Page 1 of 1

Old GoldWave 5.16 Did it EQ/resample properly?

Posted: Wed Nov 01, 2017 10:24 am
by efc1978
Hi,
I've randomly come across some info here from 10 years ago saying that GoldWave didn't do any filtering, and would just do straight resample (this was a discussion about downsampling from 96 to 44.1). It says "v5.50 and later use a high quality polyphase resampling algorithm".
So it got me thinking, back then using an old 5.16 version I EQ'd a bunch of different music (original 16bit/44.1 and saving as 16bit/44.1), so does this mean the stuff I EQ'd using version 5.16 would not have been properly resampled and contains a bunch of hiss or other noise that a more modern version of GoldWave wouldn't produce?
Was this old issue only when downsampling eg 96 to 44.1, or did it effect any kind of resampling eg EQ'ing a 44.1 file and saving as 44.1?
Thanks

Re: Old GoldWave 5.16 Did it EQ/resample properly?

Posted: Wed Nov 01, 2017 12:16 pm
by Tristan
The only thing I know is that, when I first started using GoldWave, it produced audible conversion artifacts when I reformatted my files for CD production. As time went on, the developer worked the problem out of the program. Now everything sounds fine.

Don't know about EQ'ing. If there ever were problems, I didn't notice them.

Re: Old GoldWave 5.16 Did it EQ/resample properly?

Posted: Wed Nov 01, 2017 1:19 pm
by DougDbug
Apparently, GoldWave used to measure badly in tests reported by src.infinitewave.ca. But, I've used GoldWave for a long time and I never noticed it sounding bad. (Of course, I don't resample every time I use it.)
(original 16bit/44.1 and saving as 16bit/44.1),
No resampling there. And, you only get aliasing when you downsample (i.e. Going from 96kHz or 48kHz down to 44.1kHz.) and only if there are frequencies above the Nyquist limit (half the sample rate).
so does this mean the stuff I EQ'd using version 5.16 would not have been properly resampled and contains a bunch of hiss or other noise that a more modern version of GoldWave wouldn't produce?
Aliasing doesn't produce hiss, it produces "false frequencies". That is, if you are downsampling to 44.1kHz and there are "audio" frequencies above 22.05kHz, those higher-frequency tones will be aliased-down to some frequency below 22.05kHz, and they may become audible. Thankfully, those higher frequencies tend to be low-level (so the alias is low-level) and all DSP programmers know to filter before downsampling, so those frequencies are at least attenuated if not eliminated.

Re: Old GoldWave 5.16 Did it EQ/resample properly?

Posted: Wed Nov 01, 2017 6:49 pm
by efc1978
DougDbug wrote:Apparently, GoldWave used to measure badly in tests reported by src.infinitewave.ca. But, I've used GoldWave for a long time and I never noticed it sounding bad. (Of course, I don't resample every time I use it.)
(original 16bit/44.1 and saving as 16bit/44.1),
No resampling there. And, you only get aliasing when you downsample (i.e. Going from 96kHz or 48kHz down to 44.1kHz.) and only if there are frequencies above the Nyquist limit (half the sample rate).
so does this mean the stuff I EQ'd using version 5.16 would not have been properly resampled and contains a bunch of hiss or other noise that a more modern version of GoldWave wouldn't produce?
Aliasing doesn't produce hiss, it produces "false frequencies". That is, if you are downsampling to 44.1kHz and there are "audio" frequencies above 22.05kHz, those higher-frequency tones will be aliased-down to some frequency below 22.05kHz, and they may become audible. Thankfully, those higher frequencies tend to be low-level (so the alias is low-level) and all DSP programmers know to filter before downsampling, so those frequencies are at least attenuated if not eliminated.
Hi,

The following 2 bits combined is why I asked if EQ'ing with the old 5.16 version back then would have produced undesired audible sound in my projects....

First - I thought even when importing a 16bit/44.1 file and EQ'ing it then exporting as 16bit/44.1 was resampling?, as I have always read software such as Audacity or Adobe Audition even when working with a 44.1 file does the processing at 32bit float etc., then on export to the original 44.1 the file is resampled?
So did the old Audacity 5.16 work differently to that?

Second - You said all programmers know to filter before downsampling, however I read on this forum from a very long time ago (around when I was using the 5.16 version) this comment which suggests back then GoldWave didn't filter when downsampling?
Here's what I read:
"To the best of my knowledge, Goldwave doesn't do any kind of filtering when doing a resample..it does just a straight resample...it's why if you don't pre-filter before downsampling...you get horrible results. While it's not in the help file (that I can find at least), I do recall Chris (becuase I assume he's writing the help files) saying you needed to pre-filter to avoid quantization (or some other word i can't remember in my condition) noise..this MAY have been in the manual and I remember it from the one time I read it, it may have been from a help file years ago...honestly, i've not really converted sample rates down to 8000 in a while...in fact, most of the stuff i'm doing is higher quality and i get much faster results using command-line utilties if i just want to do a straight sample-rate conversion...when i downconverted 192khz MLP DVD-A audio to 96, I just used r8brain...it was one command-lets-go deal..."

I don't have much knowledge about this sort of stuff, the 2 parts above is why I'm not sure if EQ'ing with the old 5.26 I used to use would have produced undesired audible sound in my files?
Thanks for the info.

Re: Old GoldWave 5.16 Did it EQ/resample properly?

Posted: Thu Nov 02, 2017 8:11 am
by efc1978
Are there any issues with the latest version GoldWave creating unwanted audible sound when using and saving after having used Graphic EQ or Parametric Equalizer?
Thanks

Re: Old GoldWave 5.16 Did it EQ/resample properly?

Posted: Thu Nov 02, 2017 12:01 pm
by Tristan
You're being awfully tentative in your research efforts. Wouldn't it be better just to install the latest version of GW and experiment yourself? I assume you have a license. If not, try GW in demo mode.

It's been a long, long time since I read anything about GW artifacting. We're talking years here. The EQ'ing features in GW seem essentially glitch-proof.