I exported the song I’ve been working on from Cubase 8 Pro into an MP3 file. I did it by first doing an Export Audio Mixdown to a WAV file, and having that audio imported back into my project as an audio track. Then I normalized the audio track (which is why I had the audio brought back into Cubase).

Then I did an Export Audio Mixdown from the audio track into an MP3 file.

I ran into some puzzles:

1. I wanted there to be a short period of silence (~2 beats) before the MP3 started to play. I noticed there is a Start parameter in the Project Setup dialog, but I didn’t really see how to use it for this purpose. So I ended up inserting the delay by selecting the Event containing the audio data from the creation of the WAV file, and moving its start point later by two beats. I’m wondering if this is the best/usual/easiest way of doing something like this. I suspect something more convenient is possible.

2. I normalized the audio track from the creation of the WAV file, but the MP3 file needed to be normalized despite that. I was able to do this using Audacity, but I was surprised that I needed to do it given that I had previously normalized the source audio track inside of Cubase.

I think that I don’t quite understand what Cubase does when it normalizes audio.

Hi Michael,

It’s been a long time since I’ve used Cubase, but doesn’t Cubase export exactly whatever exists between the left and right locators, to that precise length?

I’m pretty sure it would go something like this:

1 - Leave your left locator as is, at the start of your song presumably.

2 - Move the region(s) of your song exactly 2 beats to the right (to create your gap)

3 - Move the right locator along at least 2 beats to compensate so that everything still fits between the locators.

Then do your export.

It may also be possible to set your left locator to a negative value these days (e.g, minus half a measure), in which case, that’s all you’d need to do and my steps wouldn’t apply. Although, I’m confident that they will get you where you want to go.

Why are you normalizing what appears to be a mix?  There are much more musical ways to get things louder, if that is your aim.

It does indeed export what’s between the locators. I personally have all my templates saved with the left locator on bar 10 since it’s highly likely that I want some kind of intro later on.
Could it been so that your Cubase normalise had a too low level set? I.e. not -1 or 0 db(or whichever level you aimed for)?
I think Cubase only do peak normalizing, maybe Audicity uses RMS?
Not sure if that could explain the difference.

The problem with normalization is that it is an automated process and is no different from having a big volume knob.

It drags everything up and it doesn’t discriminate.

The problem with using it is that any present system noise also gets dragged up.

So, in a good mix, the engineer will squeeze every drop of volume out by recording tracks well with good levels first, then pushing those tracks up individually in the mix and keeping to an overall ‘as loud as possible without going over’ balance in order to put as much distance as possible between the instruments in the mix and any noise which may be present. 

Without getting into peak limiting etc, if your mix is quiet, instead of a robotic hammer solution like normalization, how about moving your entire mix balance up incrementally instead? Watch the metering and use your ears.

If you can’t do that or it doesn’t help, then it is likely that your tracks have been recorded too quietly in the first place.  (or you had deep fat friers present near the mic)

philwoodmusic - 18 November 2015 10:54 AM

The problem with normalization is that it is an automated process and is no different from having a big volume knob.

It drags everything up and it doesn’t discriminate.

[...]

Good points. Another is that normalization of a WAV and normalization of a lossy, compressed format (MP3) can have differing results. MP3 files typically are at 44.1 kHz, the same as a WAV (although the standard allows for 32 and 48 kHz). However, MP3 bit rate can vary significantly, and a low bit rate can sound bad no matter what normalization does to “correct” levels.

If the reason for creating an MP3 is to shrink the file size, a reduction can still be had, but with less audio quality loss, if the bit rate can be set to 128 kbps (or higher). I suspect that the Cubase default rate might be substantially lower than that.

Thanks, guys.

I’ll try those tricks with the locators.

As to the normalization, I could easily be going about what I’m trying to do the wrong way.

My problem is that due to hand problems, I can’t hit the keyboard very hard. As a result, the volume of what I play is low because the Velocity of my notes is low. When I create audio from my MIDI tracks, the vertical size of the audio envelope is only about 1/4 of the available height.

I’ve been using normalization to make my audio loud enough.

I’ve thought of some other things, but haven’t gotten around to trying them yet:

* Change the velocity response curve of the Voices that I use.

* Use Cubase to “expand” the Velocity of my notes proportionally.

* See if there’s some way to use the Mix Console for this.

I’m open to ideas from people with expertise. I’m just starting to learn to use Cubase.

Michael Trigoboff - 19 November 2015 01:49 AM

[...] My problem is that due to hand problems, I can’t hit the keyboard very hard. As a result, the volume of what I play is low because the Velocity of my notes is low. When I create audio from my MIDI tracks, the vertical size of the audio envelope is only about 1/4 of the available height.[...]

I presume you’ve already set the XF’s Velocity Curve to “soft” in Utility mode, and stored it.

I don’t use Cubase. However, many DAWs have a means of modifying incoming MIDI data. Apparently Cubase has an “Input Transformer”.

See…
https://www.steinberg.net/forums/viewtopic.php?t=23433
http://www.soundonsound.com/sos/jun11/articles/cubase-0611.htm

It would seem that multiplying the incoming note velocities would help with your situation. The value of the factor could be roughly determined by noting the highest velocity you can comfortably play, and then dividing 127 by that.

