Complete Classical Music workflow

Getting Stuff Done Hints and Tricks
#161

Yes, the space where I recorded Syrinx is small, and that of Sunayama large and high, but I also find other differences, such as that the sound is richer, with more harmonics and nuances, more present and cleaner.

I have tried to record at different input levels in my interface to see if it improves the presence of sound, if it is very low then raise the sound is something dirty, and if the volume is very high then it seems to distort, even if I do not get to make peaks. By the way, how do you raise the general level in Ardour if the input signal is low? With the master, with the mic inputs, with a limiter?

The recording was made with only two AB mics, the guitar mic was close and the flute mic a little farther away. The room was very reverberant, but even a little reverb was added to the mix.

#162

According to the mastering engineers, the best thing to do is to raise volume in several stages, for example 3dB with EQ plugin, 3dB with compressor (even without turning on any compression i.e 1:1 ratio), 3dB with Ardour gain trim and another 3dB with limiter or any other combination. There are some explanations why that’s right thing to do, but I am not engineer so I don’t want interpret the reason wrongly.

I have recorded one particular flutist with the same microphone/interface in several different spaces and the results were very different. Different frequency bands travel different distances through space (high frequencies loose energy quicker so the don’t reach as far as low frequencies) and reflect differently off of different surfaces (stone, wood, …) which all add to the quality of the sound. Even if you decide to buy a new microphone It would be useful to know what you want exactly because different mics have different frequency response regardless of the price. There are some safe bets, though.

#163

You actually need to be very careful with this advice. The reason is gain staging applies in a DAW same as it would anywhere else, and if your plugin cannot handle an overleveled signal coming into it, you can clip the signal there with no real warning. While Ardour itself, and for instance many of the plugins and processes in Mixbus will handle these situations fine, the same is not guaranteed for all plugins.

I suspect the point of that advice was more to ensure that you weren’t overprocessing at any single stage of the signal processing chain more than anything. Otherwise, assuming that the above issue I mentioned doesn’t exist, there isn’t a difference mathematically between the process you mentioned and adding in 12dB with an amplifier process at the end of the chain, they will come out identical really.

The main issue is to make sure you aren’t developing to strong a signal at any single point in the chain to cause distortion or clipping in any process. Also keep in mind if you do add in stages like this, that where any sends are placed in the channel strip will affect their level as well and you could end up with some sends significantly stronger than others. Again if you are careful about gain staging mathematically it might come out even, but it is generally something to keep in mind.

     Seablade
1 Like
#164

Generally we record a nominal signal at around -18dBFS. The reason for this is more practical on the recording side, as we want to make sure there is plenty of headroom so that we are in no danger of clipping if something is played slightly louder than anticipated (And any professional engineer worth their salt will tell you that musicians never check at the same level they perform at).

If by ‘to low’ you mean that the input signal is sitting generally well below this -18dBFS, then you should correct this at your microphone preamp at the recording stage. If you are beyond the recording stage, you can do so by adjusting TRIM or utilizing Region Gain to boost the region gain which will have the same effect to get you to a decent signal level. In general this won’t be quite as good as increasing microphone preamp level in the recording, but as long as you were close the difference should be minimal.

Beyond that where to fix level issues for the output depends on what you are trying to achieve. Whether it is most appropriate to use a compressor on the input, or limiter on the output for instance depends on what you want to do to the signal. If you want to tame only the highest brief peaks to allow some extra headroom to raise the overall signal level then a limiter on the output could be more appropriate. ON the flip side if you find the overall dynamic range of the piece to great for comfortable listening, then a compressor might be more appropriate, but where depends on the goal again and it will affect the sound. These are usually decisions made in the mastering process however, so unless you are mastering your own piece it can be best not to reach for these first and instead let your mastering engineer weigh in here (Making sure they understand that this isn’t a rock and roll piece to be squashed to hell as some ‘mastering’ engineers in name only seem to think all things are)

Not sure if I answered your question or not though, hopefully it helped.

