That is so true! The makers of these musical instruments have the same problem. I know a number of flute makers, and for most of them it’s hard to let a flute out of their shop as there’s always one more small improvement or tweak they want to make. When I went to pick up my current main flute, which was made for me in 2004, the maker asked me to go away for three hours and come back as he was still making some minor improvements to the embouchure (I play wooden simple-system flutes, the classical flutes used in orchestras during most of the 19th century before the Boehm-system flute was invented and started to take hold).
If it wasn’t already clear from my previous comments about the playing and recording, you can count me in on this too. For this particular recording, I’m just suggesting better loudness levels that approach some kind of classical recording expectation. I have no idea how this piece would factor into his whole album concept so if he has plenty of other loud pieces, this may well end up quieter than -20 LUFS integrated.
There is, of course, the question of how @Aleph’s approach would work in other situations, especially if he enjoys this process so much that he wants to record other people. At that point he may find that his use of two different microphones in an unorthodox pattern might not work. I have suggested previously that he invest in a reasonably decent pair (doesn’t have to be top tier) and learn the basic stereo arrays and other audio engineer practicalities. However, as I have also previously stated, unorthodox can sometimes work wonders but I do think an engineer needs to have the basic tools figured out in order to move outside of the box and know why.
There is also a clear difference in my head between stopping the tweaking to put out a particular recording and learning as much as you can as an audio engineer through your whole professional life. I’m simply treating @Aleph as if he might want to one day record others.
Anyhow, yes, this recording is excellent!
Sure! I’m sorry, I followed the news while rushing to work and I didn’t find time to go through more recent posts. I agree that the decision on loudness must be made relative to the other tracks in the album.
No worries and no need to apologize! I was worried more about my own writing. I realize that I’m been guilty of writing quite long responses so there’s always a good possibility that any praise I give is lost in the mix (to coin a term )
Thank you friends!!
I am happy but still the sound is a bit wheezy. Maybe the mics and the space should be improvable for a professional recording. Or perhaps I also should practice a bit more…
I will record Edgar Varèse one more time, this is one of the loudest pieces, for then decide the LUFS.
Yes please, thank you.
Of course I would like to record others, maybe not soon, but what is certain is that if I don’t learn the technique well I won’t be able to do it, even if I have a good musical ear.
I have to think seriously about it. I think I must to.
We could also talk about monitoring. Or at least to look for a good tutorial.
Are you referring to the sound of the breath? If so, natural is always preferred in my book. Don’t worry too much about getting rid of it. If it’s a technique thing then I can’t help you I always appreciate recordings in which you are allowed to hear the key clicks of the oboe, for example. It always sounds artificial to me when people use spectral editing or EQ notches to remove them.
If this is the loudest piece then that should be fine. Be sure to make a note of your gain knob settings, microphone placement etc. Don’t do any effects or further gain-staging until you have every piece edited and lined up in the DAW. Then you can paint with broad strokes and ensure the album as a whole is calibrated to your desired loudness level.
You’ll have to let me know if there is anything in particular. My best advice for starting out is to calibrate your monitors, depending on your room size, to be a certain SPL output with a sound meter. Here’s a good article/tutorial: https://www.soundonsound.com/techniques/establishing-project-studio-reference-monitoring-levels
On page 3, it gives you recommended reference levels dependent on room size. In my case 74 dB SPL C-weighted for each speaker works out very well. Trying to achieve 83 dB SPL C-weighted for each speaker is likely to be quite uncomfortable (ear damage over a long period) unless you are fortunate to own a very large mastering house
The point is to monitor with your ears and not staring at a meter. If the music sounds too loud, turn it down…if it sounds too quiet, turn it up. If you get into the habit of doing this, you’ll be surprised how closely a good level aligns with a suitable LUFS level. Obviously, for critical instances like streaming or broadcast where a stray peak or wrong LUFS integrated can result in rejection or a fine, you do need to do a final check against the meter. My advice is calibrate your monitors, mix/master with your ears and then do a final check with your favorite LUFS meter.