For example, let’s say you can hit velocity 45…
127/45 = 2.82, so try multiplying incoming velocity by that.

Once recorded, as a final tweak, you could draw in your desired velocity using the pencil tool in the velocity controller lane in the key edit window.

If I recall correctly, the information is presented to you as a stick graph, with a stick under each note.

Yeah, the stick graph is still there and can be quite useful to mediocre players like myself to correct the odd notes.
Even though I haven’t used the functions in the MIDI menu a lot myself, I think that there are some functions there that may suit a better player that just can’t hit ‘hard enough’.
Worth checking out if it’s only the velocity that needs fixing.

5pinDIN - 19 November 2015 03:08 AM

I presume you’ve already set the XF’s Velocity Curve to “soft” in Utility mode, and stored it.

Yikes! I don’t know how I missed that! (Smacks head with heel of hand.)

Thanks!

That transformer trick also looks good.

philwoodmusic - 19 November 2015 06:36 AM

Once recorded, as a final tweak, you could draw in your desired velocity using the pencil tool in the velocity controller lane in the key edit window.

If I recall correctly, the information is presented to you as a stick graph, with a stick under each note.

Excellent suggestion. Thanks. Will try it tomorrow.

I adjusted the velocities of the notes in my song tracks, as suggested. The sound envelope of the audio I exported has now become a bit wider vertically (i.e. louder), but just by a little, and not by enough.

I did some reading up and found suggestions about tweaking individual voice levels in the XF, or using a “compressor” in the Master Effects of the Xf7. None of this has really helped. I looked in the Cubase Mix Console to see if there was anything there that might be effective, but I didn’t find anything.

Is there some global adjustment that affects the output audio level? It seems to me like there should be, but I haven’t been able to find it. Without something like that, I don’t know what else I could do other than use Normalize.

I believe that the Input Transformer should work as I previously described.

If you can’t find it, perhaps this will help:
https://www.steinberg.net/forums/viewtopic.php?f=226&t=76491

If not, and if no Motifator member knows more, then I’d suggest posting on the Steinberg site.

By the way…
Changing the velocity data after it’s been recorded isn’t the best way to do things. You want to hear what you’re recording sounds like, as you record it. The character of a Voice, not just the volume level, changes with velocity. You should have Local Control turned off on the XF, send MIDI data to Cubase, and have that data returned to the XF. That way, if velocity is increased before recording MIDI data, what you’ll hear will be a more accurate representation of the final product.

Otherwise, significantly changing velocity values afterward can result in unexpected changes in the overall sound. In fact, since some Voices have velocity switching, the result can be startlingly different.

Michael,

What are you measuring the level of your files against, in terms of loudness?  Just out of interest.

And how are you measuring it?

I understand that you have to play softly, but it’s totally possible to make even softly played music fairly loud, without velocity changes.

I see your problem as two problems. 5pin’s suggestion should help your hands, but you are for whatever reasons not getting good levels into your DAW at any velocity. 

There’s really only a few places where you can change that, but you don’t mention which mode you are working in on your Motif. (voice, perf, song, pattern etc)

This is a non serious comment, but you’re not comparing the level of your files to something like Metallica’s Death Magnetic album are you?

In all seriousness though, we can’t compete with the level of commercially available music which has gone through a top mix engineer and top mastering house. 

There are many things we can do that help, but it all depends on the quality and good recording level of the music.

Also, a compressor turns peaks in audio down, not up.  They can also colour the sound somewhat.

philwoodmusic - 22 November 2015 01:57 AM

[...] I see your problem as two problems. 5pin’s suggestion should help your hands, but you are for whatever reasons not getting good levels into your DAW at any velocity. [...]

Perhaps this should be addressed first…
What maximum velocity levels can Michael typically attain? How close to 127 - are we talking 50, 80, 120? The answer can help determine what approach to take.

If somewhere near 50, there’s no question that the input velocity needs to be increased.

If around 120, increasing velocity probably isn’t going to help much, and other factors can/should be dealt with.

5pinDIN - 21 November 2015 11:59 PM

I believe that the Input Transformer should work as I previously described.

If you can’t find it, perhaps this will help:
https://www.steinberg.net/forums/viewtopic.php?f=226&t=76491

If not, and if no Motifator member knows more, then I’d suggest posting on the Steinberg site.

By the way…
Changing the velocity data after it’s been recorded isn’t the best way to do things. You want to hear what you’re recording sounds like, as you record it. The character of a Voice, not just the volume level, changes with velocity. You should have Local Control turned off on the XF, send MIDI data to Cubase, and have that data returned to the XF. That way, if velocity is increased before recording MIDI data, what you’ll hear will be a more accurate representation of the final product.

Otherwise, significantly changing velocity values afterward can result in unexpected changes in the overall sound. In fact, since some Voices have velocity switching, the result can be startlingly different.

Excellent suggestion! I will definitely route things that way from here on out.

I’m delighted to be at a point where I can actually understand what you said and do something like this.