Thursday, June 5, 2008

Audio woes

Despite Julian's best efforts, we couldn't get the sound recording set up the way we wanted for tomorrow night's session. Let me share a few painful lessons with you.

Samson C01U USB microphones are great for recording podcasts, or when there's just one of you. Really nice. But trying to connect two of them at the same time, so you can record two people having a conversation onto separate vocal tracks, is a total bitch. Don't do it. At least not under XP. (And don't even think about it with Vista or Linux. You may have more luck with a Mac.)

For a start, the audio recording software has to be able to cope with two inputs. Audacity can't, so we install an old version of Sonar LE, which came with the mics. (Sonar 4 - the current version is Sonar 7.) I'm not a big fan of Sonar, but hey, you go with what you have, right? Well, it takes Julian about an hour of frigging around before he gets both mics to input simultaneously. At all. But even when we have both inputs, one mic is really loud, and one is really quiet. Actually, that's not quite true. One appears quite loud, but when we record it and play it back, it's also really quiet. That's partly an issue with using USB mics, they just don't have the power going through them.

So we install the Samson SoftPre pre-amp software to boost the mics a bit. This comes with a somewhat ambiguous warning that basically means it doesn't work under XP if you have two mics connected at the same time; it can only cope with one mic. (In my experience, I couldn't even make it work right with one mic, but we had to try.) However, with even more frigging around in the bowels of Windows drivers, SoftPre and Sonar, Julian eventually persuades both mics to input at a reasonable level. His audio ninja skills are strong.

So now we have a stereo stream coming off each mic. This isn't what we want for voice recording, we want two mono streams. More messing about, but eventually, we get there. At this point we discover something interesting. The mic input is completely different depending on what USB socket it's plugged into. Plug it into the top back one, and it's detected as a "Samson mic". In the lower back, it's an "Unknown mic". And in either of the two sides, it's a "Samson C01U". And the input volume is different on each port. Aargh!

Now we only have one problem left. The mics have too much range, so my mic is picking up Johnnie in the background, and his is picking up my voice. We try gating the recording, but that doesn't work. So we try gating the input, and that doesn't work either. It just clips things off. Basically, the C01U is a condenser mic, and that's just the way it works. We'd need to build isolation booths.

Finally, after his third cup of tea, Julian delivers his professional opinion.
  • Buy different mics, not condensers and not USB.
  • Get some different audio software, not an old version of Sonar.
  • Don't try recording on a laptop.
  • And don't expect decent audio quality if you're recording in the dining room.
And with that, he gives up and goes home. But that's what sound recordists are for. They know this stuff. I don't feel so bad that I couldn't make it work.

Oh well, I'll have to see what I can set up later in the summer. For now, I'm going to fall back on the tried (and mostly trusted) system of using Audacity and a single C01U mic, hope I don't get too many bits with two people speaking at once, and re-record individual lines when I do.

8 comments:

Overman said...

Capturing input from multiple good-quality mics is very difficult without some kind of audio interface sitting between the Mics and the PC/MAC. It'll set you back a few hundred, but it's oh so worth it if you see yourself doing this often. Check out the Onyx line by Mackie, and M-Audio has some affordable gear as well. To get lots of channels for minimum investment, best route is probably an 8-12 channel powered analog mixer which can accept the XLR mic inputs (again, Mackie has really affordable gear here), and then an audio interface that can accept eight or so quarter-inch inputs.

If you do this, however, be prepared for the possibility that Audacity may not be the best choice to use for recording multi inputs. I'm not sure if it supports the ASIO drivers typically powering this kind of multi-input equipment. You might need to think about getting some "real" DAW software designed for this, such as Cubase, Cakewalk Sonar, or something along those lines. Another few hundred perhaps, but it'll change your whole range of sound possibilities. Finding a good vendor in the UK who sells this equipment is a great idea, as they'll be able to help advise a compatible setup that meets your budget needs.

I cringe at the thought that you've been advised to abandon the condensers, essentially because they're TOO GOOD at what they do. I understand the scenario and context, it just seems like such a shame. The kind of quality you hear from VO actors like Richard Grove are captured with the richness of condensers; I swear by them for studio VO, personally. I really think it's worthwhile to get creative with sound control techniques before switching mic types, it'll so be worth it if you can find a way. If you can just find an arrangement where you baffle the sound enough to MINIMIZE the cross-actor bleed through (not eliminate, just minimize), then you can very likely gate that remaining noise out later with a software plugin on the track.

I'm not sure if this analogy holds up, but it's what comes to mind. When I think dynamic mics, I think theatre; when I think condenser mics, I think film. Dynamics capture the broad strokes and well-projected utterances... but condensers capture sighs and labored breathing and sniffles and all those intimate subleties you often want for a machinima piece - especially in your closeups.

And I guess that makes USB mics the high school drama department or something. Stay away. Icky yucky. Unless you're recording radio chatter, they clip frequencies quite nicely for that.

Fight for those condensers, you won't be sorry.

Overman said...

Quick follow-up thought... the laptop isn't necessarily a problem in and of itself. I know pros who do field recordings with laps all the time. But they're always using them in conjunction with a setup similar to what I'm describing, where some kind of device outside the laptop is negotiating the audio's transfer into the computer, either by USB 2.0 or Firewire.

Laptop sound card? Yeah, not gonna give you the results you need. But the laptop's "laptop-ish-ness" isn't a problem at all, as long as it's got some decent muscle and RAM.

Anonymous said...

Audacity DEFINITELY won't accept multi-track inputs in the way that Overman's suggesting. We beat our heads off that for months on BloodSpell.

