Is It Impossible To Make Good Midi Renditions Of Classical Music?

[mdiemer]

I'm coming to the conclusion that it is. I work my butt off for months, make a wave or mp3 file that sounds decent on my computer speakers, but when I play it on anything else, it sounds terrible. I know you're supposed to do it so it sounds good on anything, but how does one do that? If I make a mix that works on my M-Audio-Mackie-sub setup, how is it going to sound good on my big tower stereo system speakers? Or bookshelf speakers? No one ever tells you how to do that;  just that you have to do it.

The vast majority of mixing and mastering tutorials are for non-classical music. I've done many. I've tried many approaches over the years. I still keep producing tinny, flat mixes. My samples are pretty good: Vienna Special Edition, Vienna Appasionata Strings; Cinematic Strings 2; EWQL Sym. Orch. Gold. Plus some Garritan and Dim Pro stuff. I have a UR-22 audio interface from Steinberg. I know that's not the best, but for pure midi why would I need anything  better? I am not recording any audio whatsoever.

I'm considering giving up and just using notation software from here on in. Maybe I'll decide to pay someone to make a recording that actually sounds good of my music. 

My system is an i7 with 32 BG ram. That's not the problem. Maybe my hearing is going, but then I never made any decent mixes even 20 years ago.

Keep in mind, I'm doing purely orchestral classical music, many instruments in the mix. Here's an example of what I'm up against: I know you're supposed to use more reverb to simulate depth. OK, so you have a string group, Violins 1, Violins 2, Violas, Cellos and Bass. So, the violins are in front. so, put more reverb on the others. but how much? No one ever tells you that. Is it a certain percentage? Is it that I need to just listen and wing it? I can't trust my hearing anymore. 

Then there's panning. I've spent dozens of hours trying various schemes. Also tutorials. but accurately simulating a full classical orchestra on a computer with some software just doesn't seem to work. 

Guess I'm just venting. I don't really expect a solution. Probably the best advice would be to just give up. Figure out some other way to spend my time. Maybe go back to just writing poetry. At least I have the computer skills needed for that...

 

 

 

 

 

Edited April 15, 2020 by mdiemer

[DeeringAmps]

1 hour ago, mdiemer said:

Probably the best worst advice would be to just give up

Maybe you should listen some of Jerry Gerber's work.

He has several in the songs forum right now.

Here's the thread for his Symphony #10

Often he makes the score for his works available as well.

IIRC he also teaches virtual orchestration.

t

[Brian Walton]

1 hour ago, mdiemer said:

I'm coming to the conclusion that it is. I work my butt off for months, make a wave or mp3 file that sounds decent on my computer speakers, but when I play it on anything else, it sounds terrible. I know you're supposed to do it so it sounds good on anything, but how does one do that? If I make a mix that works on my M-Audio-Mackie-sub setup, how is it going to sound good on my big tower stereo system speakers? Or bookshelf speakers? No one ever tells you how to do that;  just that you have to do it.

The vast majority of mixing and mastering tutorials are for non-classical music. I've done many. I've tried many approaches over the years. I still keep producing tinny, flat mixes. My samples are pretty good: Vienna Special Edition, Vienna Appasionata Strings; Cinematic Strings 2; EWQL Sym. Orch. Gold. Plus some Garritan and Dim Pro stuff. I have a UR-22 audio interface from Steinberg. I know that's not the best, but for pure midi why would I need anything  better? I am not recording any audio whatsoever.

I'm considering giving up and just using notation software from here on in. Maybe I'll decide to pay someone to make a recording that actually sounds good of my music. 

My system is an i7 with 32 BG ram. That's not the problem. Maybe my hearing is going, but then I never made any decent mixes even 20 years ago.

Keep in mind, I'm doing purely orchestral classical music, many instruments in the mix. Here's an example of what I'm up against: I know you're supposed to use more reverb to simulate depth. OK, so you have a string group, Violins 1, Violins 2, Violas, Cellos and Bass. So, the violins are in front. so, put more reverb on the others. but how much? No one ever tells you that. Is it a certain percentage? Is it that I need to just listen and wing it? I can't trust my hearing anymore. 

