What's new

Question about mixing in audio vs. midi

Ok thanks. Yes I know it’s very low volume, it’s straight from a very quiet section in the piece. I should’ve made it louder and can correct that tomorrow if it simply can’t be heard.

No, no Event Input plugin. Will check the multithreading setting and possibly send the project, or part of it, tomorrow as it’s very late here.
 
Yes that makes sense, although I'd have to do some testing to determine how audible the changes in room as opposed to instrument levels actually are when adjusted. Worth exploring though for sure.

Been working all day so just about to wrestle some more with this strange issue regarding having to select legato tracks to make them play properly. Not sure there's much point in sending the project file because as I said the same thing is happening in a brand new project with nothing in it except the offending track. Unless I send that track but first I want to see what happens if I create a new track rather than import it. At this point I'm thinking the problem is most likely to do with the way the Mac and the PC, or Logic and VEPro, are talking to each other, especially when Play is involved, rather than anything going on in Logic itself. However I guess anything is possible.
 
Its somewhat academic though. We already use all manner of artificial effects while sculpting our sound. You can do everything "wrong" and you'll know if it sounds bad. if it sounds good, then there is nothing wrong with the approach, including freezing tracks with burned in room sound.

If you really think about it, what is the difference between sending a dry signal with a higher or lower signal, through an ER generator....or adjusting the level after it has been sent through the ER generator? It only makes a difference if the ER generator is doing something different to the sound depending on the level. Will the early reflections be different for a low volume sound compared to a high volume sound src? Possibly theoretically, or maybe not, but I don't you will notice much difference in actual practice of mixing it all together. There are other things happening in mixing though, EQ, etc. So this will depend on what you're trying to do whether you need the close dry sound to work with in addition to the ER'd sound. If you only have burned in samples, then your options or more limited. But I think in many cases you can adjust the ER'd room sound on an instrument by instrument basis...and while it may not be strictly the way a real orchestra would be recored with real engineers, if it sounds good...fine by me.
 
Not sure there's much point in sending the project file because as I said the same thing is happening in a brand new project with nothing in it except the offending track. Unless I send that track but first I want to see what happens if I create a new track rather than import it.

If you have an extremely simple project with just one track having the problem that is even better. Please PM me the project and I'll have a look at it.
 
Labor in Vain,

Yes really you are being academic. If that works for you then keep doing it! But its not the only way to get good sound. Who cares about whether its "realistic"? Half of what you hear from hollywood is not!

If that is your aim, to realistically and accurately reproduce a particular location, then I agree with you. I use MirPro and that is one reason why I like it, but even in MirPro there is no rule that says you have to conform to anything in the name of realistic accuracy of the physical universe. The goal is to make good sounding music and you can definitely get great sounding, orchestral results..without getting so far into the weeds that you feel its not possible to bounce your instrument tracks prior to mix down.

EW HO is a classic example of how its possible to get great sound using sounds with ER baked into the samples. People are doing it every day.

Think about it some more. What is the difference between bouncing down the output from EWHO versus playing it from the instrument?

Answer = None.
 
Interesting discussion there guys though I don 't agree that labornvain's point is merely academic. However I'm not sure that I'd make an issue of it either in terms of avoiding mixing with audio tracks.

Anyway..

So I've created a new track with the ostinato figure written fresh into it and I'm getting the same anomaly. I've started a new Server project with just the one instance with Play on it hosting the one oboe legato patch. Same problem. Changed the audio output to built-in rather than my interface. Same problem. Setting multi-threading to Playback Tracks makes no difference.

Yes Dewdman I'll send you the simple project if you wouldn't mind trying to replicate the issue I'm having, or seeing what you can find. Bear in mind this is a 'vanilla' project, i.e. it is not based on a template or anything else so there shouldn't be anything going on in the background to complicate matters. On the VEPro side I have an instance of Play with the OB Leg Slur VS Full NI patch loaded. Try the playback with the midi track selected and then with the audio track selected and see if there's any difference. I'd be very interested to hear what you find.

Thanks for taking the time to do this. Appreciated.
 
I can't figure out how to send it. There is an option to upload a file but it won't let me select a Logic project, only an audio file. Do you know how to do this?
 
also if you can send an audio of what its SUPPOSED to sound like..if you are able to even create that... I will take a look either way
 
The audio that sounds as it’s supposed to is the first example I posted earlier, if you can hear it. I’ll try and make it downloadable so that you can raise the level.
 
