Inspired by Complete Classical Music workflow thread, I wanted to share some new recordings from the concert that took place on this Sunday, 11/10/19, directly from the oven, so to speak. I hope further discussion on mic placement, pick up patterns, recording venues, plugins, editing options, etc. may develop from this couple of real-world examples.
Recording took place in small town of Kutina, Croatia, in a local Cultural center hall (a combination of cinema and theater hall) with surprisingly nice acoustic.
The ansamble consists of flute, oboe, clarinet and bassoon in first example and flute 2 oboes, 2 horns, 2 bassoons, and 2 clarinets in the second.
This is the third movement, Scherzo, from the wind quartet by Zoran JuraniÄ. I didnāt specify composer or title because I shared it solely as basis for audio technical discussion, without asking for permission for publishing it. In unlikely event of anyone wanting to use it for a commercial purposes, be warned that it is copyrighted.
This is the third movement from Gounodās Petite Symphonie for winds.
Recording gear is as simple as possible: a pair of Rodes NT5 with omni capsules, into E-Mu0404USB to Raspberry Pi Zero (headless, controled by Huawei phone), imported to Ardour 5.12 on Dell Inspiron i3 laptop running Debian 9 / custom fluxbox edition
I served both as audio engineer and the bassoonist/performer - nerve-wracking combination - I advise to avoid it if you can.
Very enjoyable to listen to. The recording chain doesnāt get in the way of the performance at all. It is refreshing to hear live music like that. I personally have heard enough āsolo vocal in front of a computer generated karaoke trackā music that seems to be the normal thing on the radio these days.
Yes, this is a lovely acoustic and very well recorded (and played)! You say you advise to avoid serving multiple roles and I concur. It doesnāt stop me doing it and also getting a huge thrill when everything goes well. I need a stiff drink afterwards though.
The Rode NT5 omnis prove that for a very good price, excellent results can happen (particularly fine at picking up the clarinet timbre). I feel the same about @Alephās new flute recordings with his sE8 pair (more on that when I post in that topic). My one request is I wonder whether you can share even just a snippet of the lossless versions. Soundcloud, as good as these recording are, has a tendency to use less-than-ideal bitrates for their lossy codecs. Itās why if Orfium actually fixed the upload issue I had, I will use that due to the default 320kbps streaming rate. Thereās even talk of allowing lossless downloading (like Bandcamp).
Anyhow, these really are excellent. Perhaps for a āstudioā recording Iād add spots to each instrument (or pair in the nonet) but I probably wouldnāt use much of it as the spaced pair is so good. Could the bassoon benefit from a little bit more presence? The recording seems to use a very good amount of the stereo spread and has the intimacy that I would associate with such a chamber performance. A smidge of hall reverb might be a nice addition to further sweeten the space but Iām sure youāll get to this given this is ārawā. Do you ever record the rehearsal so that you can patch sections of the concert that you are not happy with? JEG did it to marvelous effect for his Bach Cantata Pilgrimage series. Some of the tuning in the early concerts was horrendous but fixed by pre- and post-concert rehearsal patching.
Iād be interested to know the spacing you used, distance from ensemble and the spacing of the players. Also, the Pi headless is just plain cool
Good chamber group, are you all members of the orchestra?
Itās not that I want to please you for free, but you play very well @vaskq. That bassoon sounds very good.
It sounds incredibly good, but maybe a few individual mics and try a little more reverb, as @anon60445789 says for a studio recording, would be nice.
With a similar configuration I recorded years ago this wind quintet CD:
Hopefully this doesnāt take us off topic but this recording has a pleasing naturally ācompressedā sound and a kind of analog warmth. Reminds of why I enjoy Kurt Sanderlingās recordings of Mahler. Do you know what equipment was used?
Indeed, Soundcloud upload really takes some quality away. Here, I uploaded originals on Google drive, I hope it will be easily accessible:
I am not huge fan of spot mics, although I have to admit they may make life much easier and I would sure use as much of those as needed should someone pay me for recording them. However in my performer/engineer scenario it would be logistically very cumbersome to have up to hundred meters of cables, lots of stands, and the recording gear in the backstage unattended. I would probably need an assistant, which is currently not possible. And then, any time I actually used spot mics I had trouble getting the sound I heard from the audience. The reason is probably the fact that with woodwind instruments sound exits from the holes all around the body and it forms the way we like to hear it at least 1.5-2 meters away. With spot one needs to decide which part of the instrument to āamplifyā relative to the others - that comes into play more with bassoons than high woodwinds. There is workaround for that too, but with properly placed stereo pair you can have it all with just a pair of mics. Of course, real world situations come in the wayā¦ in this particular concert we had a piece with solo clarinet so I put the mikes much further back than I wanted so the bassoons in the bottom of the āhorseshoeā formation sounded the quietestā¦plus the fact that we played some modern pieces with extreme piano at places (A.Part - Fratres) and we decided to use softer reeds.
