Recording an orchestra

Thanks so much Tom, this is great help. First quick reaction; I’m doing this as a favor, so I’m not investing in battery-powered recorders at this point. I’m using my old laptop and the Steinberg MR816CSX, which both need to be plugged in. And I’m using AT-2020’s, so it’s quite difficult to estimate the real recording angle, since all I can do is look at the back of the mic and try to align it with the angle I’m shooting for. Since I had three ensembles to record, I stuck with 110 degrees based on the template I printed out. I can imagine in the future I would have small diaphragm condensers and the angle would be easier to play with. But, to be honest, they would be cheap! Nobody’s donating to this cause.

On the mix, thanks for the comments. I think I might be able to move the tracks toward center to make the stereo less extreme. I may have misunderstood, but isn’t ORTF intended for cardiods rather than omnis? I’ll give another hearing to the high frequencies and their impact on hall space. I did a little a/b with some classical videos and I noticed that they were decidedly more mid-range than what I captured, so that’s why I didn’t feel so bad about cutting the highs.

Much appreciate the suggestions about how to handle those pesky cymbal and timpani peaks. I think my version of Wavelab can do that. As well as do the RMS analysis.

And I would love to hear your take on it. Let me know where I should post the giant files.

I did a poor job at expressing my thoughts in my prior post. I’ll try to add some clarity here.

It seems like you did the ORTF just fine. The overall stereo image is fine. I would not mess with it. My comment was designed to address the SRA (Stereo Recording Angle), not the 110 degree angle of the mics in ORTF. Even though the mics are at 110 degrees, using ORTF yields a 96 degree SRA. Any performers outside of the 96 degree mark will be pinned to the side channel. It sounded to me like some performers were beyond the 96 degree mark. However, this is very very minor. So no worries here. Probably best to ignore my comment.

Do not pan the channels in. They should be panned hard left and hard right. This is how ORTF works. What you have is good.

You have it right. ORTF is designed to be used with cardiods, not omnis. The point of my comment, which I should have made clearer, is as follows. When ever you pick an array, mics, distance - everything is a bunch of trade-offs. To compare ORTF to a standard AB configuration, here are the trade-offs. ORTF uses mics with a cardiod pattern. The laws of physics usually cause a low end freq drop for a card pattern. However, ORTF presents a great stereo image, is very forgiving, is hard to screw-up, and sounds good even in rooms that don’t have great acoustics.

If you went with an AB config, it would typically be omni mics spaced 50-60cm apart. The omnis typically don’t have a low freq roll-off like cardiods do. So more base is captured with omnis. In a great hall, omnis can sound glorious. However, there are downside trade-offs. If the room isn’t great acoustically, omnis will capture this unattractive sound and reproduce it. Also, an AB setup with only omnis relies only upon timing differences to create a stereo image. So the imaging isn’t as quite as precise as it is with ORTF (which uses a combination of timing and intensity/volume differences). Everything is balancing tradeoffs based on the equipment you have, the performers, and the recording space.

I believe that ORTF was the way to go. The slight low freq roll-off is a small price to pay for all that ORTF brings to the table. ORTF is a very versatile array.

OK. difference of opinion here.

I’ll pm you - I have a drop box account. Might work to do it there.


OK, interesting additional note, the local professional orchestra would like to have me record them on the basis of my recordings of the youth orchestra. They are looking for a recording for their archive, and nothing more. My main concern is redundancy of the recording, as I don’t have equipment for that. As mentioned earlier, I’m recording on a very old sturdy and reliable Thinkpad (into Cubase 4.5), but anything could happen. So very short term, I was thinking to get a Zoom as a backup to ensure that they have something for their archive. Eventually I’ll get a new laptop, but this laptop is essentially my field recorder, with two hard drives, no internet connection (most of the time), tuned for Cubase, and a record of solid success.

I intend to buy a new stand, as has been suggested by Tommyboy and MatJones, and maybe I’ll spring for a couple of new SDC mics before the gig. Ideas about which mic are welcome.

Congrats on getting a nice recording gig!

Redundancy is very important in live events. My two cents: Figure out what will happen if any one part of your signal chain fails. Build redundancy to eliminate failure.

Use passive transformer splitters. Art is OK. Radial and Whirlwind are better. Signal should be mic-splitter-recorders. Put the direct through to the recorder with the best pre-amps and ADC. The transformer side to the lesser recorder. A simple Art S8 will cut the mustard. If you save up money down the road, I recommend Radial and Whirlwind splitters. If you can get them with Jensen trannies, even better. Radial JS2s and JS3s are very nice.

Recorders - Using a zoom is fine. Even better if the zoom can run on batteries and not be reliant on system power, or automatically switch to battery if the main power stops working.