Then there's panning. I've spent dozens of hours trying various schemes. Also tutorials. but accurately simulating a full classical orchestra on a computer with some software just doesn't seem to work. 

Guess I'm just venting. I don't really expect a solution. Probably the best advice would be to just give up. Figure out some other way to spend my time. Maybe go back to just writing poetry. At least I have the computer skills needed for that...

 

 

 

 

 

How about posting an example of your current results.  That might provide some insight.

 

Traditional Classical Recordings are just a Stereo Pair,in the room recorded live.

Having everything in the box as a separate track has its own challenges to get a similar  style of recording.

However, getting a "big" recording should actually be pretty easy.    Translation to other systems is a challenge.    Many mixers will use stereo monitors.  Adding a sub can be deceiving for translation in some other formats. unless you have a great ear and learn how it will translate.  

[mdiemer]

15 hours ago, DeeringAmps said:

Maybe you should listen some of Jerry Gerber's work.

He has several in the songs forum right now.

Here's the thread for his Symphony #10

Often he makes the score for his works available as well.

IIRC he also teaches virtual orchestration.

t

I've heard Jerry's stuff before, even emailed him once or twice, but never really studied how he does it. I'll check out the thread you posted.

[mdiemer]

15 hours ago, Brian Walton said:

How about posting an example of your current results.  That might provide some insight.

 

Traditional Classical Recordings are just a Stereo Pair,in the room recorded live.

Having everything in the box as a separate track has its own challenges to get a similar  style of recording.

However, getting a "big" recording should actually be pretty easy.    Translation to other systems is a challenge.    Many mixers will use stereo monitors.  Adding a sub can be deceiving for translation in some other formats. unless you have a great ear and learn how it will translate.  

OK, you asked for it:

https://app.box.com/s/s2mb484mubt2zxg9o1gq2zke17b9uu6f

It's the first movement of Three Easter Scenes, which I just completed.

Edited April 15, 2020 by mdiemer

[Starise]

Hi mdiemer,

Tim here. Good to hear from you.Maybe you remember me from another forum.

That was a nicely put together orchestration. I listened on my ATH-M50x headphones along with correction by Sonarworks. Wife is working at home and on the phone so I can't use my monitors right now.

I listened in mono/stereo and in bass tilt mode. I took off the -6db ceiling to make sure i could turn it up. The issues I hear are easy to correct....no really. I only hear two issues that jump out at me right away. 

- Not enough bass

-Not loud enough, even for classical music.

The main reasons for this are usually you hear too much bass through your monitor system which causes you to ( falsely) think you need less bass when you really need more bass. The answer to this is a more accurate monitor setup. Possibly using ARC2, or at the least balancing the bass in the space to match the bass you really are hearing on other systems. Just seen an interesting video on the idea that many mixing engineers over use the high pass filter when sometimes it's important to keep bass information  in a lot of tracks people are rolling off, at least some of it.

The loudness issue could be corrected by adding stages of gain in small amounts here and there, then adding a nice limiter to the master. 

The loudness of the monitor system is important too since it should be set to a level that isn't too loud. If it's too loud you'll make the mix too soft. The K-system uses pink noise to get a level set. It's a good system. I don't always use it. What I sometimes do instead is mix to a place where anything past 60% volume on the computer volume control is getting loud and I'm within db, LUFS and K limits for louder material. It won't kill a mix if you go over in small amounts here and there. The best thing though IMHO is to find the few offending tracks and compress/limit them before going to master. 

Feel free to PM me any time on this. 

Edited April 15, 2020 by Starise

[mdiemer]

Hi Tim, of course I recognize you, both from here and the Ning forum. Thanks for all the great suggestions. I'll take them into consideration. I figured I needed to re-set my sub-woofer setup. I have also noticed the lack of bass, but when I'm working it's so overwhelming I have to keep it low. I usually use a the Sonar LP-64 EQ on the master, and increase the bass there some. 

I recently downloaded the free Ozone elements suite, courtesy of Sweetwater.  but it seems like over-kill, and much more designed with loud pop/rock etc in mind. So I have yet to use it. I mean I've tried it, but it just made everything extremely loud.

[craigb]

?

[InstrEd]

Put a Darth Faper head  on the mixing board and all your mixes shall be cured ?

