So, I'm working on this piece:

https://files.catbox.moe/18ek7n.flac

This is just the 1st Phrase from Shostakovich's 5. It's bone dry with just Tree Mic (I can add Close, AB, Surround). How does one go about mixing or mastering this.

Few Ideas I'm thinking of:

1. Adding some vinyl noise crackle to the first bar to make the first notes not come just from nowhere.
2. Pass the stems through Acustica or Nebula Pres.
3. Add Reverb. [Probably try mics especially surround ones before this]

It sounds quite artificial, I'd like to give it some naturalness. Any of you fellas worked in classical before? All help appreciated.

I just added Ozone & some Reverb and Panning to fuck around and it gave out this:
[x] Server Error when uploading, will post later...

Damn, it makes things loud...a proper chain can surpass it without doubt, but this took me couple seconds.

 

Last edited: Oct 17, 2024

I haven't worked on classical projects per-se, so take my opinion with that in mind.
To me, the sounds are already wet in the sense that I can hear reverberation pretty clearly. I don't see how adding saturation would help it.
What I think makes it sound "artificial" are probably velocities that are too consistent, pitches all in tune and everything right on the grid (just speculating that these are samples). Also in an example from a real orchestrata it seems like I can somewhat hear the lower notes resonating amongst the different "airboxes" on stage:

 

Last edited: Oct 12, 2024

If you don't like it, it might be better to use another high-quality sound library. So I would like better source material.

 

Last edited: Oct 13, 2024

Is this Berlin strings? I agree in that I'm not sure I would start resorting to processing at this point, I think there could be more improvement in the programing itself. Case in point, looking at the score, and you can clearly hear it in the performance, there are accents on the down beat

 

Last edited: Oct 12, 2024

Lol! I thought the thread died! Anyway I didn't get any notifications.

Yeah! I improved the programming but upon comparing it with recordings there's just something magical in the recording; in the thundering basses, flowing textures, etc. that just isn't possible to reproduce using these samples. As for other libraries, I don't know any other than VSL or Berlin that have enough articulations even to make a thing like this happen. Also, I thought these things come pre-panned but Berlin surely needs some panning.

Nah, I used a Tempo Map.

 

Last edited: Oct 17, 2024

First of all bravo for your choice. Great piece. I have recorded some classical but much more baroque. I have never programmed classical pieces though. It is a very difficult endeavor hehe. Errr so here are my thoughts.
1. You need more players if you are aiming to re-produce the symphonic piece. As many discrete channels as possible for realism. Use all mics available.
2. Pan/position them as they would sit on stage.
3. Much more dynamics please.
4. Big reverb and a lot of it (lol). Altiverb or other convolution will do, but Altiverb has a wealthy choice of symphonic halls.
Good luck
PS: Whatever you do, do not compress anything whatsoever. You need every bit of dynamics you can get. A bit of subtle eq (~2db) could work nicely especially on double basses and cellos.

 

Agreed with @taskforce, if you're using Berlin you have to double up the sections with the transpose trick. Berlin has small section sizes (8/6/5/5/4). And definitely use all the mics.

I am not afraid of compression on strings, but I only compress (independent LR) the attacks on shorts (and some accented longs like espressivo, but with a lighter touch) on the close mics at mf and above. It's not a recommendation however, not many are as neurotic as me.

 

Last edited: Oct 17, 2024

As it is from 1937, you have to add 78rpm gramophone record noise and crackle to be sure it's how Shostakovich intended it to sound.

 

Overall Longer Attacks on compression (especially at bus level), go at the most half as hard as you might think on any processing you plan to do in general. Remember, back then the goal was solely to capture the "performance" as was intended and as was performed as clean as possible. Especially with the limited resources (and obviously track amount capability), this naturally created mixes that were the engineer was more so "out of the way" on just by default.

