Low level on exported Mp3 file from Sonar

[Robert Bone]

There is SOOOO much headroom available in the digital world of Cakewalk, that you never need to record with a 'hot' level that approaches unity.  You would almost certainly be clipping digitally if you did.

Most VU Meters are calibrated, by default, using a sine wave 1k signal, so that a reading of 0 dBVU = -18 dBFS.  VU meters work in measurements that are an AVERAGE of the level within a 300 millisecond time span, whereas Peak meters work in measurements that are ABSOLUTE - not averaged, but the actual Peak.  Two different kinds of levels.  When a VU meter is calibrated to the standard of -18 dBFS, then when the VU meter is at 0 dBVU, the full-scale equivalent (dBFS) is really down at -18 dBFS.

That means that when the VU meters are hitting around 0 dBVU, you really have  a  full 18 dB of headroom - which is  bunch.  and, because the noise floor in the digital world is so low, you can raise the signal way up (like to do the export with 10-12 dB of added gain), without adding any  audible noise).

SOOOO - keep on recording at your -15 dBFS levels, you are FINE - and just remember to temporarily bump up your Master Bus Gain just prior to export, so that your exported level is set to about -3 dBFS (that is the PEAK value, not the average/RMS/dBVU value), and turn the Master Bus Gain back down so you are back to playback at the -15 dBFS level you were at prior to the export,  and all will be fine.

Bob Bone

 

[Robert Bone]

Here are a few videos that will explain the calibration of VU meters, and gain staging in general, and dbFS and how the human ears hear, and all of that sort of thing - and that may help you to understand what levels to record at, and why, and when it makes sense to use VU meters instead of Peak meters, as well as when to use Peak meters instead of VU meters, etc....

https://youtu.be/DuWhMl8oMwY  Video on Gain Staging with VU

https://youtu.be/gyJPRQOqk1g  gain staging at -18 dBFS or 0 dB VU

https://youtu.be/sCj5wrPvLs8  Video on Digital Gain Staging

 

Bob Bone

 

 

 

[James See]

With all due respect, I think the respondents to this post are missing the point.  I have the same problem as Dusan.  In my final mix, the master level averages at -9 and peaks @ -3.  You can't go much hotter than that.  When I export the project to mp3 and play that file on my Windows media player the audio level is is just fine.  As soon as I play that same mp3 file on a different system the audio level is so low that the system has to be turned up almost to  max volume to hear the song at a normal listening level.  So while Dusan may have some recording or mixing problem as well, I do not believe that is the root of the problem he and I and what I am beginning to understand, a lot of other people are having.

[Kalle Rantaaho]

On 6/23/2020 at 3:05 PM, James See said:

With all due respect, I think the respondents to this post are missing the point.  I have the same problem as Dusan.  In my final mix, the master level averages at -9 and peaks @ -3.  You can't go much hotter than that.  When I export the project to mp3 and play that file on my Windows media player the audio level is is just fine.  As soon as I play that same mp3 file on a different system the audio level is so low that the system has to be turned up almost to  max volume to hear the song at a normal listening level.  So while Dusan may have some recording or mixing problem as well, I do not believe that is the root of the problem he and I and what I am beginning to understand, a lot of other people are having.

If you open the exported file in CbB, how do the level readings look?

The different sides of getting a loud song are well explained in this thread.

I believe most commercial products  of  rock music and similar today are limited/compressed to louder than -1 dB, usually at least around -0.5- -0.3
How do you compress and limit your project ? The difference between -3 dB peak and -3 dB average is very audible, not to mention between  -3 and -0.3, which is huge.

If the audio level in WMP sounds fine, but much too quiet in other systems, it only means that the WMP has it's volume set loud. There is no other explanation. If the average level of the audio file is, say, -3 dB, then it is -3 dB no matter where you play it.  

As already mentioned in this thread, at the end you don't build the loudness with tracks level sliders, but with compression/limiting at master bus, after per track volume automation and some per track compression.  That way also a low level recording can be made loud. 

Edited June 24, 2020 by Kalle Rantaaho

[John Vere]

There' just a lot of things typical of novice recording engineering done incorrectly that all add up to quiet recordings. 

First the low level recording. I too have a Scarlett 6i6 which is similar in design. Yes there are no hardware input levels as those inputs are designed to be teamed up with hardware that has a master volume, like a keyboard or a mixing board. Put the tracks in record standby , play as loud as you will be playing and adjust your keyboard master volume to keep below clipping at ALL times.  There is also the Mix control software but that takes a very steep learning curve to master. 

Once you've got a better level on the track you then need to push that at the master without clipping. 

All it takes is one little peak to clip the output. This is were a good mastering limiter is required. I use the BT Brickwall set at    - 0.9 and this allows me to hit the master a little harder. 

But then I always master my songs in Wave Lab because it's critical to producing a proper finished release. 

The recording process : In the real world this is done by 3 different people who are very good at what they do with years of experience and training. Please don't expect a computer to bypass any of this process and get the same results they do. 

