Hi Terry

See this description of normalisation - MEP uses peak detection (maximum level) and in VPX there is also the option to use EBU R128 normalisation

In very short clips there are insufficient peaks or no significant peaks for a true assessment of what the actual level compared to a longer clip.

HTH

John EB

Hi John EB,

What is the solution?

I have taken to letting the recording continue on for about 10 seconds and then trimming. It seems to help.

Or should we not normalize until the end, and do a mixdown of the audio track. I just tried this and it seems to be the solution.

1. Record with Normalize after recording set to off;
2. Make sure that all recordings are on one or two tracks, with no other audio on them;
3. Mute all other tracks that have audio that is not to be normalized or mixed down, or simpler, solo the track or tracks to be normalized;
4. Combine audio (under Edit) or Shift+D. Note the name and location of the file to be the mixdown;
5. Note the filename of the resulting mixed down wave object. It should be red and the original audio clips will now be gone;
6. Undo to remove the mixed down object and get the original audio clips back;
7. Navigate to the location of the mixed down audio file that you noted and drag it onto the timeline.
8. Select it, right-click, Normalize (maximum level). In VPX, select Normalize (EBU R128), which will give a lower output.
9. If using the Maximum level, reduce the audio level by about 6 db. Easiest is to use the slider in the Mixer.
10. Mute all corresponding source audio tracks, as per 3 above.
11. You now have a mixed down audio file, normalized, and the possibility to go back and delete/change recordings throughout. If this needs to be done, delete the mixed down file, unmute the other audio track, fix up the source, and start again at point 2.

It sounds difficult, but it takes longer to explain than to do.

Comments?

Thanks both.

(Edit: Also now seen your post, John CB. As you say, it seems complicated!)

It's odd, but there seems to be some other factor involoved in my case. I've just made a series of test recordings. About half were 7-10 secs and others 2-3 secs. On this occasion there was no significant variation; all of them were satisfactory. Similar recordings with Normalise disabled were also consistent, regardless of duration, but all were too quiet.

But yesterday, not for the first time, with Normalise enabled I had a series of 'too quiet' recordings, which was what prompted my post. I tried removing and replacing the USB plug at the mic several times, and also sliding the mic's base cut and -20dB swiches (both normally disabled) on and off a few times, but no change.

The mic is mounted, not hand-held, about 16 inches from my mouth. FWIW the volume setting is at 41.6 dB.

Terry, East Grinstead, UK

Hi Terry,

What type of a mic is it? What do the instructions say about distance from the mic? Do you have an anti-pop filter/windscreen on it?

Mine is an iTrack Studio CM25B condenser microphone with a cardioid polar pattern that connects using an XLR cable and connection to my M-Audio M-Track audio interface, which is connected to a USB port. It uses phantom power. I added a foam filter and have found that 6 inches is the optimal distance from my mic. The gain is turned up close to max on the M-Track and the input level in Windows is, oops, at 67. I thought that I had it at 100. I just boosted it. Now I will have to test again. Maybe that is why my levels without normalizing are so low.

Time to do some yard work, more later.

Hi

The method I use for MEP which uses peak normalisation is:

1. Import or record required audio
    
2. Clean up each audio object if required
    
3. Normalise each audio object if required
    
4. Adjust the relative volume levels to the levels I require for each object
    
5. Add any volume curves (ducking) necessary
    
6. Use MDynamic EQ in the Master effects to adjust the final audio level audio to -6dB.

I skip adjusting volume at track level because my audio can be on any free track and grouped with an image or video object, collage or PIP element on the timeline.

Combine / Mix down the audio - never use it.

@ Terry

. . . . On this occasion there was no significant variation . . .

A classic example of Murphy's Law.

. . . .sliding the mic's base cut and -20dB swiches (both normally disabled) . . .

Bass cut being on for speech is OK - it should help filter out any mains interference.

The -20dB switch - is this to cut the level by a further 20 dB or limit the mic max output to -20 dB ?

Have you added mic boost in the Windows Sound, Recording devices - I have to use 30 dB boost with my desktop mic.

HTH

John EB

HTH

John EB

Hi Terry,

What type of a mic is it?

Hi John,

AKG Perception 120 USB. Here's the User Guide:
https://www.dropbox.com/s/egh7tysh8lqk88s/AKG%20Perception%20120%20USB%20Guide.pdf?dl=0

 

What do the instructions say about distance from the mic?