[craigb]

2 hours ago, InstrEd said:

Put a Darth Faper head  on the mixing board and all your mixes shall be cured ?

Darth Faper?

Finding yourself bored with all the free time in isolation there Ed? ?

[InstrEd]

35 minutes ago, craigb said:

Darth Faper?

Finding yourself bored with all the free time in isolation there Ed? ?

I'm taking the 5th on that one ?

[Amicus717]

Hi Michael,

I totally know where you are coming from -- I do only orchestral music; I'm 100% midi; and over the years I have struggled to get decent mixes. 

I figured I'd write up my own approach to recording my symphonic music. For what it's worth, over the past two years I have gotten results that I have generally been happy with, and thought maybe some elements of my approach might be useful. However, I am also an amateur with tons of things still left to learn, so take all this with a grain of salt, too. 

For my stuff, I usually use only one reverb on the mix. I have found Nimbus to the absolute best reverb out there for classical/symphonic stuff. It is easily the most open and airy reverb in my arsenal, and to my ears gives the orchestra some really nice air but doesn't thicken the overall mix too much. I add Nimbus as a send, and give the majority of patches and sampler instances the same general settings -- to my ears it gives a unified sense of space if the instruments all run through the same reverb. I may vary the amount that gets sent to Nimbus from section to section, depending on what I'm writing, but it is never more than a few percentage points. FWIW, I believe Nimbus' smaller brother, PhoenixVerb is on sale for $10 over at the Deals forum -- which is an absurdly low price for a top flight, neutral and clear reverb, and I'd totally recommend you grab it and try it out. I have a ton of reverbs (I have something of a weakness for them), but the only one I fire up these days is Nimbus. I haven't touched the others in years. In my opinion, it is THE software reverb for Orchestral stuff. 

I also put a lot of work into building my orchestral template, and I have deliberately selected instrument patches that inhabit similar sounding acoustic space (or can be made to do so, using various mic mixes). I've also spent hours fine-tuning the template so that each instrument in the entire symphonic assembly sits where I want it to across the acoustic landscape. 

I use a really broad mix of libraries:

1) Orchestral Strings: AlbionONE (sustains, tremolo, short articulations and pizz); 8Dio Anthology Strings' legato patches for dramatic lines; Sonokinetic's Modal Runs; Hollywood Strings Gold runs patches; Sonokinetic Da Capo's legato string patch (for really quiet passages, as I like the sound of their strings when the mod wheel is dialed really low); Palette's Trill patches; NI Symphony Series Ensemble Strings Harmonic patch

2) Woodwinds: VSL solo woodwinds, all from the Synchronized Special Editions; Sonokinetic Ensemble Woodwinds Standard Edition

3) Brass: VSL solo and ensemble brass, all from the Synchronized Special Editions; NI Symphony Series (for low brass)

4) Percussion: NI Symphony Series Percussion (for snares, bass drum, some timpani, tamtams, suspended cymbals, mark tree); VSL Timpanis; Kontakt Library VSL Glockenspiel; Hollywood Percussion Gold (for orchestral chimes); EastWest Storm Drum 2; AlbionONE's Darwin Percussion Easter Island Hits; August Forester Grand Piano from Kontakt Library

5) A various scattering of other libraries: Kontakt Library's VSL Harp, Lacrimosa Choir patches from 8Dio, Mercury Boys Choir Elements, Embertone Recorders, and a handful of Eduardo Taloronte's Era Libraries (Era II Vocal Codex, Dark Era percussion and flutes, some Forest Kingdom flutes, etc). 

If I need to push an instrument back or bring it forward, I usually start with the microphone mix settings built into the library in question, or the reverb built in the sample engine/library. This last point applied mainly to the original VSL Special Edition libraries, which were quite dry and center panned, and needed Vienna Ensemble's built in reverb to push them back into the acoustic space, and Vienna Ensemble's mixer and stereo width slider to position them properly across the acoustic landscape. When VSL released the Synchronized versions of its Special Editions, the reverb and positioning of the default mix for those instruments worked a lot better right out of the box, and they fit into my template's acoustic space with minimal adjustment. 