Tracking

Mixing 

Mastering 

[chuckebaby]

Dont trust windows media player for any kind of measuring stick. Mix to -3db peak. Then master it. Which should bring the level up to -0.3db (for digital overs never mix to 0db). Mastering is the process that gives you the loudness while still retaining dynamics.

[John Vere]

Not sure because I've never used the export to mp3 but anyone see a difference in encoding algorithms?

I always export as a wave so I can master the song in my wave editor.  Once the wave file is 100% then I'll convert to MP3 using Gold Wave. The MP3 are amazingly identical to my ears. I use 192 kbps. 

Here some examples https://sites.google.com/view/cactus-studios/original-songs

 

Edited June 25, 2020 by John Vere

[James See]

Just for some clarity, using a VU meter referenced to -18db the Cakewalk mix at the master channel is averaging between -3db and -2db.  I created a wave file and the average level was about -2db. It was actually a little hotter than the mix.  The mp3 version came in between -6db and -5db.  This is a considerable drop in level.  In theory, all three versions should have the exact same level.  Given that the level at both the master and the wave file version are plenty hot, that would lead me to say that any considerations about mixing or mastering would be moot.  What does seem to be happening is that Cakewalk is producing the mp3 file with a lower level which it certainly should not be doing to my knowledge.  I have heard of many people having this problem but clearly it's not everybody.

[Bob Greaves]

I have had this same problem with SONAR in different versions. I had this isue in X1 and X3 but not X2. Upon export to either an mp3 or wav, the resulting exported file is about 12 dB lower. If I then import that exported file  back into the SONAR project, that track plays back in the project about 12 dB lower than the project itself if I go back and forth.  If I raise the level of the master out on export, then the exported file will be clipped and yet not at full potential volume. It clips with 10 or 12 dB yet to fill its potential. Somehow the translation from 24 bit in the project to 16 bit in the output is over reducing the the amplitude of the result. 

In the versions that did this I have never found a solution. Reinstalling - NOPE! Updating drivers - NOPE. Somewhere there is a setting I am missing.

[John Vere]

Peak level is only part of the issue. Unless it's a sustained note it doesn't tell you how "loud" something really is.  You need to know the RMS average or LUFS. What I'm using now with very good results is the Youlean meter. The paid version allows you to select an output format and reports when you go over that limit. It also allows drag and drop analyzing.  The basic version is free but I highly recommend paying as it's a very good investment if you are mastering your own songs. 

https://youlean.co/youlean-loudness-meter/  

Span is also free and is useful. 

https://www.voxengo.com/product/span/

And for Mp3 conversion I use Gold Wave which is free to use but will ask for payment and is well worth the $50. 

http://www.goldwave.ca/

[James Dziezynski]

Hey Dusan! 

The answers here, while right, don't really help with your problem. I'm no expert, but I think I have a solution that could help you. 

1) Export your track to a stereo wave after finding mix levels you  like. Your volume levels can be nice n' low (-8 to -12 db is what I do). Then close your CWB file. 

2) Open a new EMPTY project, add an AUDIO track and FILE > IMPORT your stereo wave file. 

3) On the left you'll have your Pro Channel modules. I like to delete them all, except EQ (which you can't delete). Set your EQ to what pleases your ear by double clicking the "screen" on the EQ module (again, this is just an exercise in improving volume). I usually set a sharp slope (48) and freq to 24, but again, that's just me. 

4) For ease of use, next add a Compression Module in your Pro Channel after (below) the EQ. (Add it with the plus icon at the top of your channel strip, still on the left of the screen)

I like the PC4k Style Bus. To start, you I like to tweak threshold and attack. Turn thresh all the way up (no compression) and gradually lower it about -2 more to cut off possible high clips. You can use the attack as well to tweak the tone. I set the release to 0. 

5) Finally - the coup de grace - ADD (plus symbol) a final module > Style Dial Knobs > Max. This is a limiter and should come at the bottom, end of your chain in your channel. It's good for dummies like me because it's just one big knob. Find the loudest part of your wave and turn the knob until you are at 0 or even +3 DB on your channel. Make it loud, as long as it's not clipping. This is where your quieter recording pays off - if you record too loud, a limiter can make a throbbing, pulsing sound. Because you kept your levels open to headspace, that shouldn't be an issue!

This will be the volume you're looking for. Export to wav or MP3 and you'll have some color to your song, but also a "real recording" volume. 

There's a zillion more nuances in these plugins but they are fun to figure out. The Style Dials seem to be made for amateurs like myself and so I put them in from time to time as the audio changes are very apparent. Many other plugins are much more subtle. 

Hope this helps!

PS: Here are two examples I recording this morning to prove the point (not to show off anything in my writing or playing)

Before mastering: https://www.mountainouswords.com/wp-content/uploads/2022/03/Facing-Up-1.mp3
After: https://www.mountainouswords.com/wp-content/uploads/2022/03/final-boost-master.mp3

Edited March 23, 2022 by James Dziezynski
minor spacing error