I think in this case you are talking about the tone of the flute in the recording, right? It’s perhaps not sounding as pure as you like and there’s some high-frequency hiss in places? I could only detect a few places where the tone sounded a little airy and less focused, but given your mastery of the flute I’d be surprised if that was lack of tonal control on your part. You could try playing a little with EQ cuts in the high frequencies, although you have to be careful to avoid the common mistake of recording engineers who try to make the flute sound “soft,” you want to retain the bite and attack and fullness of the sound, and I think that means retaining as much of that high-frequency content as possible. It’s one of the reasons I decided to try recording myself, since engineers at studios where I’ve recorded in the past never produced a sound that was satisfactory to me, even when I gave them recordings I liked as a reference.
Yes, exactly. You are also flutist…
I tried to do some EQ cut on the 7k-8k Hrz, but it doesn’t get much better. It seems electronic distortion, or saturation in the high and loud notes (Not specifically in Varèse around minute 3. That particular note is very strident, it’s like that…). Maybe the quality of the mics is not the best… also depends a lot on which headphones or speakers you’re listening to it with. Anyway, I like the sound more than before
Today I recorded the piece with the widest dynamic range. The difference in LUFS is quite wide with the same settings:
Syrinx: -24,47 Integrated / STM -19,18 / MM -15,37
Varèse: -13,86 Integrated / STM -6,93 / MM -3, 56
After reading I think I will continue to use headphones for a while longer
Has your Audient come home yet?
Just arrived an hour ago. Looks like a have a slight issue with the KSM141 pair with a seized up stereo bar and case foam stuck in the ends of the microphones. Not a happy camper. Got far enough to plug in the Audient to see if it lights up (yes!). At this point with Halloween for kids, might be tomorrow at some point before I can seriously investigate.
Honestly, unless you know exactly what you are doing, EQing is probably not a useful step at this point for a beginning engineer. I mean no offence by this. Focus on the recording part for now getting the best sound you can and perhaps try to learn from some local engineers in a decent studio where you can hear everything properly. I’ve yet to meet anyone brand new to engineering, particularly classical, who can just go in and make everything right.
What are your true peaks on these? The Varese seems quite loud in comparison to Syrinx. Did you record both pieces in one recording session with identical preamp settings on the interface? Sounds like they were. I’m a little surprised but if that’s right, no worries!
EDIT: Also, if your momentary max is -3.56 that’s awfully high for a 24-bit recording when you are able to leave significant amounts of headroom…
Ugh, I hope you can solve it.
Yes, exactly the same. The only diference I did Syrinx on MXLinux and Varèse on Antix, but the settings are the same.
A good advice, don’t worry, I will try to do it.
Hmm. Then we can’t be entirely sure. My point on the high values on the Varèse are that if the LUFS momentary max is that high, I’m worried that your true peak level has clipped. Could you post the file and I can analyze?
I was wrong, the default settings in Ardour are 44.1 and 32-bit
A did. It is above.
Looks like you were lucky and hit about -1dB true peak You should analyze some of your favorite recordings of the piece and see where the loudness lands. Perhaps with those high notes in the Varese do skew the loudness that much? Tidal would suggest that your loudest piece should hit between -18 and -20 LUFS integrated. Or, let the whole album play out and then adjust. Don’t worry about it until you get your entire album lined up. I will say that for the actual recording you should set your preamp gain to your absolute loudest section on the album. Align that with -12 dB peaks and you’ll be all set and comfortable knowing that even if you play with extra passion you won’t be near clipping.
I think there’s no distortion in those few high notes, and there won’t be many more like that in the other pieces. I added +2dB on the left channel and +4dB on the right channel, I think the maximum peak was in real -6dB. If I adjust to -12 maybe the rest of the album is going to be very low and I’m going to have to raise everything.