The only time I ever fire up a second reverb is for some of the Era patches, if I need to give them a particularly distant or washed-out sound (and for that I usually create a second reverb send with Valhalla's Shimmer). There were a few occasions where I tried using Wave's TrueVerb with it's early reflection room/space simulations, and would run various sections through it first in attempt to manipulate their positioning, but I never found it more effective than just whatever settings I could manipulate in the various libraries themselves. 

As far as getting good mixes is concerned, I bought the best monitors I could afford (a pair of Adam F7), spend a LOT of time with them so I have a good sense of their sound and how it translates, and when in doubt I tend to trust the ears of the engineers who made the samples. I rarely do much beyond the slight cutting of frequency here and there -- usually in attempt to clear out some mud and honk from the mix. I almost never add anything via EQ, and I usually do minimal compression unless I have a specific solo instrument that really needs to cut through the mix. And before I compress it, I usually try to clear space for the solo instrument by adjusting the orchestration first. 

When I have a final mix from Cubase (my main DAW for midi work), I usually take the exported track and bring it into Cakewalk and use ProChannel to master it (which for me usually means giving it a bit of boost using ProChannel's Concrete Limiter, and adding ProChannel's Tape Emulation and/or Console Emulation to give it some analog mojo). 

I have no idea if the above things are right or wrong, or outright heresy, but right now I get results I am generally happy with. I've linked to two tracks below, both of which were built using the exact template setup listed above. For both, there is only one instance of Nimbus reverb, very limited EQ, and Cakewalk's ProChannel Concrete Limiter and Tape Emulator over the final mix - basically just four plugins across two DAWs, and absolutely nothing else. To EQ these tracks, I used IK Multimedia's White Channel on the "Cut Boominess" setting, and dialed back the amount it was actually cutting by a fair margin. I find it effectively takes just a bit of thickness and smear out of the low end, and to my ears that seems to go a long way to clearing up the whole mix without gutting the bass.

I find these particular mixes sound great on my monitors, very good on my iPhone earbuds and my bluetooth headphones, acceptable in my car (nothing sounds particularly good in my car, frankly, so its a good test), and the folks I've sent it to have never made a negative comment about the mix quality, so I assume it sounds decent enough on their systems.  

As I mentioned, I do not presume to have all (or any) answers to the various issues that crop up for us folks doing symphonic music on computer. But what I outlined above has worked for me, and I figured I'd offer it up as some potential ideas that may work for you as well.

https://soundcloud.com/amicusaudio/the-soldiers-hornpipe-ver-4

https://soundcloud.com/amicusaudio/a-song-for-the-trillium-queen-adrielles-theme

I hope some of this helps, or at least inspires some different approaches. But for what it's worth, I totally get how you feel, and I've ranted a bit myself on this topic. 

Good luck! And if you want to connect about this stuff -- send me tracks to listen to, bounce ideas back and forth, etc -- please let me know. We orchestral folks need to stick together

Regards,

Rob

Edited April 17, 2020 by Amicus717

[Amicus717]

21 hours ago, InstrEd said:

I'm taking the 5th on that one ?

Don't you mean "I'm taking the Sith on that one"...?

 

Sorry. I'll see myself out...

[JoeGBradford]

This is a really interesting thread. I've only got Garritan PO and am just in the process of upgrading from Miroslav Philharmonik 2 CE to the full 2 thanks to the current IK Total Max 2 deal but there are a lot of interesting tips and ideas here so thanks everyone

[JoeGBradford]

It might be worth watching some of the Spitfire videos, especially the ones about the BBC SO VSTi - I've found those interesting

[mdiemer]

On 4/15/2020 at 9:14 PM, craigb said:

?

Great! Where can I get one of those?

[mdiemer]

On 4/16/2020 at 2:40 PM, InstrEd said:

Put a Darth Faper head  on the mixing board and all your mixes shall be cured ?

If only it was that easy...

I've been doing some research, and one thing I saw was to put the basses closer in, having them too far to the side loses too much signal.