Actually I already added some very subtle reverb to fill empty space at rests (ir.lv2 with āGreat Hallā impulse @-30dB). The venue itself was ābeautifully dryā (meaning lacking room modes). I will experiment with some more, but lately i am more inclined to drier recordings because reverberant recordings tend to sound good in some listening environments and not-so-good in the others.
Yes! I always do that. The nonet recording is also mildly edited. The part of the problem is that we rarely rehearse all of the pieces at the venue because we spare energy for the concert, but this pieces we do play through are invaluable to me as the patching material.
I donāt know exact measurements. The mics are separated some 45 cm. This is a picture of our setup:
Thank you!
Clarinetists are music school teachers and the rest of us are all members of orchestras: first oboe and both bassoons from Zagreb Philarmonic Orchestra, second oboe, second horn and contrabass (not on this recording) from Croatian Radio Symphony and the first horn and flute are from Opera house in Zagreb. It is always a problem to arrange rehearsals with the people having regular jobs at different times.
Thank you so much, it means a lot to me.
This is fantastic playing and recording. Very interesting music, too. Thanks for sharing!
Please, feel free to post any track you may find interesting for reference or comparison. Also, feel free to comment and ask questions about any of those.
This is a good point. Reverb is a very personal decision too. I donāt think thereās a right and wrong way. You can always experiment with dry, a bit of added reverb and a little more than you are usually comfortable with and have a vote
Reverb is certainly an interesting topic for discussion. I am somewhat limited in choice because I only have what is available for Linux without access to fancy Bricasti or Lexicon stuff. Of all commercial plugins I only bought Mikeās SR2 which can be tuned nicely and I used it a lot for some non-classical stuff (got paid for this so the investment paid off ), but it is not always my go to reverb. I use Tomās ir.lv2 reverb most of the time and sometimes Fonsā Zita-rev. I have yet to explore Dragonfly reverb and LSP Room Builder. Good thing about Ardour is that one can stack different reverbs and a/b/c them easily to decide which fits the best.
For me, there some issues one needs to be aware of when applying reverb to classical music recorded on a acoustically non treated location.
reverberation of the recording space sometimes clashes badly with digital reverb so when selecting type of reverb one should be careful to match digital and natural reverb.
Some venues, especially smaller ones have distinct room modes or standing waves (the sound wooes on certain notes/frequencies). Those should be detected and EQed before applying reverb.
In my experience digital reverb affects different instrument colors and dynamic in a different way. That is especially evident with woodwind ensembles because each instrument has its distinct color (flute has very few overtones while the bassoon sound is all overtones with hardly noticeable fundamental tone). And so when I apply reverb it sometimes make one instrument sound beautiful and the other sound like the sound coming from a tin can. I am not sure why some reverbs mess up the frequency response but that may have something to do with the fact that the algorithm treats all frequencies equaly while in natural spaces high frequencies vanish more quicklyā¦ or is there some other explanation? The same goes for dynamic, some reverbs add the same 3 sec tail to the piano and forte notesā¦
To overcome some of this issues I sometimes create additional reverb bus with 100% wet reverb which I then EQ to my taste (so I stack the reverb and then EQ in the bus plugin section). I then create a send in the Master bus (or the track I want the reverb to be applied to) and decide the amount of reverb with the send volume.
Anyway thank you for your feedback, I will certainly try various reverbs and decide which one āgluesā the music the best and Iāll post the results, hopefully by the middle of the next week. Unfortunately, around here it is practice with professional sound engineers to apply some crystal-hall-4-sec-tail reverb and call it a day. I would like to offer some tastier alternative, so I am giving it much thought.
Please make a comparison between the reveb of Robin Gareus LV2 Convolution Stereo and the IR of Tom Szilayi.
With the same Vienna quad file (from Bricasti) Robinās reverb is clearly more natural.
In my experience, Tomās reverb affects the color of the soundā¦
This might belong in the ācomplete classical music workflowā thread, but Iām curious about how you manage these recordings for patching, and whether you do it differently if youāre using stereo pairs versus using spot mics.
