Strange bug with audio modulation overlay on high frequencies

Milosz-Nasadowski wrote on 5/8/2022, 2:26 PM

I am doing video about bird sounds and I know I have heard it before but this time I noticed it's Vegas' fault.

Basically for I think mainly high frequencies if you place the audio track in Vegas it sounds with modulation overlay. A bit like some of the old basic kid's keyboard synthesisers.

There is no audio effects on the track and this is not acceptable sound for me, any ideas?

Attached example with 5 sec Original audio and the same 5sec clip after placing in Vegas.


rraud wrote on 5/8/2022, 4:56 PM

Hi @Milosz-Nasadowski, Is your Vegas 'Audio Properties' sample rate set the same as the file's and is the bit depth at least 16?
Does the rendered file have the same 'abnormal' sound quality as Vegas preview? (played back in another device)
What format is the original bird sound file?
Can you post the original sound file in DropBox, GoogleDrive. OneDrive or elsewhere?

DMT3 wrote on 5/8/2022, 5:24 PM

Make sure the effx that default to each audio track are not enabled.

Milosz-Nasadowski wrote on 5/9/2022, 3:00 AM

YES! Found it. It's not the bit depth but the sample rate set to 48k by default in Vegas. Apparently for those high frequencies it matters. Strange I never have the problem when I convert in audio software! Thank you very much!

todd-b wrote on 5/9/2022, 5:01 AM

48khz is a audio standard for Video, it's the default across all NLE's. I confirmed the anomaly in Vegas with your audio file at default settings. The audio doesn't play correctly on timeline, and the same distorted audio when rendered

In Premiere Pro, the audio also doesn't sound right, however it renders correctly, the way it should sound.

In Davinci Resolve the audio sounds normal on timeline, and normal when rendered. At a uninformed guess, Vegas is not resampling on timeline or render, Premiere does no resample on timeline but does on render, and Resolve does a real time resample and it's renders are resampled

Can any audio boffins tell us what's going on here, and what is unique about the bird call?

(You can download his recording from the forum, it's been trans-coded by the website I assume but still shows the problem )

EricLNZ wrote on 5/9/2022, 6:31 AM

@Milosz-Nasadowski What is the sample rate of of your original recording?

rraud wrote on 5/9/2022, 11:14 AM

FWIW, Bird calls are mostly high-frequencies which are sample-rate dependent. IF the smaple rate needs to be changed, a pro-quality resample process is recommended. The bit depth affects the dynamic range. 24 bit is usually recommended for nature sounds (floating point 32 bit, if possible, is recommenced for unpredictable SPLs..) I see the posted file is an MP3, so there is no bit depth parameter for MP3 files. In any case MP3s should be avoided if at all possible it is a final end-user format. Inherently the maximum encode bit rate is 320kbs and CBR (Constant Bit Rate) is generally better than VBR. Joint stereo is slightly lower quality than stereo... though if the file has no stereo spacial content, (spoken word) encoding a single-channel mono mode will double the apparent resolution 320kps=740kps. Of course this applies to encoding from a high-quality PCM file or newly recorded material. Otherwise 'garbage-in-garbage out' applies.

rraud, Magix Sound Forge forum moderator

Milosz-Nasadowski wrote on 5/9/2022, 11:35 AM

FWIW, In any case MP3s should be avoided if at all possible it is a final end-user format. Inherently the maximum encode bit rate is 320kbs and CBR (Constant Bit Rate) is generally better than VBR.

rraud, Magix Sound Forge forum moderator

Both, bit depth and MP3 compression is irrelevant in this case. Mp3 compression sounds more like a lisping sound. Somehow Vegas doesn't know how to handle audio files with different sample rate, adding something that sounds like phasing at high frequencies with sudden modulation, which is rare because probably only birds can do that naturally.

todd-b wrote on 5/9/2022, 5:15 PM

I've been able to replicate this effect on a 44.1khz audio timeline with your 44.1khz file by doing the following

So I think what vegas is doing is playing back the audio faster and then implementing a pitch correction that is low quality or maybe it is suitable for other audio but not bird calls. The reason I sped up the audio by that amount is due to the difference between 44.1khz and 48khz being 1.0884

This still doesn't explain why Premiere and Vegas play this audio the same on timeline, but premiere renders the correct audio, or how Resolve can playback the audio correctly and render correctly. What are they doing differently

EDIT: That's obviously not a 1.0884 increase in speed, my maths had a stroke, but it did duplicate the effect

Audio experts?