 

In addition to many of the excellent ideas listed here... the last thing I would do would be to add crackle or something that makes it sound like you just did a vinyl rip; it's unpleasant to hear and probably isn't really going to help, but what you can do is go out and find a good room tone sample. They're available lots of places, just Google. Put a whiff of that in the background and it will solve your digital "coming out of nowhere" problem and help tie the whole thing together.

 

no need for anything fancy, less is more,
things to keep in mind:
phase coherence across tracks, sensible denoising, natural reverb, don't hesitate to disable tracks you don't need (like spot mics),
mastering should be done in accordance to target platform, focusing on good compromise between rich dynamic range and acceptable loudness

 

I interned under Dan Wallin - engineer for many famous scores. He would set everything up so that it was mixed straight to tape (in this case they actually tracked to 4 different sources all in sync. a nightmare lol). The processing he had for an entire orchestra was so minimal it's unbelievable.

Pair of pultecs doing a smiley, couple UA comps's doing some very minor compression, Fairchild pair. i thinkkkk he would use neve pre's with eq engaged doing maybe 1-3db cuts and boost, GML patched in. Thats it. maybe one or two more things, but the point is, it was all about mic placement, and less about processing. He would spend a good 30 minutes setting up 8 or so mics. getting up on ladders and adjusting them from a birds eye view.

If you have the ability, try experimenting with placing the parts of the orchestra in the room. One of those panning plugins that fakes a room placement / depth. Besides that, yah, like you mentioned, Nebula or old acqua plugins. Brown would be good. El Rey comp on master too.

 

I think OP was talking about samples, not a recorded performance, that was played under specific direction to best realize the composer's intentions. Not to say your post doesn't offer any insight, it does and thank you, but I think context is important here.

 

ah i see. its just one library being used. Ryu, if there is a way to separate the elements of the orcehstra, perhaps by duplicating the library and soloing / muting certain elements, you can try messing about with artificial placement of the players.

 

I'm working on a classical orchestra performance (in a church) 4 mics recording (2xORTF + 2xAB) mixing/mastering...

...just one word: EQs.

 

I've seen it done using MIR 3D, haven't got my hands on it yet. I use A1 Stereo Control on stems of each section to pan these [it completely lost it's balance after panning and required gain-staging], passing through 13-15% Altiverb Berlin Philharmonic in Serial with 2C-B2.

As for doubling, I already layered multiple articulations to get the sound...there's already more than 2x players playing the phrase to make up for the small section size...I'm constantly A/B-ing with a reference recording to get the balance right, I realized it's not gonna be completely realistic...though I bet a normal ear would distinguish it [It's realistic at a good volume but when listened at low volume, it comes very artificial]...no wonder you never hear anything other than dumb sustains/adagios in most orchestral mock-ups.

I'll process more after I've orchestrated at least a decent chunk. I just wanted to test from a small phrase that if some mixing can do it good, it'll be a waste of time to just orchestrate and realize some things are better solved with processing...

I looked into Room Tone...will try placing it in silent chunks...

 

Last edited: Oct 23, 2024 at 10:58 AM

Yeah, try to emulate whole console path, probably some AlexB one, have a feeling you will dig the sound of AlexB Modern Flagship Console Golden Edition with azzimov skin.

 

Can you post an updated version? I just went back and listened to your first phrases and compared with the youtube excerpt I posted and I noticed that your version seems a lot slower and the attack less "whipped" or aggressive, as if the first note on every phrase was prolonged in yours. The reference video sounds more like TA-DAAANNN with a staccato first note whereas yours is more like taaadaaaAANN kind of leaning into legato, it if makes any sense.

 

I never listened to Tilson Thomas's version. The one I always listen to is by Kurt Sanderling, I'lm also familiar with Rostropovich & Mravinsky recordings.

 

This one for example?

Even when comparing with this one, I still think that in the very first phrase (right at 2 to 3 seconds) your violins (violas?) are still dragging the first note on each phrase. Also, try cutting around 630 Hz, to increase the perceived brightness, I was just trying out some settings on my mixer comparing your flac file to the youtube recording and that EQ cut seemed to approximate the feeling a bit more.