Seablade

#165

Back home and couldn’t wait to jump in!

Ensuring that your device and DAW are set for 24-bit bit depth and, say, 44.1k sample rate, you should be aiming for about -10 or -12 dB peak level in the DAW for the microphone capture. After much research, I can safely say that this is good and solid advice. No need to aim higher and, actually, you can aim lower and still not have any problems in 24-bit. This way you have a good safety zone for if any performers play with a bit more vigor after sound check :wink: I believe this, and much of what follows, is in alignment with what @seablade is advocating. For a new session in any DAW be it Ardour or any other, all your faders etc will be at unity and at this point you are simply using the gain knob on your audio interface, field recorder or whatever, to adjust to this recommended peak level. In post, you will raise the gain via region gain or gain plugin on the master bus to, say, -6 dB true peak for solo flute and call it good, or, listen back and figure out your desired LUFS level be it integrated, max momentary and the like. Here, as previously discussed, -18 to -20 LUFS integrated across the whole album with -1dB true peak is good. For a well-recorded album this could be all you need to do to put out a very nice recording. Please note that Mixbus, in particular, has a lovely workflow for gain-staging via the mixer. There’s a great video from Harrison Consoles explaining it here. There’s an earlier video too.

Often with classical music you want to retain all dynamics (or much of it) and so you won’t necessarily want to be reaching automatically for a compressor, limiter and the like. I would definitely only recommend a limiter for really, really loud stuff like orchestral climaxes. As discussed previously, for a wide dynamic range that may not fit in the intended listening environment some very slight compression 1:1.1 to 1:25 might work. Even better is parallel compression.

The point here though is that gain staging should not need to be happening via compressor, limiter and the like at capture. If you are not able to get a ball park peak of -10 dB or so for your fortissimo passages during recording after only setting your physical gain knobs, something might be wrong in the chain.

It might be covering old ground but I return to microphone arrays. Strict ORTF, DIN, NOS, EBS etc are standards for a good reason: under the right circumstances they have a good chance of producing excellent recordings. The Williams “stereo zoom” paper gives graphs to show how near-coincident pairs can be altered via separation and/or angle to best fit the recording angle you’d like/need. This way your workflow is finding the best spot in the room and then figuring out the most appropriate stereo array formation. Same for spaced omnis. I will say that Williams has much narrower omni configurations than traditionally thought of but I think he is correct in his outcomes. It goes without saying that a matched pair would be excellent but two of the same microphone bought separately from a company like Audio Technica will be extremely close and give very even results across the stereo field.

Room is also key, obviously. Record in the best rooms you can and adding reverb will sweeten things ever so slightly versus trying to cover bad things up. Advice given to me has often been to use cardioids in bad spaces and omnis in good ones but I realize it isn’t always that simple.

Anyway, back to the main point. If you are not getting -10 to -12 dB peaks as your input levels with just your interface preamps (and all DAW faders at unity i.e. untouched) check the following (the first two can probably be dismissed easily):

  1. Do you need phantom power (+48V) on your microphones and have you set this on your interface?
  2. Perhaps your microphone is faulty?
  3. Does your interface have a software control panel and are all faders at unity? I realize this is probably not an issue for Linux but you may want to check in alsamixer or similar that your faders for both playback and recording are all set to unity.
  4. Are the numbers looking good but headphone/speaker output low? Again, check physical knobs on your interface. At this tracking stage for a stereo pair without any plugins added, what goes in should be coming out at the master bus at the exact same level.
  5. Are your headphones/monitors working correctly? I’m very surprised that the Audient preamps are “distorting” before you even peak! Something doesn’t feel right…

There may be others but I can’t think of any more right now. I apologize if I’m stating the obvious but I assure you that I’m just trying to cover all the bases no matter how easy it might sound.