Mics - your AT2020s are fine. Consider a backup pair. Could be built in mics on zoom recorder. Reason for this is if the power goes out (from a loose cord, etc…), you need something. Could use a 4 channel zoom - two channels for built in mics, two for split

SDC Mics - if you want to get SDCs, consider saving up before putting money on lesser mics. Consider mid-priced mics (or better if possible). The mics used make a difference. Mics and mic position impact sound more than the pres and ADCs. Ideally, you’d have a slight HF lift on the mic to compensate for the distance btw the group and main pair. My best pair of SDC mics are Schoeps MK5s, which sound beautiful. My medium priced SDC mics are Sennheiser e914s. Sounds good, but not as good as the Schoeps.

Some people like Shure 141s and you’ll sometimes see these recommended on forums as a starter mid priced mic. I have a pair and really don’t like them for main pair on orchestra because they are dark sounding mics. Recording orchestra with these results in a dull sounding recording.

If you can only swing lower priced SDC’s, maybe consider R0de’s M5 and the Audio Technica 4041s. I have no experience with these mics, so I can’t comment on sound quality. I would also not try these mics without using the AT2020s at the same time to have an insurance policy. It would be good to to be able to return them if their sound didn’t suit the recording being done.

For mic’ing - keep it simple. If you do decide to experiment with other mic configurations or mics, consider setting up a simple ORTF array in addition to whatever you experiment with. That way you’re covered if you don’t like the other config. ORTF almost always sounds good. It might not sound best, but will generally sound good.


Thanks, Tom! Sage advice as always.

Hi all, I’m re-examining my mixes, particularly of the concert band. The problem here is that their set went from a flute quartet to a piece featuring two drum kits (both at stage front, left and right, showing off their two drummers). If I normalize the whole set, then the first three pieces are very quiet, and the mastering settings have barely any effect, because the ensemble doesn’t even get to the threshold settings. What do you guys do in that situation? I’m thinking ride the gain on the first three pieces and hope listeners don’t actually notice. It was a bit unusual. Just before the concert band got on stage, there was a small string ensemble, and I had normalized their part, but now they’re louder than the first three concert band pieces. I guess there is some happy medium.

And, I have to ask another question, thanks for your patience… why are SDC’s better than LDC’s for an orchestra/ensemble? I can think of two reasons not related to sound: hardly visible to the audience, and much easier to set up the ORTF or other method angle! I tend to use my LDC mics (AT2020) for anything where high frequency is an important qualifier, such as acoustic guitar, acoustic bass, and drum overheads. For guitar amp, snare, toms, I’m using SM 57’s. Nobody cares about the high frequencies there! Thanks to all.

I would suggest mastering each song/piece on its own. For maintaining cohesiveness of the set, aim for different average volumes for the pieces (but not that different). If the flute piece is super quiet and drum piece is super loud, master these so that the flute piece is a little quieter than the drum piece. It’s a balancing act to maintain a difference, but to make them both listenable.

Some peeps who do classical would call that blasphemy, but I think it makes things easier to listen to. Gain riding that you suggest is a fine way to approach it.


SDCs are generally better for recording an orchestra for the following reasons:

  • Less visible (very important for live performances)
  • Generally better (more sensitive) for transient response
  • Generally better (more accurate) for off-axis pick up. This is a key reason, especially if you’re using ORTF where you’re picking up critical elements of the orchestra off-axis. Even when using a spaced pair (AB), SDCs will usually win out over LDCs because the reflections (reverb) they pick up will not have their off-axis responses skewed (freq) as an LDC might.
    • these are all generalizations. One needs to really analyze on a mic by mic basis.

      LDCs can have a slight edge in lower noise. However, good SDCs have such a low noise floor that this is irrelevant.

      All this being said, AT2020s are still fine for recording an orchestra. And if that’s what you got, it should be fine. The high freq lift helps for use as a main pair for an orchestra. As you go further down this path, you can consider upgrading mics. I used AT4050s at first, then (after a while of saving) bought a used pair of Schoeps mics (MK5 caps) from Vintage King. There is a big difference. It’s easier to dial in a great sound with Schoeps. Less monkeying with EQ or other measures (in fact, I almost never use EQ on the Schoeps). Sounds natural.


Thanks, Tom, I bought the AT2020’s based on affordability at the time. They’ve been good drum overhead mics. I am thinking SDC’s for reason number one: not so much in the way! And I have no experience with them, so I’d like to advance, eventually. My sense at this point is that the mic is far more important than the pre-amp, but the pre-amp is also important. I have the MR816csx for preamps, and I am sure there would be arguments about requiring a tube preamp, but I think it’s not my weakest link.