Here's another brick wall I'm running into: I use different orchestral libraries. Most are recorded in-situ or position, so theoretically you don't have to pan them. Unless, you mix them with other libraries that have different seating arrangements. Like, Vienna tends to have the basses on the left; most others have them on the right. There can also be differences with the two violin groups. But moving them out of their "native" positions makes me nervous, as now I'm changing their position in the room they were recorded in. but that's what I do, I can't have the basses sometimes on the left and sometimes on the right.  It seems the more I try to experiment, the more confused I get.

Edited April 17, 2020 by mdiemer

[user 905133]

3 hours ago, Amicus717 said:

I totally know where you are coming from -- I do only orchestral music; I'm 100% midi; and over the years I have struggled to get decent mixes. 

I figured I'd write up my own approach to recording my symphonic music. For what it's worth, over the past two years I have gotten results that I have generally been happy with, and thought maybe some elements of my approach might be useful. However, I am also an amateur with tons of things still left to learn, so take all this with a grain of salt, too. 

Thanks for taking the time to share your sample library choices and your recording/production insights and for the links to your wonderful music. 

 

[mdiemer]

3 hours ago, Amicus717 said:

Hi Michael,

I totally know where you are coming from -- I do only orchestral music; I'm 100% midi; and over the years I have struggled to get decent mixes. 

I figured I'd write up my own approach to recording my symphonic music. For what it's worth, over the past two years I have gotten results that I have generally been happy with, and thought maybe some elements of my approach might be useful. However, I am also an amateur with tons of things still left to learn, so take all this with a grain of salt, too. 

For my stuff, I usually use only one reverb on the mix. I have found Nimbus to the absolute best reverb out there for classical/symphonic stuff. It is easily the most open and airy reverb in my arsenal, and to my ears gives the orchestra some really nice air but doesn't thicken the overall mix too much. I add Nimbus as a send, and give the majority of patches and sampler instances the same general settings -- to my ears it gives a unified sense of space if the instruments all run through the same reverb. I may vary the amount that gets sent to Nimbus from section to section, depending on what I'm writing, but it is never more than a few percentage points. FWIW, I believe Nimbus' smaller brother, PhoenixVerb is on sale for $10 over at the Deals forum -- which is an absurdly low price for a top flight, neutral and clear reverb, and I'd totally recommend you grab it and try it out. I have a ton of reverbs (I have something of a weakness for them), but the only one I fire up these days is Nimbus. I haven't touched the others in years. In my opinion, it is THE software reverb for Orchestral stuff. 

I also put a lot of work into building my orchestral template, and I have deliberately selected instrument patches that inhabit similar sounding acoustic space (or can be made to do so, using various mic mixes). I've also spent hours fine-tuning the template so that each instrument in the entire symphonic assembly sits where I want it to across the acoustic landscape. 

I use a really broad mix of libraries:

1) Orchestral Strings: AlbionONE (sustains, tremolo, short articulations and pizz); 8Dio Anthology Strings' legato patches for dramatic lines; Sonokinetic's Modal Runs; Hollywood Strings Gold runs patches; Sonokinetic Da Capo's legato string patch (for really quiet passages, as I like the sound of their strings when the mod wheel is dialed really low); Palette's Trill patches; NI Symphony Series Ensemble Strings Harmonic patch

2) Woodwinds: VSL solo woodwinds, all from the Synchronized Special Editions; Sonokinetic Ensemble Woodwinds Standard Edition

3) Brass: VSL solo and ensemble brass, all from the Synchronized Special Editions; NI Symphony Series (for low brass)

4) Percussion: NI Symphony Series Percussion (for snares, bass drum, some timpani, tamtams, suspended cymbals, mark tree); VSL Timpanis; Kontakt Library VSL Glockenspiel; Hollywood Percussion Gold (for orchestral chimes); EastWest Storm Drum 2; AlbionONE's Darwin Percussion Easter Island Hits; August Forester Grand Piano from Kontakt Library

5) A various scattering of other libraries: Kontakt Library's VSL Harp, Lacrimosa Choir patches from 8Dio, Mercury Boys Choir Elements, Embertone Recorders, and a handful of Eduardo Taloronte's Era Libraries (Era II Vocal Codex, Dark Era percussion and flutes, some Forest Kingdom flutes, etc). 