Do you set up alternative tracks and playlists for the rehearsal takes? For example, if Iām recording multiple takes or rehearsals of a duo or small ensemble with spot mics, I set up alternative tracks for each instrument or voice and within each track I have multiple playlists, one for each recording. That works well, although I havenāt yet discovered an efficient way to mark passages in individual playlists when Iām reviewing them. The NLE I use for video, Davinci Resolve, offers both timeline markers and clip markers, which means you can not only create markers for your session but you can also make markers for individual regions that stay with those regions if you change playlists or move the region around. I love that feature. Does Ardour have anything similar? I know you can save a region to the region list, so maybe thatās the best way but Iād love to hear how others approach the task of reviewing and choosing passages from multiple recordings for patching into your master recording.
I mostly do live recordings in which there is the performance and only one rehearsal so I have developed rather simple workflow. I place the performance first and then rehearsal material for patching next, linearly down the timeline. After that it is all about putting some meaningful markers and rename them to make navigation easy. For example, I mark performance movements/songs with range markers, rehearsal songs with regular markers renamed with the song title, and the places I want to edit with the regular markers containing numbers only. I suppose if I had 10-15 takes for each song iād have to think of something else, that would be easier to navigate. Iād probably stack all the regions containing the same material in the same track on the same place in the timeline, rename each of those the way it would be easy to recognize (probably name of the song and take number) and then Iād think of some convenient shortcut to āChoose Topā¦ā which essentially creates a sort of playlist-dialog. As for spot mics, I just put all of the tracks in the same group - that allows me to edit multiple vertical tracks the same way I edit only one.
I have trialed something similar and certainly this was my workflow in Sequoia placing the preferred performance followed by rehearsal material and/or other concert performance nights. My current favorite way is to listen back either looking at a score or closing my eyes and just hit TAB every time I hear something that I think needs to be fixed (and labelled with a pleasing yellow āwarningā color). Then it is a case of auditioning all the other potential patches. The actual editing is still far too many clicks ānā drags compared to good olā Source-Destination but I am learning to live with it (so is my wallet!).
The playlist is a great idea. I generally record everything in a row versus starting a new playlist or layer but worth experimenting with in the future.
Good call on the grouping of tracks, too. Ripple editing is much more fun when regions across all the tracks move as one
EDIT: One aspect that Iām still getting used to in Ardour coming from Sequoia is that CD markers do not stick to the regions when the regions are moved. However, Iām beginning to think that Ardourās behavior is to be preferred in the long run. It would be fun to be able to change CD marker order in the locations window and bring the audio with them (like in Sequoia and Pyramix).
Recording everything in a row is indeed the typical approach for classical music, at least based on everything Iāve heard and read, and of course it can lead to sessions that are many hours long. The big advantage to that approach over what I do (using playlists) is the ability to mark things. With my approach, when I switch playlists the markers I have in my session no longer make sense because they refer to a previous playlist. So I think Iāll switch to the standard classical approach of one long session with multiple takes, using markers to navigate through them.
The biggest challenge I have with patching is dealing with microphone bleed. Itās not an issue if youāre replacing everything (e.g., if you recorded using a stereo pair with no spot mics), but my current project is recording a singer accompanied by guitar, with spot mics, in the same room. If I replace a phrase in the vocal track, thereās so much vocal bleed in the guitar track that a bit of the original vocal is still audible. I tried using Izotope RXās de-bleed tool on the guitar track (to attenuate the vocal bleed) but wasnāt satisfied. Of course I could simply patch in the corresponding guitar take on the guitar track, but itās trickier to do that seamlessly than it is with the vocal because each take was at a slightly different tempo and the guitar has a lot of sustained notes.
Agreed that it can be tricky. If you group the tracks or do a mixdown of the vocal and guitar before you edit, Iāve found thereās always a way to get what you want. You might have to replace a slightly longer section and/or play around with the fade type and length. Itās difficult to know without hearing/seeing an example so I suppose we could start a new topic to investigate? If your takes are at different tempi it is another reason to treat them as grouped. Sustained notes by themselves shouldnāt be too much of a problem if my pipe organ editing experiences are anything to go by. It can just take a little longer to find the magic fade type/length combination.
Good idea, but let me try first and see how it goes; I might be able to make it work. The tempi are no dramatically different. If I get stuck and canāt make it sound convincing I will start a new thread. Thanks! There is one phrase in our favorite take where the singerās voice warbled a bit, and itās a loud passage so her voice bled quite a bit into the guitar track. So the best solution is to replace that section of the guitar track with the corresponding guitar track from the alternative track from which I replaced her vocal on that phrase, assuming Iām happy with how the guitar was played in that take.