On the other hand, if I bought better mics, and fed them into a Zoom H6 (I only mention this because I was recently at a really good classical concert where the guy recording was using a Zoom H6 with Beyerdynamic SDC’s (I don’t know the model)) and I wondered whether the better mics would be a greater improvement than the presumably worse preamps of the Zoom H6). As I mentioned earlier, I’m wary of depending on my old laptop, and thinking of backup, so might want to buy a Zoom or equivalent as long as it would take external mics. But I don’t want to take a step backward from the MR816csx.

And thanks for your thoughts on mastering. I’m trying to find a middle course.

Recorders - If you’re set on Zoom, consider an F4 or F8 over any H series. Better preamps. F4 currently on sale at Sweetwater for $599. However, for $50 more, you could get a new Sound Devices Mix-Pre3. If you think you’ll need more channels, consider the Mix-Pre6. Preamps on F4 will be better than H series. Preamps on Sound Devices will be better still. If you could stretch your budget, I’d pick Sound Devices over Zoom for the following reasons:

  • Limiters. Zooms are digital (meaning AFTER the A/D converter). SD uses analog limiters BEFORE the A/D converter. Bottom line - if you have unintended peaks, the SD will handle them much better with far less distortion. I ALWAYS record with limiters on when on location at live events. You never know what’s gonna happen. I’m good at setting levels, but the safety net is important. If you always get sound checks, maybe this is less important. I have gigs where I don’t get good sound checks, so limiters are key.
  • Support. I have called SD support a few times when issues cropped up with my 788T. Every time I’ve been quickly connected to a knowledgeable SD engineer who helped me resolve my issue. With Zoom - you’ll not have anywhere near this level of support.
  • Sound Quality - it doesn’t get better than Sound Devices. I paid about $8K for my SD 788T with CL-8 attachment. The fact that you can get SD quality now for under $1K is phenomenal.
  • Reliability - pros use SD for a reason. They work, are rugged, and reliable.

Another recorder comment - I also have a Roland R-44. It’s a nice recorder. Very easy/simple to use. And i actually use the built in mics for recording some school events my kids are in. However…the sound quality is not on par with my SD 788T. Different league. R-44 isn’t bad, just not as good. Also, my R-44 has digital limiters, which i’ve bumped against from time to time. They do not handle overs as well as the SD analog limiters. You can hear some distortion on the R-44.

tube Preamps - I would not use tube preamps on location. One - I don’t want to have to plug into anything. I’m free of power outlets and all of the headaches that can accompany that (getting clean power, conditioners, finding outlet, more cabling to manage, etc…). Two - SD preamps are very clean (but not sterile) - so no need for external pres. F4 or F8 pres are fine also. Three - tube pres are often used to color the sound (although this often comes more from transformers than the tubes), which is not what I’d want from a classical recording.

Which is more important - mics, pres, converters? Mics are first. Pres second, as long as they’re not too noisy. Converters last. Converters are quite good these days and it doesn’t take a lot of $$$ to get decent conversion. However, mics are different. It generally pays to get more expensive mics.

If you can hold off on buying mics for a while - save up for a used pair of Schoeps with MK5 caps. Versatile as can be omni or card. Sound is fantastic. You’ll never regret having a nice pair of schoeps mics. And they hold their value.

If I had your setup and was looking to plan an upgrade path, I’d do it as follows:
1 - Get an SD recorder. Less headaches with power. Would also eliminate using a laptop and separate interface. Less to carry. Easier to use. AT2020s, while perhaps not ideal, still turn out respectable results. You’d still need laptop, etc… if you used it to be redundant. However, getting at least one recorder off of the power grid helps for stability/redundancy if somebody trips over your power cord (or if breaker blows), the battery powered SD keeps on going. If you bought another recorder down the road, you could ditch the laptop.
2 - Buy a passive mic splitter.
3 - Mics. Get SDCs with a slight HF lift to be used as a main pair. If you could save up for a while, consider a used pair of Schoeps with MK5 caps (benefit of omni and card patterns). If that is beyond budget, consider other small SDCs.

Just my 2 cents.


I was not set on the Zoom, it’s just that I met someone who was using one for this purpose, and was thinking of redundancy at low cost. I was concerned about mic-pres though. So your input is very useful. The Mix-Pre3 would seem to be a much better choice for the money for this purpose. I even wonder if it is a better primary solution than my laptop with the Steinberg MR816. The MR816 is supposed to have decent preamps with phantom power, and given my usual recording of my band, the eight inputs has been very useful to me. I don’t have examples of a bunch of these devices to compare, but I used to use a MOTU interface, and the Steinberg seems to me to have much better sound.

