!! Quote
Philter Posted June 30, 2010 Report Posted June 30, 2010 When quoted like that, I'd like to be sure I'm clear. I wasn't thinking that there is compression within the input of the RedCam. But that the effect of over driving the input was to compress the level. It's the only way to explain the meter behavior with my tests. It was kind of redundant of me to even point it out. Not to defend the camera, it certainly has issues for sound, but I fully expect the sound on the Red with my pads in line to be excellent. My tests in the end sure sounded like "professional quality." I'd never do a gig, no matter the budget, knowing that the sound shouldn't be used. But I most certainly will follow up here to let all know how it went. p Quote
Marc Wielage Posted July 1, 2010 Report Posted July 1, 2010 I wasn't thinking that there is compression within the input of the RedCam. But that the effect of over driving the input was to compress the level. It's the only way to explain the meter behavior with my tests. It was kind of redundant of me to even point it out. I don't think there's any dynamic compression going on. My suspicion is that the Red camera is just using real cheap analog preamps that don't have much headroom. I've been doing tones at -20 and max peaks at -10 (or whatever looks close to that on the on-screen meters), and I won't work on a Red project where they won't pay for double-system. I also tell them to consider the camera sound to be "only a work track," and my clients have agreed so far. I understand that there are run and gun situations (especially news and docos) where the camera audio has to be the only audio, and I sympathize with that. But I think the Red is being used primarily for features, TV, and shorts, generally situations where double-system sound is needed and preferred. --Marc W. Quote
Philter Posted July 1, 2010 Report Posted July 1, 2010 Sorry, I think I've hijacked this thread... I left out an important logical step as I re-read. With my pads in line, the meters on the RED were only a few dB lower than without the pads in line. Almost a "normal" looking level. I measured 12dB drop in those pads between my two over the shoulder mixers so likely the real drop as they go into the RED could be somewhat different. Impedance differences and all. However the point is they should have been quite a lot lower than just a few with the pads. And the only sense I could make of that is there is too much level for the pre-amp or some part of the input stage to handle. Hence the compression comment. I regret the red herring, guys. No one questions whether it is much safer(at least more comforting) to record your own tracks. However, shit happens. And more often than you think. Wish I had more control over the work I need to take. After 38 years in pro sound I still have to suck the hind tit. For this MOW I'm being paid half my usual rate as it is - and that's twice what ANYbody is getting on the production. I still expect the sound to be distortion free on the RED with this set up. We'll see. How sweet the electronics sound? Hard argument to make. Phil Brouwer Quote
Jeff Wexler Posted July 1, 2010 Report Posted July 1, 2010 I left out an important logical step as I re-read. With my pads in line, the meters on the RED were only a few dB lower than without the pads in line. Almost a "normal" looking level. I measured 12dB drop in those pads between my two over the shoulder mixers so likely the real drop as they go into the RED could be somewhat different. Impedance differences and all. Phil Brouwer It is possible, depending on the wiring of the cables (adapter cables?) you were using, and the pads (inline connectors?) feeding the RED, you may not have been actually attenuating the signal at all. Just a thought. - Jeff Wexler Quote
studiomprd Posted July 1, 2010 Report Posted July 1, 2010 " the effect of over driving the input was to compress the level. " no! it does not compress the level, ... It clips, that is it shears off the tops of the waves. " there is too much level for the pre-amp or some part of the input stage to handle. " exactly, and so it just cuts off the wave tops! now, I'd be curious: what real network would order a MOW so cheaply made?? how many Executive Producers are there ?? (that could be where all the $$$ went!) anyone have any real experience?? Quote
Angelo Waldron Posted July 1, 2010 Report Posted July 1, 2010 Yes, I remember my first time......but that's another story. This just sounds to me like a classic case of line into mic level. Not helping you is the fact you recorded one channel -20dBfs and the other -30dBfs.. as FCP upon import pans all channels to center... something editors can't seem to understand. Quote
Sonny Posted July 2, 2010 Author Report Posted July 2, 2010 I'll certainly never claim to be mistake free, but line into mic level isn't the case here. Sat down with the AC and marveled at (what I thought to be) the needlessly roundabout audio menus as she set the camera inputs to line level for me. That FCP center pans everything at import is an interesting point. I'll have to follow up on that and see if getting rid of the hotter channel helps at all. Either way, we've got our clean files and our work flow figured out. Quote