Also, I just purchased the Audient iD44 myself so I darn well hope the preamps are good for classical :wink:

#166

They are! I know a guy in my town doing excellent classical recordings with this Audient 8-channel adat box, asp880, plugged into RME Babyface pro.
Could you, please, test if iD44 works with Linux, and if so what kind of functionality is available through alsamixer?

#167

Of course, I’d be happy to. On a brief Reddit post the user claims every channel works in Ardour but I will do a more in-depth report as soon as it arrives at the end of the week. In other news, as part of the same order, I picked up a pair of the Shure KSM141s, pair of M40x headphones as well as the upgrade to RX 7 (vs going with Acon for now, at least).

I’m definitely now at a point where I feel that the only next logical step is to move into the upper tier of microphones such as a pair of Schoeps but, to this day, the best recording I ever made was with a single matched pair of ADK TLs in omni mode. They are magical microphones. I still have doubts whether paying through the nose for Gefell, DPA, Schoeps etc is worth it in the grand scheme of things. Pity I sold the ADKs when I moved across country :frowning:

#168

Thank you very much for your advices @anon60445789 @vasakq and @seablade!!!

Today I tried to record myself one more time, now following some of your recommendations. Before I was trying to get close to 0 when I was recording, between -3-6db and that might saturate the signal. The Audient controls were almost at maximum and I was about 2 meters away from the microphones.

Today I recorded a piece with a great dynamic range and I tried not to exceed -18db in the most extreme notes. The Audient controls have been almost in the middle.

To know the signal level I have added in the Master the plugin “True Peak and RMS stereo” which leaves marked the maximum signal peaks. I can’t see the laptop while I play…

To raise the signal I have done it in three ways, following the advice of the sound engineers ( @vasakq ). ALT+6 (I did it 5 times) on the tracks. I added the a-compressor on each track (Threshold -20db Makeup +3db) I added a mono LSP limiter ( 1:1 Treshold -3db).

I don’t know if everything has been too complicated and could have been done more simply.

The result of the recording has been better, it has much less wheezing, I think, although today my sound has not been the best :frowning:

The setup of microphones and the recording space has been exactly the same as in the other recordings.

Here is the result in LUFS out

Captura%20de%20pantalla_ardour4
Captura de pantalla_ardour4.png926×767 298 KB

And here the wav file:

#169

Thank you!!! :slight_smile:

I think so.

#170

I think @seablade is referring to -18dB RMS? I’m advocating for -10 to -12 dB peaks. In any case, -18dB peaks with 24-bit and good preamps should be fine, if a little conservative (given you know your own playing habits and aren’t likely to overshoot). I’ve recorded even lower when I couldn’t do a sound test beforehand will no ill effects…

Again, this doesn’t sound right to me. You shouldn’t be needing to compress or limit for solo flute to get a healthy signal. Simply raise the entire gain so that the peaks hit -6 dB or so. No compression or limiting unless on listening back you feel the quietest parts are too quiet. Compressed solo flute can sound as bad as compressed solo piano and it is almost never done. And are you not recording into a stereo track? Why a-compressor “on each track”? And why a mono limiter when you are dealing with stereo?

I’ll be honest, listening back to your latest file, it sounds like the life has been squeezed out of it. I much preferred your original.

I’ll try to be clear:

  1. Record (with zero plugins engaged) and reach c. -12 dB peaks just by using the preamp gain knob on your interface.
  2. After successful capture, raise gain in Ardour on the stereo channel or via a gain plugin on the master bus so that a) the true peak meter reads -6 dB max peak or b) if aiming for a particular LUFS level, use a LUFS meter and then adjust gain accordingly after initial reading.
  3. I would advise you to record everything for the album before any other processing. You want uniform gain changes across the entire album so that you don’t end up with your flute sounding louder or softer because you tried to make every single track the same loudness.
  4. Avoid compression and limiters on solo stuff if you can help it. It is going to take away from the sound most of the time for classical music, at least.