(FWIW, all of BloodSpell was recorded on a Presonus Firepod with SM58 vocal mics.)

If you don't have isolation booths, you're going to get some bleed-over from one mic to another, no matter what mic you're using. As mentioned, we were using SM58s, which have a very small pickup range, and we still got the same kind of bleed you're describing.

It's also worth noting that we recorded onto a laptop in just the way Overman's describing - from the SM58s to the Presonus to the Macbook Pro.

Personally, not being an audio guy but having a fair amount of experience setting up audio recordings for Machinima work, I'd go with condenser mics (the non-USB versions of the CU01 are quite affordabout, but you'll need shockmounts, stands and pop shields too, which will rack the price up a bit), and a USB mixing desk rather than a multi-track system. My opinion, which you should promptly ignore if Phil thinks differently, is that for most Machinima, a single vocal track mixed down from the mics is fine. It gives you a little less flexibility, but also a lot less editing hassle.

Feel free to phone or email if you want more suggestions on this!

snorkel said...

Hey overman!

You are right in two respects:

1. Condenser mics are great for voice.
2. USB mics suck.

A condenser powered properly off 48V phantom will have superb signal-to-noise. A condenser powered off USB with a stinky driver will not. And as Matt found, not all USB ports are the same. It may well be the case that the motherboard comes with 2 ports off the main power rail, whilst the remainder may live on an auxiliary card or secondary rail, hence the weirdness with record volume and plug-and-play ID.

For laptops I would go for something like this: http://www.m-audio.com/products/en_us/FireWire410-main.html (presuming you have a Firewire port!). For towers, an internal card such as http://www.m-audio.com/products/en_us/Delta44-main.html is good, though because it has no mic in you'd need a mixer with phantom and bussed outputs. In fact, I'd go further and suggest getting one of these (per mic) too:

http://www.bonnersmusic.co.uk/buy/Behringer/Behringer_Composer_PRO-XL_MDX2600

For £89 you get a pre-amp, gating, compression, a de-esser, some adaptive EQ and some (faux) valve warmth.

Run those into a mixer, and the mixer out into the sound card.

I use Cubase in my studio. I can't say that it's better than Sonar or Logic or Samplitude or whatever, but there is a light version and it's half-decent. If you wanted me to do the engineering (which I'd be happy to) then I'd prefer to use Cubase.

An audio laptop requires a big hi-res screen (because DAWs such as Cubase require masses of screen estate); they require a fast video card to drive that screen (a slow card will lead to crackle as it fights the streaming audio for access to the bus); they need LOTS of RAM (4GB would be good); they require fast, quiet hard-drives. 2.5" HDDs are rather poor, you really need to record to a 7200 RPM unit. USB2 may not cut it, and you're already using your Firewire port for your audio card. So laptops are not great for recording. But some are better than others. Eg http://www.shop-sonica.com/cat_livetracker_lite.cfm?gclid=CNf9pOLB35MCFQTtlAodtAy1WQ

I prefer tower systems for this reason. You can get fast, quiet hard drives (eg Samsung Spinpoints), super-quiet PSUs, sound insulated cases, PCI cards that don't suffer from cable latencies... Also, you can set them up so that they are optimal for audio tasks (no wireless for one thing, I used to get occasional pops on network activity during recording).

Anyway, that's enough talk for now. Hope that helps!

Overman said...

Saw this tonight and thought of you, Matt. Interesting to note that they don't seem the least bit concerned with bleedthrough on their group scenes.

Oh... and it appears to be condensers all around. :)

http://www.youtube.com/watch?v=t926G1FFzL4

Where bleedthrough IS a real problem, though, is when what you're hearing in the other mics is the ROOM, moreso than the direct noise from its source. Then it creates some real cleanup headaches for you later. That's probably what your guy was getting at with regard to the "dining room" comment.

@snorkel - Good to hear from a fellow Cubase enthusiast. You're right about the screen real estate (I use a dual monitor setup to keep it manageable)... however, for the recording session, I think a single regular-sized screen would probably perform just fine, because you don't need as many windows for recording. I'd never suggest *mixing* with DAW software on a laptop though, for that screen reason as well as the CPU and RAM required for the VST effects you'll likely employ at that time.

Matt Kelland said...

Thanks for the input everyone. That Family Guy clip is interesting - good find! The difference though is that they can probably just slap the audio of all the voices onto the footage as is, and bleed isn't an issue. However, because Moviestorm uses the audio for automated lipsync, it's vital to have clean audio, or the character will end up with extraneous mouth movements.

However, it all worked out fine with a single mic, as I noted in today's post: http://worldofmongoose.blogspot.com/2008/06/audio-joy.html

Unknown said...

Yeah, this microphone is strong oponent!

I connected this microphone to laptop, an old Dell latitude x300 wint a tiny intel chip running ubuntu studio 8.10. I tried to record voice and mix it in realtime with music from a player similar to winamp. I finally at about 1.30 AM brought it to effect!

But when tried to repeat this on my friend's desktop PC I had to resign. He has some audigy and other motherboard architecture. In fact when using USB microphone you connect to your computer another soundcard. And your OS should support two sound devices simultaneously. I succeded it till now on the laptop.

Unknown said...

Well - I succeeded setting up samson c01u at another pc. The matter was I forgot to set up samon as a source of signal... I used jack audio server under Ubuntu linux. Now I found at Ubuntu forums there could be a possibility to connect some more usb mics:
http://ubuntuforums.org/showthread.php?t=1128748