WAV Amplitude prior to decoding

[ClarkNovak]

I've been futzing with this desktop decode of Carole King's "Music" from QS-encoded CD using Kempfand's Au3.0 script. I finally got the phasing figured out (yes, it was as simple as inverting the back two channels), but: I'm hearing some artifacts in the decoded audio that sound like audio clipping at FSD (full scale digital), yet the waveforms are nowhere near 0 dB.

Could this clipping be a result of the source being too hot during the decode? Should I decrease the volume of the source WAV before I run the script on it?

 

[Bob Romano]

I always though you needed to decrease the wavs by 3db for SQ decoding but not QS. With Lucanu's SQ script, no decrease is necessary. I have never had to decrease anything during a QS decode.

 

[Old Quad Guy]

I thought we didn't need to decrease the SQ wavs by 3db after the release of Adobe Audition 2.0 & 3.0. Am I wrong with that assumption?

One thing I've discovered with SQ and QS decoding is I’ve gotten much better results decoding with a much lower recording into the computer. With the peaks not even hitting -20db with some recordings. Then raise the volume after decoding. Try lowering the volume on the song’s file that’s causing the audio clipping and decode again.

 

[ClarkNovak]

OK, running a new decode now. I'll let you know what I find.

 

[kap'n krunch]

I always though you needed to decrease the wavs by 3db for SQ decoding but not QS. With Lucanu's SQ script, no decrease is necessary. I have never had to decrease anything during a QS decode.

Ditto here.

Even though I record pretty hot (some shaving off peaks here and there),
usually after decoding (with Lucanu's SQ script with PhaseBug) the signal is much lower-about 3-4 db....

 

[ClarkNovak]

Well, I ran the decode script again, first reducing the level of the input WAV file by 3dB; nothing peaking over -30. There was no change in the quality of the output audio; the clipping distortion is still present in both rear channels.

I think, as the saying goes, that the demux process has "revealed the limitations of the original source material" -- I think maybe occasional clipping on background vocals was present but buried in the stereo mixdown. It's all I can really think of to explain what I'm hearing.

 

[Key]

Well, I ran the decode script again, first reducing the level of the input WAV file by 3dB; nothing peaking over -30. There was no change in the quality of the output audio; the clipping distortion is still present in both rear channels.

I think, as the saying goes, that the demux process has "revealed the limitations of the original source material" -- I think maybe occasional clipping on background vocals was present but buried in the stereo mixdown. It's all I can really think of to explain what I'm hearing.

Yes this will happen. And it will eff with your mind thinking it's the decoder causing distortions. What can happen is certain low level out of phase distortions can be introduced at the mixing and mastering stage with compressors amongst other processing. If it is out of phase the decoder will send it to the rears. Other things like tape splicing and tape distortion can show up in the rears as well. I am weird I could always tell the difference between different kinds of distortions like analog vs digital and preamp vs tape etc... So I just learned to ignore it and trust my math and decoder when I know I am not causing the distortions.

But also be aware that with compressed and maximized CDs all bets are off in terms of where to set a decoder volume. Because of the mastering techniques used on most all CDs now they actually take the volume 5dB or more above 0dB. I thought I had perfected the level setting on a decoder until I found this one CD that just expanded like crazy when decoded and then later found others. To be clear on this sometimes you will be internally clipping the signals even though the output is technically not clipping. Any mistep in a complicated decoder can cause that one step in the chain to clip because it should have been dropped in volume 6dB before the processing. In order to fully insure you are not doing this you should drop the volume 6dB -12dB or MORE before running it through your decoder.