I wonder if something is being lost in translation here?

#171

Correct on all counts, (Except RMS is a bit of a misnomer when discussing complex audio signals). -18dBFS nominal means it is ok to peak a little above it on occasion, so peaking absolutely at -12dBFS would be fine and probably a very similar result. Heck peaking above -12dBFS is fine, that is why we set that low to give headroom for when things happen.

Recording peaks at -18dBFS is a little conservative but on even semi-decent 24bit converters, it should be fine honestly, and it is always better to be conservative.

      Seablade
#172

Yep. I know colleagues who record really low for opera because you never know what might happen. I have personally found that choral concerts always overshoot rehearsal settings because singers don’t want to over-sing before the performance. There have been a couple of instances when I thought the field recorder meter had clipped because a whole bunch of alums appeared unexpectedly from out of the audience for the final number but thankfully when I got home I only found -6 dB peaks. Phew! I forgot the exact peak level you could theoretically record at and still have lower noise than a 16-bit recording (anywhere north of -48dB?) but the point is that even ridiculously low as -24 dB peaks would be fine and of no consequence for a decent preamp recording at 24-bit. I’m looking forward to using the 32-bit float recording option on the MixPre6 ii…

@seablade For classical recording, I have always based preamp gain settings on peak level because the music is so dynamic. How would you suggest setting them to a nominal level? A mezzo-forte section? I’m used to thinking this way for mastering given Bob Katz’s work…

EDIT: And also, my limited understanding of nominal level is that it is mainly as a reference for setting up your entire audio system. A bit like calibrating a Katz meter, you would be using your ears on a properly calibrated system to set preamp level, final mastered level etc versus worrying about using your eyes to look at a meter. The more I think about it, I feel I should be using peak meters for capture and then K-meters / LUFS meters for mixing/mastering. By default I have my home system calibrated to K-20 and then do a final check against a LUFS meter.

#173

Nominal in this case can also be referred to as ‘average’ etc. It means your typical levels. I ran a downlink site once for an opera, walked in to find the venue with lots of problems, and the audio tech misunderstood nominal with peak. There isn’t necessarily a proper way to define it, which is why you get so many metering standards for average levels etc. I generally reference meters that are similar to VU ballistics or peak meters and base it off my gut though, and yes peak is a great thing to keep an eye on during tracking for exactly the reasons you mentioned, above all else you don’t want to clip.

Seablade

#174

I started recording on two tracks because I have two very different mics. I thought that maybe I should equalize them a bit different or maybe to move the gain for compensate if necessary. However the looks very similar and yes, I could use a Stereo track. I will do now.

[quote=“bachstudies, post:170, topic:101827”]

  • Record (with zero plugins engaged) and reach c. -12 dB peaks just by using the preamp gain knob on your interface.
    OK i will do it. -12dB peaks.
  • After successful capture, raise gain in Ardour on the stereo channel or via a gain plugin on the master bus so that a) the true peak meter reads -6 dB max peak or b) if aiming for a particular LUFS level, use a LUFS meter and then adjust gain accordingly after initial reading. I was watching the Mixbus tutorial. I didn’t know that Mixbus is loke Ardour…
    They do three things to move up the gain: a) Alt+6 on the tracks b) turn right the upper bottom right c) move up the gain control. What to do? The simplest way I understand is to move up the gain control but to do Alt+6 is also useful because increase the wave form and could be easiest to do the crossfades looking the wave. I don’t know the difference between the upper button right and the gain control.
  • I would advise you to record everything for the album before any other processing. You want uniform gain changes across the entire album so that you don’t end up with your flute sounding louder or softer because you tried to make every single track the same loudness.
    Yes. It’s the most sensible thing to do
  • Avoid compression and limiters on solo stuff if you can help it. It is going to take away from the sound most of the time for classical music, at least.
    [/quote] OK

mixer-1-strip-numbered
mixer-1-strip-numbered.png110×770 17.9 KB

#175

It’s gonna be good for sure. My sound recording problems are due to my clumsiness…

#176

I’m sorry, I didn’t read this post before.
Perhaps could be better don’t do this on Ardour and move only the gain control.