"The microphone should be 20 to 40 cm (8 to 16 inch) from the speaker's mouth in order to obtain the best results."
I used to hand hold it, at about 10", but now have it in its stand behind my keyboard. (Actually one of its three legs is on the keyboard, to get it as close as possible.) So, at the extreme of that range, I'd like to get the highest undistorted volume.

Do you have an anti-pop filter/windscreen on it?

No, indoor use only.

Mine is an iTrack Studio CM25B condenser microphone with a cardioid polar pattern that connects using an XLR cable and connection to my M-Audio M-Track audio interface, which is connected to a USB port. It uses phantom power. I added a foam filter and have found that 6 inches is the optimal distance from my mic. The gain is turned up close to max on the M-Track and the input level in Windows is, oops, at 67. I thought that I had it at 100. I just boosted it. Now I will have to test again. Maybe that is why my levels without normalizing are so low.

Time to do some yard work, more later.

I expect your 'yard' is a lot cooler than our 'garden', which is around 27 C at this time ( 20:12).
 

Hi

The method I use for MEP which uses peak normalisation is:

1. Import or record required audio
    
2. Clean up each audio object if required
    
3. Normalise each audio object if required
    
4. Adjust the relative volume levels to the levels I require for each object
    
5. Add any volume curves (ducking) necessary
    
6. Use MDynamic EQ in the Master effects to adjust the final audio level audio to -6dB.

I skip adjusting volume at track level because my audio can be on any free track and grouped with an image or video object, collage or PIP element on the timeline.

Combine / Mix down the audio - never use it.

@ Terry

. . . . On this occasion there was no significant variation . . .

A classic example of Murphy's Law.

. . . .sliding the mic's base cut and -20dB swiches (both normally disabled) . . .

Bass cut being on for speech is OK - it should help filter out any mains interference.

The -20dB switch - is this to cut the level by a further 20 dB or limit the mic max output to -20 dB ?

Have you added mic boost in the Windows Sound, Recording devices - I have to use 30 dB boost with my desktop mic.

HTH

John EB

HTH

John EB

 

Hi

The method I use for MEP which uses peak normalisation is:

1. Import or record required audio
    
2. Clean up each audio object if required
    
3. Normalise each audio object if required
    
4. Adjust the relative volume levels to the levels I require for each object
    
5. Add any volume curves (ducking) necessary
    
6. Use MDynamic EQ in the Master effects to adjust the final audio level audio to -6dB.

I skip adjusting volume at track level because my audio can be on any free track and grouped with an image or video object, collage or PIP element on the timeline.

Combine / Mix down the audio - never use it.

@ Terry

. . . . On this occasion there was no significant variation . . .

A classic example of Murphy's Law.

. . . .sliding the mic's base cut and -20dB swiches (both normally disabled) . . .

Bass cut being on for speech is OK - it should help filter out any mains interference.

In its new position, mounted on its little tripod, happily I'm having no serious problems with mains noise now.

The -20dB switch - is this to cut the level by a further 20 dB or limit the mic max output to -20 dB ?

The former. See also user guide if you're curious:
https://www.dropbox.com/s/egh7tysh8lqk88s/AKG%20Perception%20120%20USB%20Guide.pdf?dl=0

Have you added mic boost in the Windows Sound, Recording devices - I have to use 30 dB boost with my desktop mic.

Pretty sure we concluded I don't have that option here? Just looked again but...

HTH

John EB

HTH

John EB

 

Terry, East Grinstead, UK

Hi Terry,

It's 27C and humid in the garden this am, going to over 30C.

The pop filter is to help remove or reduce plosive p's and s's - Peter Piper picked.... and She sold seashells, and heavy breathes. Mine helps with this.

I'm off to wash the trailer and do more yard work.

As a recent convert to VPX, I've just tried the discussed EBU R128 normalisation option on a single clip lasting 1hr 24m +, and am still told the object is too short! At the very least the error message to the user is unhelpful....

The Help PDF is not much help as to how and when to use the facility! - Mike -

Hi Mike.

What happens if you click "Continue"? I've just tested this on a video, numerous clips, total playing time about 58 minutes. I received the same error message for some of the clips (which were admittedly very short). Clicking Continue seemed to apply the effect to at least most of the clips (I haven't had time to check all of them).

Jeff

Hi

@Recycler

I get the same message, after a 'little' testing, it appears that a video or audio clip longer than 45m 27s 10f will 'fail' and give this message, however clicking Continue does apply the normalisation which 'sounds' correct.

IIRC the shortest audio which can be normalised is about 8 secs.

Looks like it is a bug or the wrong message.

HTH

John EB