I am a total Newby to Audio editing. But I love it to make ORTF stereo recordings of classical concerts with my two AKG P 220 mics (or Rhode Nt 5) and my Lexicon Omega Audio interface.
The only problem is: I don’t really know what is the best way to mix/master these recordings. I think mixing is not necessary because I only have one stereo channel and not more channels that have to mixdown to stereo. Or am I wrong here?
I saw a tutorial about mastering and tried to follow this for editing my stereo recording. So at the moment I use the following effects (mostly Calf Plugins):
DC-offset with SWH Plugin
Normalize the audio and amplify it to -0,3 db
a first, light limiter-operation with Calf Limiter (attentuation under 2db)
Equalizing
Compressor (Multiband is too hard for me at the moment)
Add tube warmth with Invada Tube Distortion (Do you know a better plugin? TAP doesn’t please me very much.)
A little bit exciting with Calf Exciter
A gentle Ambience reverb
Then HighPass and LowPass Filter
And last again limiting with Calf limiter
Is this a good way to go? Do you know a tutorial about editing stereo recordings? What could I do better?
You can find a first try with a piano solo piece here:
In my opinion you don’t need any of the processing above. Compression is particularly harmful for classical music because it kills all of the nuances and musical expression. If you need to reduce dynamic range, better use gain automation. I guess some limiting (sparingly!), a nice convolution reverb, and very little EQ (only for room resonance, not for enhancing sound) could pass. All you need to do is put two nt5s in ORTF configuration placed half a meter outside piano box pointed towards open lid (if you can I suggest you buy omni capsules for nt5 for AB mic placement). No need for mixing, left mic goes to left speaker, right mic to the right speaker.
Thank you very very much for your tipps and suggestions. The Chopin was recorded with my AKG mics and a similar placement as you recommended
Do you concretely find that the dynamic range is too limited through the compressors? Please feel free, to criticise my recording. This is the only way to learn! I tried very few settings with all plugins but a little bit the dynamic has to be reduced, I think. Thank you very much for your tip with the automation. The next time I will try this…
By the way: Do you think that a compressor is also harmful for a classical singer, too. Here I liked the smoothing the compressor made.
Do you have a tip for a good convolution reverb in Linux. The Calf reverb plugin seems not to offer such a reverb type…
In this example I cut with the EQ some frequencies at 500 Hz because the natural reverb in the room blurred the piano sound too much. Additionally I raised a little bit the high frequencies above 5000 Hz.
Cutting the room resonances was almost impossible because there were too much of them for my ear. I think a piano has naturally very much resonances? Perhaps I have to exercise my ear a more to distinguish what a ressonance is and what is normal…
Again thank you very much for your suggestions…
Max
First of all I think your playing is beautiful (and I am not just a classical music lover, I am full time employed at philharmonic orchestra as a bassoonist and I listen to the world class pianists at least ten days per month) so I can neglect technical issues and enjoy the interpretation. But if you want my purely technical opinion, then yes, I think compressor is harmful. if you listen for example from 15-22 sec or a chord at 3:04 when the transient is faster than attack time and the sound is abruptly attenuated immediately after thus making a sort of a “pumping” sound… as opposed to quiet parts below compressor treshold which are stunningly beautiful. And to my ears EQ is largely unnecessary. Lift above 5000Hz makes the sound unnaturally bright (I guess no one would like to listen Chopin very close to the instrument.) and I miss some bass. Don’t get me wrong, I don’t think your recording is bad (or crap or rubbish as some people describe things they don’t like), I just think It would be much more beautiful, and more intimate, less aggressive (more Chopin-like) without that much processing. If you wish you could post raw recording, I am curious.
I know some people apply compressor to opera singers but with cautious settings. For example treshold = -15, ratio = 1:1,2, attack = 15 ms, release =250 ms (not as a strict rule, of course, it all depends on many factors). That would be hardly noticeable, but helpful.
As for convolution reverb, on Debian based systems there is a package on repositories called ir.lv2 by Tomas Szilagy. Impulse responses can be found around internet, some places are suggested in the package description ( apt show ir.lv2 ).
Room resonances can be easily calculated: frequency = speed / wavelength. So if two parallel walls are 4 meters apart then 340m/s (approx speed of sound) / 4m (length of a standing wave) = 85 Hz which is roughly e or f (mi -fa) two octaves below middle c (and probably the first harmonic 170 Hz). The same formula applies to room length and height. Those notes should sound “woofy” and should be attenuated with narrow q EQ by arround 4 dB.