#177

Looking forward to getting the iD44 on Thursday. I wouldn’t say “clumsiness”! We all start somewhere with this and most audio engineers don’t have the musical talent you clearly have. Use your musical ear to your advantage! I think what we are suggesting in this particular section of the conversation will clarify how to think about things. I remember when I first started, I was making mistake after mistake. The worst was getting a well-paid recording gig with a very talented pianist and not being able to get any sound out of the interface and headphones for monitoring. I blamed a faulty interface on location but it turns out I simply hadn’t rotated the monitor knob from the “off” position. Quite embarrassing but such is life :slight_smile:

I would say that this is good advice on any DAW! If you have successfully recorded with peaks at around -12 dB peaks just use the trim control (upper button right) in the channel strip to increase your gain the desired amount to hit your loudness or peak target. Starting with trim also means that for more complicated recordings you could then use the fader (vertical slider) to make much finer adjustments to the balance if, for example, you have more than one stereo track or spot microphones to blend.

A trim knob is generally used to achieve a basic balanced mix so that 1) You can easily return to this basic balance by quickly moving all faders to unity and 2) because you get much finer control near the unity mark of a fader given it is not linear.

And, more importantly, trim knobs gain-stage the sound coming into the channel structure and the fader controls the volume after effects etc coming out at the end of the channel structure. Then, if you send a group of tracks to a mixbus, you could adjust input to that mixbus with another trim knob and so on. This is gain-staging through your whole signal chain until you reach the master bus. For a simple stereo classical recording you might not necessarily take advantage of all this stuff but it’s good to start thinking in this way. If I’ve misspoken about this someone please jump in as I want to clarify not confuse!

#178

Thank you for yours kind words @anon60445789!

I had hoped not to be much worse sound engineer than flutist, but right now I have my doubts… :slight_smile:

I think I have more clear now all the things, but I still have one more question about gain staging. On the Mixbus video the used also Alt+6 for increase the wave. When could b useful this technique?

Now I have recorded one more time Syrinx from scratch. Nothing about equalization, compression and limitiers. Only added with the trim 2dBs to the left channel and 4dBs to the right. I used two traks because after to do some tests the Sennheiser mic is a bit less sensible and I prefer control with two tranks. I hope I am doing well…

The Audient knobs are aroud the 60-70%

All the mics, distance and space are the same.

Can you listen the result, friends?

#179

I’d say that this would be useful for balancing different tracks. Obviously for single flute you probably don’t want to do this but you might for, say, a mixed album of solo piano, then choir, then orchestra etc. You’d at least want to get the levels closer so it wasn’t so jarring. The other way to do this is to put each on a separate track and use trim and fader. Same difference but depends on whether you want to use common plugins etc.

A benefit to using identical microphones is that you would normally set the gain exactly the same on each channel and you shouldn’t need to tweak after that. I’m assuming you did it here to balance the stereo image due to the different microphones. In this case, separate mono tracks are fine but be sure to pan fully L and R. For identical microphones, definitely use a stereo track.

On initial listening via my cheap and nasty desktop speakers it sounds great. What were your peak levels while capturing and what were the final values for peak and LUFS as you exported? On a quick analysis on the command line I’m getting -24.49 LUFS integrated with a -9.31 dB true peak. I think if you increased the common gain on both tracks by 4.5 dB you’d be sounding in the right ball park! That would get you to -20 LUFS integrated.

After you record the whole album, you could analyze the entirety and then adjust gain to reach -20 LUFS integrated and see where the peaks land.

#180

It sounds great to me too. Excellent playing of course, but the sound is excellent too. I listened through headphones and it sounds as good as any other classical flute recording I’ve heard (and I’ve heard many).