I wonder if anyone else who runs Logic and VEPro over a slave system and has EW Hollywood Orchestra, or even just the woodwinds, would be willing to have a listen to this test project and let me know what you hear. Dewdman42 doesn't seem to be hearing the issue I'm talking about so I'd love to get another opinion or two. If so, please follow the guidelines at the end of the last paragraph of post #48 above, it's quite straight-forward and will only take a couple of minutes. The sort of difference I'm hearing is exemplified in the two audio files in post #39, although there are some CC's affecting those which have been erased from the test example to keep things as simple as possible. Many thanks in advance.
 

Attachments

  • Oboe test.zip
    9.1 MB · Views: 2
Dewdman42 couldn't hear your audio files before, its not that I couldn't hear what you are talking about. The files were inaudible for me. I have not had a chance yet to look at your updated sound files. But I will today...
 
Dewdman42 couldn't hear your audio files before, its not that I couldn't hear what you are talking about. The files were inaudible for me. I have not had a chance yet to look at your updated sound files. But I will today...

Well I didn’t mean anything by that, I was just saying that you weren’t hearing the issue, one way or another. Whether that’s the audio files or the midi playback is not really important. I’m simply asking that someone else take a look as well, no negative intended to you. I do appreciate you taking a look at this at all.
 
Alright, sorry for the delay. The updated soundcloud files are much easier to hear now. I imported them both into LPX and I can hear that the timing is off between the two versions.

But just looking at the bounced AIF's you provided, presuming they were done sample accurately with the same start times in each AIF, The one produced while your audio track was selected is delayed by 14,615 samples compared to the first one. So first off without looking any deeper yet...is that because you bounced them that way into the AIF or is one bounce procedure introducing latency that the other is not?

That being said I time-aligned their start times and I can also hear subtle differences and observe the differences visually in the bounced wave forms, though its difficult to hear a consistent pattern, but clearly they are different.

compare2.jpg

So the timing difference is one thing..not sure if that was your doing or something about the bounce being different when one track is selected vs the other. But the fact that the wave forms are different AT ALL other then that is a concern!

The question is why.

Honestly I'm having a hard time hearing a specific midi-performance-related pattern difference that i can identify and try to figure out, but they are different for sure.

At this point I am pointing the finger at LPX as to why its happening, only because I know LPX does weird things when you have track headers selected.

By the way, how did you bounce the midi track while the audio tack was selected? Were you doing offline bounce or real time recording through a bus or something of that nature?

Unfortunately I am in the middle of installing a bunch of large VSL libraries and my elicensor won't let me start up VEP until that's all done, and possibly I have to reboot. So I will try to replicate your procedure after that. But anyway, what was the procedure you used to produce these two bounces?
 
Last edited:
another test for you to try, what happens if you put PLAY directly in LPX and do the same two bounce operations? (eliminate VEP for this test)
 
Ok thank you. The start times and some of the visual differences are probably moot because I simply edited both files separately to delete the silent parts at the start and also to raise the levels. So they are not consistent with each other in those respects.

At this point the issue is not so much the variation between the midi and the audio files as the fact that the playback sounds different, meaning the timing is out and certain notes are not ‘triggering’ properly, when a different track is selected vs. the midi track that is playing. I hope that’s clear. Those recordings do still represent the differences I am hearing though so some of what we can see in the waveforms probably reflects that.

It seems that the ‘weird things Logic does when a track header is selected’ might be a factor here. I’d be interested in hearing more about that..

I did them using a real-time recording through a bus but the result is the same if I bounce the project in real-time with that track soloed. Again, now I do seem to be getting an accurate recording of the midi track, it’s more the midi track plays back differently depending on what is selected, whether I record it’s output or not. Again once again, I do hope I’m explaining the issue clearly enough here!
 
another test for you to try, what happens if you put PLAY directly in LPX and do the same two bounce operations? (eliminate VEP for this test)

Yes, I thought of that yesterday but haven’t done it yet because it means moving files and licenses etc. around but I should definitely try it. It’ll be tomorrow though because it’s late at night again here!
 
Top Bottom