Thank you so much for your detailled analysis. This really helps me so much. I see that I have to train my ears more (for example I am at the moment not sensitised for the pumping of the compressor). On weekend I hopefully can look more into your posting (I am at work under the week).
But yet today the raw audio file: https://we.tl/sAne4jRFJO
To my ears raw recording is much more beautiful then the processed one.
If I you want my analysis… when I listen to the treble line (right hand) it is much more pronounced then the bass (left hand) - as it should be, but in expression, not in tonal quality - and distinctly centered in the stereo image, while bass comes more from the right. It seems to me that you placed microphones somewhere in the line with mallets or some 10-15 cm to the right.
Now, you will have to experiment with the microphone placement, but I think that if you move microphones around 1 meter to the right (towards the tail of the piano) and perhaps another 20-30 cm away from the piano (to get mellower sound, cardioid microphones tend to focus sound a lot, omnis are much better for classical music) you will achieve much more balanced and musical sound, no digital magic necessary. That is just a small adjustment, otherwise I like the raw recording a lot.
I just want to say thank you again. Your input really was a guide star and so important for me as a beginner!! Now I hear, that my processings was more harmful than useful, too.
I have deleted all compressor and with the eq I only cut ressonance frequencies (I found only one at about 262, 515, 787; cutting was not really necessary I think). It is still too loud, but I work on it. Nevertheless a little bit reducing of the dynamic range I want to achieve (the audio is for a video), even if it is not really necessary I think (I read that the RMS level in classical music shall be -18dbfs at loud parts. This has already the raw recording)
Where do you get all your knowledge? Do you have any tip for a book about recording and editing classical music? I saw that in the german book “getting pro” are some hints for classical music. But apart from that every literature is only concerned with Pop music etc.
Best wishes,
Max
PS.: Oh yes, the mic position was a little bit stressful. We forgot a cable and did not have enough time for the positioning before the concert. But thank you for your hints. I saw in the book “getting pro” additionally different possibilities…
PSS.: Did you ever use parallel compression? I read that this is used in classical music, too. Perhaps for a opera singer this could be a option, too?
I am glad I could help.
I started off just like you. I was a bassoon student looking for a cheap way to record my own recitals and exams so I bought an affordable audio interface (at that time very popular E-Mu 0404USB. Still working…) and two microphones (rode nt5). My WinXP installation kept crushing every six month or so, besides, antivirus software used to slow down machine significantly (at that time 1.8GHz single core processor) so I switched to Ubuntu (soon after that, Debian) and Ardour (at the time version 2.7). Luckily the drivers for the E-Mu card were already available.
As for learning the basic knowledge of operating a DAW and learning the functionality of plugins, I enrolled to an online course in audio production (Audiomasterclass.com by David Mellor). I totally recommend it although there are probably many others just as good.
As you can perceive yourself they barely touch the issue of recording classical music. So i search the internet myself and found pieces of puzzle all over the place. One of the best examples would be this article:
but there are many others (look for “recording chamber music” at gearslutz.com and similar forums a hint: you don’t need to buy extremely expensive gear as gearslutz suggest). The best source of knowledge for you would be to work with an audio engineer who is experienced at classical recordings. You can offer your expertise: reading scores, knowledge of the piano literature, sensitivity to pitch, intonation, balance, etc. I profited the most in a role of coproducer in a Cello/Piano CD recording project. My task was to read the score and mark all the mistakes to facilitate further edits, but I watched closely to all the moves of audio engineer and leaned a lot in the process. The other way to learn a lot is to get acquainted with a circle of audio engineers. You can listen to each others recordings and projects an comment them in a constructive way… there are lots of opportunities to learn.
I don’t have any experience with parallel compression, but it seems to be a good way of fine tuning tonal quality of a track.
Too many how-to articles on mastering are geared towards much heavier music. For something like this a much lighter touch is needed. Clarity, ambience, and especially keeping everything balanced spatialy now that most people consume music with earbuds are really important. You will also do battle with the room noise with recordings like this. There is simply no way to mask it like you can with louder genre of music.
For stuff like this you want light multiband compression to keep the dynamic range in check, but don’t crush it like people do with heavier content. Use harmonics to fill in stuff that gets lost on the high and low ends. Level out any EQ bias in the mid range areas. Try to widen out the spatial image when it’s a stereo recording so it doesn’t sound like it’s coming from inside your head. And add light ambience, which is not quite the same as reverb: you want it to sound like the early scattered reflections you get in a live space which is much more subtle than reverb.
I probably over did it a bit, but A/B this with your original to pick out the differences. (Adjust for volume differences between the two when comparing. Volume differences when doing A/B comparisons can be deceiving.)