Then I’m wondering, if I have a Mix-Pre3, do I really need redundancy when recording live? No moving parts anywhere, battery powered, it must be extremely reliable. Meanwhile on the laptop, there’s operating system, Cubase, hard drives, power cords, firewire cable, etc. that are all vulnerable.

I never even considered a tube pre-amp, and I’m glad you confirmed what I already thought, which is mics most important, pre-amps next, and converters third (nowadays). I got those AT-2020’s for next to nothing, so I really have to consider an upgrade, and you are no doubt right that I should wait and save up.

Thanks again, Tom.

And I also wanted reply to your comment on limiting. In my current setup, any limiting would be in my laptop, so I’m pretty sure if I go into the red on the MR816, my only recourse is to try to cover it up in Cubase. So that’s why I was trying to be very conservative on the input levels, given 24-bit dynamic range. I didn’t hit red once last time out. I hit red every so often when the band is recording, but nothing so ridiculous as to be a problem with our music. But on that Mix-Pre3, having an analog limiter in front would be really useful in a concert recording. Anyway, I think I understand that correctly!

If you are recording a live event with no second takes, i would recommend redundancy -no matter how good the gear is.

OK, getting ready for the local pro orchestra this weekend. I did buy the QuickLok A50 in the meantime, so at least that issue is solved! Man, that thing is a heavy beast! I think someone would be more likely to break a shin than to knock it over.

I took a look at “The Stereophonic Zoom” pdf, and I think I have a better understanding of the SRA (I hope) so I’ll set up my ORTF to keep that in mind – basically by moving the stand back a row or two if necessary. I still have no redundancy, which is unfortunate, but the orchestra management is not too worried since this is just for archive, and nobody seems to have listened to any of the archives in years! I am truly appreciating the advantages of battery power, so maybe before the next one, I’ll be there.

One consideration that occurred to me is that while I’m recording in Cubase, I don’t have to sit there with my headphones on – I can see the wave forms as they come in, so I know if anything has gone wrong. With a portable digital recorder, I suppose the operator really must wear headphones the whole time?

Another thing I’m wondering is if 32-bit float would provide even more headroom. I’m using 24-bit now. But if I used 32-bit, wouldn’t I be able to take in a very low signal and still have tons of headroom? I know it takes more disk space, but that’s not going to be a problem. At least in this particular instance, I’m going to be able to record the dress rehearsal, so I should have a good idea of the peaks.

I’ve listened to the recordings of the orchestras from around the area, and I’m not too fearful. I think my mics are not going to be the weak link. Redundancy definitely is. But the ORTF approach, subject to consideration of the SRA, should provide a decent recording. As far as I could tell, the last guy who did this job used one stereo mic, and I assume inside it’s an X-Y array (what else could it be), so maybe I can do something a little better, I hope, if the laptop keeps running.

Love to hear anybody’s advice at this point, thanks.

Glad you got the A50. I’ve had mine for many years and it’s still going strong. As I get older, I ponder how long I can keep carrying that thing around. It is a beast.

Portable digital recorder question - nope, you don’t have to wear headphones the whole time. Meters indicate signal being recorded, conceptually similar to wave form drawing you look at in Cubase. I don’t wear headphones much when recording live events.

Bits question - 24 is enough as it picks up plenty of detail while leaving a lot of headroom. I haven’t recorded with 32, so I can’t comment on this based on experience. Maybe somebody who has used cubase to record at 24 and 32 bits can weigh in.

Mic point - you’re right. Stereo mics are often XY. You’ll do much better with your ORTF array.

Redundancy - you’re right here too. This is your weak point.


Was thinking of the bits question. I’ve never seen or heard of an AD converter that operates at higher than 24 bits. Maybe a 32 bit converter exists, but I haven’t seen it. So if the converter is at 24 bits, would it make any sense to have cubase recording at 32 bits?

Thanks Tom… I’ve recorded in 32-bits float (a project that was passed to me). In general, I guess it’s overkill. But if, as you think, the converter is working at 24 bits, then I can’t think of a reason to record higher than 24 bits. Anyway, 24-bit is huge headroom as it is, so I guess I’ll stick to that. Wish me luck!

You know, I was wondering, as a sort of side issue, what makes one pre-amp better than another, assuming they’re solid state. I know lots of people want tube pre-amps for coloration, but that’s clearly not applicable for a live classical performance (I think - maybe somebody feels it’s a subject to discuss). I am wondering what an expensive solid state preamp has that an inexpensive solid state preamp doesn’t have. I can only say from experience that my Steinberg MR816CSX seems to have a better recording sound than the unit it replaced, at MOTU 8-pre. I still have that unit, but I only use it to record additional percussion tracks!

This is great thread. Thanks for all the good advice and ideas.

It may be generally a quiet forum, but there are a lot of smart and genuinely helpful people here!