Help me fix this audio

Hello to everyone,

I’m trying for the first time to upload my games play to Youtube.
The tools: recorded with Fraps; splitted in two parts with Virtualdub; encoded with Vidcoder normal preset; trying to adjust the audio with Audacity 2.0.

The issue: the audio sounds like metallic, I don’t know how to better explain it but it’s different from the the original gameplay.
I think that the problem started from Fraps because I extracted a sample from the huge (68GB) raw video clip and it has the same issue.
I uploaded an encoded version (4.40MB) online which features the issue at this link:

Could anyone please have a look at it and suggest the best steps in order to make the sound similar to the original?
Thanks in advance.

Tinkly, metallic audio is almost always caused by multiple compressions. Audacity doesn’t edit in native format. It pulls audio into itself in its own high quality format and then exports a new version of the show when you’re done. Unless you exported the show in an uncompressed format like WAV, the show will experience Audacity’s compression in addition to the original game compression and in addition to whatever Fraps is doing.

Compression systems are all designed to start with a perfect, crystal clear, uncompressed original show and produce one compressed and delivered show for enjoyment, not for opening up and more editing or production.

A recent poster whose whole show was editing on-line audio ran into trouble when they wanted him to compress it for internet delivery. The show turned to tinkly mush when he tried. Too many compression steps.


Thanks for feedback but I don’t understand.

I agree that it’s hard to understand and to deal with multiple compressed files, however, I extracted a wav sample from the raw Fraps file and the issue is there right after Fraps recording, I didn’t do any multiple compression.
There is NO other way of recording raw gaming other than Fraps, I think that it’s the only software which applies a very light compression (hence the resulting huge avi file) to the recording while other software will definitely compress it to the bare minimum.

So what trick can I use in Audacity?

The demo file seems to have 0 kb.

That seems strange, I downloaded the file and shows 4.39MB and can play in Windows Media Player, maybe the antivirus blocked it.

The file is behind a graphics “captcha” layer. I emailed the file to you.

It’s rough to know what’s normal and what isn’t. It sounds like normal game play to me.


That might be browser related issue.
Download 1.jpg
Download 2.jpg

I emailed the file to you.

I don’t understand this, where is it?

It’s rough to know what’s normal and what isn’t. It sounds like normal game play to me.

I agree, you need to listen the audio from within the game then you’ll feel the issue.

I sent the sound file to Robert. He may not be able to make it past a captcha. He can’t see them.

I agree, you need to listen the audio from within the game then you’ll feel the issue.

So we’re resolving a distortion issue on a fuzz guitar.

It sounds fine to me and if there’s nothing to compare it to, we’re stuck.


Thank you Koz, I’ll trust your judgment.
As you say, it is important to have an ideal recording to determine the difference. Is there perhaps some frequency range that is cut away?

From a spectral point of view, the distribution of the frequencies is nearly ideal.
A bass boost of 6 dB (bass and treble effect) gives 6 dB roll off for the whole range.

However, that’s not the problem. I think that the audio plays too fast.
Set the rate in the track drop-down menu to 44100 Hz (48000 Hz is checked).
Doesn’t that sound much better, in particular the female voice?

This is a sample recorded directly with Audacity.

Edit: This time I recorded 16bit and 96,000 Hz both settings in Windows and Audacity, then the sample reduced to mp3 default 128bit

I emailed the sound file to you.

So I’ve seen, thank you Koz.
It is difficult to nail down the difference. I presume that the Audacity recording is the target sound.
In direct comparison (so far as it is possible), the mp3 seems to have more sound effects (sword clanging and ringing) than the duller wave file.

This might be due to different channel balancing/-mapping. I don’t use Fraps so I can’t tell.
Using the bass and treble effect should help somewhat (+10 dB treble and perhaps a bit of bass too).
The speed is apparently the same.
By the way, 96 kHz is way too high for such a recording imo.

I can’t technically tell it but for sure Fraps added a strange effect to the sounds

Using the bass and treble effect should help somewhat (+10 dB treble and perhaps a bit of bass too).

Tried it (struggled to find the filter as I was using older Audacity 2.0) but I think that that doesn’t effect the specific issue.

By the way, 96 kHz is way too high for such a recording imo.

And here I think that you anticipated my following feedback.
I recorded at 96 KHz just to see if there was any change so indeed for gaming recording and Youtube 44K KHz is plenty of quality.
By the way, when I tried to upload the sample on the forums I faced with the size limit so to reduce its size even more I reduced the bit rate to 64bit then I listened to it and discovered that the issue I’m talking about was empathized even more.
I artificially created and empathized the audio effect (which I’m complaining about) even more so that is why I recorded at 96 KHz, I don’t know how to technically explain it.

To make a selection fit on the forum, the usual technique is to reduce the length of the show, certainly not the quality which is usually the whole point of the posting. An alternative not mentioned in the posting instructions is to post a very high quality MP3 (128 Mono or 256 Stereo instead of a WAV. You can get many more seconds of show time posted and most times, the quality difference doesn’t matter.

If we’re trying to diagnose a technical or bit-by-bit wave error, then yes, you need to post the actual WAV.


Yesterday I had a problem with Audacity 2.0 (working on 2.0.5 now) always saving same files size (which was too big) it didn’t matter how much I was cutting it then I gave up because it was a bit late.
However, here are for clarity both samples from Fraps and Audacity recording saved at 256 bit.
I think that there is some difference.
Fraps sample:

Audacity sample:

It’s a rather subtle difference isn’t it?
(the spectrum differs only marginally)
You should actually record the exact same battle scene with both, Audacity and Fraps.
Of course, this may not be possible due to driver restrictions/exclusions.
Can you tell which input devices serve as source in Fraps and Audacity respectively?
Fraps uses normally a automatic detection but you can set it to a manual input assignment in order to be sure.
Changing the fps rate might also have some impact.

Fraps settings:
Fraps settings.png
Audacity settings:

In the mean time I’ll try a different game play recording software.

I struggled so much with this piece of audio, my purpose was to upload game play to Youtube but also I’ve had chances to practice with Audacity.
I played the audio on other computers and it sounded differently, I would say just normal.
Then I checked the computer I’m working on, the Windows Media Player equalizer’s bands from 2K to 16K were set to the extreme top!

I would recommend people sharing their computers to create a personal user name profile and always log in with a password, also one person should act as an administrator and set the proper limitations to all users so that they may not change system’s settings.
When dealing with video and audio editing it’s of vital importance to know your computer and mostly important not to let anyone change any setting.

Audacity is an amazing software and it’s high recommended in professional private institutes and universities, last but not the least this forum is a vital core for the support of the software.
Thanks everyone for giving professional support here and for your patience!