(That was the piece where I used a bit of compression in the other version)
I have a Peter-Lukas Graf CD, its version is one of our reference versions, analyzing it reports the following LUFs data:
- Intg: -13,34
- STM: -5,86
- MM: -3,51
- Int: -13,86
- STM -6,93
- MM -3, 56
OK, that’s better! Still no need to be recording that loud in 24-bit…
I want to be clear…I’m only talking about the tracking/capture of the raw audio. After that you can raise the volume to appropriate levels as part of your editing/mastering. We won’t know for sure what that should be until your whole album is recorded. Another trick is to use an SPL meter and figure out how loud those notes are in the Varese (you might need an assistant ). People like to listen to classical music at natural SPL levels they would experience in real life. For audiophile recordings, Bob Katz used to set -20 dB/RMS as his forte which means that about -16 dB/RMS on that scale is fortissimo (loudest an orchestra or big band might play). Note this is RMS and not peak readings (which can get much higher). Calibrated monitors are a key component of this although at a pinch you can just read the meters. Of course we all have volume knobs on our music players but we’re aiming for consistency and not shocking someone when a solo flute sounds louder than an orchestra
Now we have LUFS Katz has said k-meters are no longer necessary but for classical I still believe thinking about SPL, forte/fortissimo is extremely useful.
I think it’s my problem and my horrible English, I’m sorry…
I did the tracking/capture of the raw audio at -12dBs, and then I added +2dB on the left channel and +4dB on the right channel, so that the process would be identical to what I did with Syrinx.
Syrinx Lufs levels were very low:
-24,47 Integrated / STM -19,18 / MM -15,37
You never have to apologize for your English. I don’t speak any Spanish (a little French here and there) but I can appreciate that this isn’t going to be easy given the use of audio terms that probably don’t translate very well.
If you tracked at -12 dB peaks, how did you end up at -1 dB if you only added +2 and +4 dB? Maybe the difference between regular peak and true peak perhaps? Anyhow, doesn’t really matter. Syrinx isn’t low in the grand scheme of things. Plenty of orchestral recordings end up there or lower on certain quiet movements. Again, don’t worry about this too much right now. Record all of your album first using the exact same preamp settings all through the process. We can then talk about appropriate final levels and whether any dynamic adjustments need to be made.
Any other classical workflow questions on your mind?
I called Sweetwater this morning and they are shipping me a new set of the KSM141s after inspecting them first. Hopefully this will solve everything. The microphones seem lovely and sturdy and should do very well on distance recording of classical ensembles in tandem with my existing microphone collection.
I asked a colleague yesterday about a decent (but not super expensive) microphone pair to record flute and she seconded my idea of the sE8 pair (essentially a KM184 clone). Given that the first thing you should probably do is have matched or same model microphones, these would be relatively inexpensive and produce great results. She also recommended whatever microphones you use to experiment with ORTF plus an extra spot microphone aimed at the end of the instrument either in front or directly in line with the body of the flute. In a good room you might be able to get away with just the ORTF pair.
This raises a question: I usually aim for -18dBfs when tracking, should I be aiming higher?
One note on distortion: I don’t know where your mics are placed, but I’ve had some problems with air from the flute causing distortion. I normally place the mics up above the flute to be sure the airstream goes well underneath them. Years ago in the only classical-style recording setup I’ve played in, I played with a harpist and the mics were set up quite a large distance away (we were recorded by some classical recording engineers from WGHB radio in Boston) but only slightly above my eye level. At one point I must have lifted my head a little and the compressed airstream leaving the embouchure hole traveled straight across the room right into the mic and you can hear a moment of distortion in the recording; I was surprised they used it.
-18 dB peaks are fine according to theory! I believe you’d have to go lower than -40 dB peaks to be the same or worse off than 16-bit? The only question is whether hitting preamps that low gives you the best sound. For any decent ensemble that rehearsal around about the same levels as they perform, I’ve found -12 dB peaks work for me. As I mentioned previously, a lot of opera recording engineers aim quite low as there are plenty more unknowns. This is why for location recording I’m thrilled with the new MixPre
Nice, thanks, I’ll give this a try next time. These days I’m mostly accompanying a singer and she does have a tendency (which is common, of course) to sing much more loudly when recording or performing than when doing a level-check, but I’ve been perhaps too conservative.