If I need to push an instrument back or bring it forward, I usually start with the microphone mix settings built into the library in question, or the reverb built in the sample engine/library. This last point applied mainly to the original VSL Special Edition libraries, which were quite dry and center panned, and needed Vienna Ensemble's built in reverb to push them back into the acoustic space, and Vienna Ensemble's mixer and stereo width slider to position them properly across the acoustic landscape. When VSL released the Synchronized versions of its Special Editions, the reverb and positioning of the default mix for those instruments worked a lot better right out of the box, and they fit into my template's acoustic space with minimal adjustment. 

The only time I ever fire up a second reverb is for some of the Era patches, if I need to give them a particularly distant or washed-out sound (and for that I usually create a second reverb send with Valhalla's Shimmer). There were a few occasions where I tried using Wave's TrueVerb with it's early reflection room/space simulations, and would run various sections through it first in attempt to manipulate their positioning, but I never found it more effective than just whatever settings I could manipulate in the various libraries themselves. 

As far as getting good mixes is concerned, I bought the best monitors I could afford (a pair of Adam F7), spend a LOT of time with them so I have a good sense of their sound and how it translates, and when in doubt I tend to trust the ears of the engineers who made the samples. I rarely do much beyond the slight cutting of frequency here and there -- usually in attempt to clear out some mud and honk from the mix. I almost never add anything via EQ, and I usually do minimal compression unless I have a specific solo instrument that really needs to cut through the mix. And before I compress it, I usually try to clear space for the solo instrument by adjusting the orchestration first. 

When I have a final mix from Cubase (my main DAW for midi work), I usually take the exported track and bring it into Cakewalk and use ProChannel to master it (which for me usually means giving it a bit of boost using ProChannel's Concrete Limiter, and adding ProChannel's Tape Emulation and/or Console Emulation to give it some analog mojo). 

I have no idea if the above things are right or wrong, or outright heresy, but right now I get results I am generally happy with. I've linked to two tracks below, both of which were built using the exact template setup listed above. For both, there is only one instance of Nimbus reverb, very limited EQ, and Cakewalk's ProChannel Concrete Limiter and Tape Emulator over the final mix - basically just four plugins across two DAWs, and absolutely nothing else. To EQ these tracks, I used IK Multimedia's White Channel on the "Cut Boominess" setting, and dialed back the amount it was actually cutting by a fair margin. I find it effectively takes just a bit of thickness and smear out of the low end, and to my ears that seems to go a long way to clearing up the whole mix without gutting the bass.

I find these particular mixes sound great on my monitors, very good on my iPhone earbuds and my bluetooth headphones, acceptable in my car (nothing sounds particularly good in my car, frankly, so its a good test), and the folks I've sent it to have never made a negative comment about the mix quality, so I assume it sounds decent enough on their systems.  

As I mentioned, I do not presume to have all (or any) answers to the various issues that crop up for us folks doing symphonic music on computer. But what I outlined above has worked for me, and I figured I'd offer it up as some potential ideas that may work for you as well.

https://soundcloud.com/amicusaudio/the-soldiers-hornpipe-ver-4

https://soundcloud.com/amicusaudio/a-song-for-the-trillium-queen-adrielles-theme

I hope some of this helps, or at least inspires some different approaches. But for what it's worth, I totally get how you feel, and I've ranted a bit myself on this topic. 

Good luck! And if you want to connect about this stuff -- send me tracks to listen to, bounce ideas back and forth, etc -- please let me know. We orchestral folks need to stick together

Regards,

Rob

Wow, thank you very much, Rob, that was really informative. I would love to have my mixes sound as good as yours. sometimes I wonder if the style I compose in has something to do with it. I don't have much dynamic variation in my music. I tend to write stuff that's on the quieter side. Or maybe my samples aren't good enough. 

Anyway, I will keep re-reading your advice, there's a lot there. I've already read it three times, and again I really appreciate your taking the time to help me out here.

[JoeGBradford]

On 4/15/2020 at 6:25 PM, mdiemer said:

OK, you asked for it:

https://app.box.com/s/s2mb484mubt2zxg9o1gq2zke17b9uu6f

It's the first movement of Three Easter Scenes, which I just completed.

Nice piece! Know what you mean about the sound though - yours sounds much better than the little I have composed but, as you say, it still has that unrealistic sound even though you have some great libraries