Philter Posted July 2, 2010 Report Posted July 2, 2010 It is possible, depending on the wiring of the cables (adapter cables?) you were using, and the pads (inline connectors?) feeding the RED, you may not have been actually attenuating the signal at all. Just a thought. - Jeff Wexler Nope. For sure it's padding the signal. It's a standard H-pad. That's in an in-line connector. The signal cleaned right up with it. For sure it was doing it's thing. More below. p " the effect of over driving the input was to compress the level. " no! it does not compress the level, ... It clips, that is it shears off the tops of the waves. " there is too much level for the pre-amp or some part of the input stage to handle. " exactly, and so it just cuts off the wave tops! now, I'd be curious: what real network would order a MOW so cheaply made?? how many Executive Producers are there ?? (that could be where all the $$$ went!) anyone have any real experience?? You may want to read carefully the quote. It had the effect of compressing the signal. Digital stuff clips(or shears) and the sound is unmistakable. Over driving electronics usually isn't so harsh. Which is what this is. In effect, as it rounds off the signal it compresses it.(and distorts as it does - often not even audibly) It's the only explanation I have for the behavior of the RED's meters. I don't think that's a stretch. Sorry if I've been hard to follow. This is no real network job. It's an indy. The guy's own money, doing it cause he's driven. I feel bad for him. Throwing his cash away like that, thinking he's got something here. He's done some acting in Hollywood, England and now this. And I don't think this is terribly unique. Lots of us are bottom feeders. You guys forget. Phil Brouwer Quote
studiomprd Posted July 2, 2010 Report Posted July 2, 2010 " It had the effect of compressing the signal. Digital stuff clips(or shears) and the sound is unmistakable. Over driving electronics usually isn't so harsh. Which is what this is. In effect, as it rounds off the signal it compresses it " Clipping is quite different from compression, as are other types of distortion, although to your ear they may sound similar. the effect of overdriving was distortion (clipping); having the effect of compressing it would mean it was compressed, what you are trying to say is that you thought it sounded like it was being compressed... OK, maybe we could say that by clipping off the tops of the waves, the signal was being 'compressed into the available dynamic range', but clipping is an unacceptable, terrible sounding, brute force compression scheme, and even RED would not call it "compression"... thanks for the update: sounds like this is a variation on all the inexperienced wanna bee's making "feature films", but he is making a potential MOW, which will probably end up on You Tube, or the like... Quote
André Boisvert Posted July 4, 2010 Report Posted July 4, 2010 Apologies for being OT. Welcome Phil. Here's hoping one of your next posts will be in "Who I Am". If you're ever in Montréal and need extra gear, give me a ring. I still have an AT4462 debt that I owe you. Quote
ericwallace Posted July 30, 2010 Report Posted July 30, 2010 This red audio overmod issue is something that has been driving me nuts watching episodes of a show I recently worked on. Yes...these overmods went to broadcast because they didnt want to sync the masters we recorded. Quote
Scott Smith Posted July 30, 2010 Report Posted July 30, 2010 The clipping issues with the RED are well known. As with any digital system, when you reach 0 dbFS, you are out of headroom. End of story. When you go above that, the electronics in the RED will also tend to compress the signal, but at that point you are well into clipping anyway, so it doesn't much matter. Although I have had some moderate success with an interface with variable attenuators I designed specifically to work with the RED, I still tell the producers it's just a scratch track, as I have no control over what happens to signal once it leaves my cart. There's a reason we paid all that money for good digital recorders... --Scott Quote
ProSound Posted July 30, 2010 Report Posted July 30, 2010 This red audio overmod issue is something that has been driving me nuts watching episodes of a show I recently worked on. Yes...these overmods went to broadcast because they didnt want to sync the masters we recorded. Eric I am in the same boat I watched some of my show on air and 95% of the RED Audio is ok but not great but 5% of it is just horrible. I phoned post and asked what was going on and they said everything was fine. I went to the post house myself and double checked that my audio would sync correctly and it does. What it came down to is that the editor didn't want to have to mix the iso tracks so that is what he won't sync them. but then i explained that the 788T also records the mix track as well!! No such luck they don't want to sync so I let it go as no one is complaining to me Quote
Recommended Posts
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.