Skip to main content

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: Are downsampling and upsampling in ReplayGain Scanner a matter of "either/or"? (Read 1393 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Are downsampling and upsampling in ReplayGain Scanner a matter of "either/or"?

Hi,

I've done a quick search because it's AGES since I used ReplayGain (decades maybe) and I forgot all about it.
I saw that there are two main options in Foobar, one to have everything downsampled (because allegedly the inaudible sounds in high resolution material would interfere with the scan), and the other which does the opposite, and upsamples everything to find the true peak.
In a tutorial on another website it's recommended to deselect the downsampling and select the true peak instead, with upsampling at 4x. So I thought that you either select one or the other.
But the settings in Foobar allow for both things to be selected together.
So now I'm unsure, is it a mistake in Foobar and it should not be allowed to have the downsampling and upsampling selected at same time?
Or does selecting them both simply make the process longer because there would be a pass with downsampling and another one with upsampling?

And how do I know what to choose?
Thanks
Or would that just

Re: Are downsampling and upsampling in ReplayGain Scanner a matter of "either/or"?

Reply #1
The True Peak scanning and downsampling are separate settings and can be enabled at the same time.

If you use clip prevention you will have better results by using True Peak scanning. If you don't care about clip prevention then this option doesn't matter to you. This upsampling is only used when scanning peaks, it doesn't affect the gain calculations.

If you care about using ReplayGain and achieving audibly same playback level, then I'd recommend using the downsampling option. Loud inaudible content in hires material can otherwise skew the scanning results and make the RG lower the loudness too much.

All resampling options will reduce the scanning speed. The faster the used resampler the less the speed will be affected.

Re: Are downsampling and upsampling in ReplayGain Scanner a matter of "either/or"?

Reply #2
The faster the used resampler the less the speed will be affected.
Thanks.
From your answer I understand that these settings also affect any additional 3rd party RG plugin that I might use, like i.e. the Alternative Replay Gain?
Otherwise I can't put your answer in context, because there is no such a thing as a clip prevention option in the built-in RG scanner.
If there is, please point me to it, because I don't see it. I only see the choice between scanning per-file track gain, or as a single album or as an album by tag.

Thanks

Re: Are downsampling and upsampling in ReplayGain Scanner a matter of "either/or"?

Reply #3
The options affect other components that use the built-in RG scanner service with proper flags. The peak options for example fully affect the foo_dsp_replaygain component, but my True Peak scanner component that can also do RG scanning has its own peak scanning options. That is unaffected by the built-in RG scanner's peak scanning settings.

Clip prevention is not a scanning time decision. With built-in ReplayGain you control clip prevention with ReplayGain options found under 'Preferences' -> 'Playback'. The 'Processing' dropdown menu allows turning all processing off, just applying gain, doing both gain and clip prevention or performing just clip prevention.

Re: Are downsampling and upsampling in ReplayGain Scanner a matter of "either/or"?

Reply #4
With built-in ReplayGain you control clip prevention with ReplayGain options found under 'Preferences' -> 'Playback'.

Ah, fuck, I totally missed that section of the preferences!
I actually saw it several times but today I could not find it, and I forgot that it exists.
Thanks.

I was about to ask you what is the component that you mention, but I've found your page foobar.hyv.fi :)
So, let me understand, in that page you have a mix of your own work and other people abandoned 32bit components (like Luigi's alternative RG) which you transformed into 64?
Any thought about adapting the Dolby Headphones Wrapper to 64?
That component still has lot of fans.
I personally keep using Foobar in 32 bit just because of that component...

Anyway, I will try your True Peak. Thanks.

Re: Are downsampling and upsampling in ReplayGain Scanner a matter of "either/or"?

Reply #5
My page lists most of my components and some component recompiles/modifications I have made. But luigimercurio's "Alternative ReplayGain" (foo_arg) is not one of them. That component should not be used by anyone as it breaks all ReplayGain operations. It replaces the native ReplayGain functionality with something giving entirely incorrect results and prevents third parties from using ReplayGain service entirely.

There are no sources for Dolby Headphone wrapper and I'm not a reverse engineer expert like Yirkha was. I have no idea what to do with the functions exposed by the dll. Also personally I have no use for that functionality, I hate headphone sound that tries to mimic speakers.

Re: Are downsampling and upsampling in ReplayGain Scanner a matter of "either/or"?

Reply #6
Oh, I love virtual soundstage. At first it's a bit weird but after a while when the brain gets used it's for me the new normal and then I can't do without.

My bad, when I first mentioned the alternative RG I was referring to Luigi's, and when you answered mentioning foo_dsp_replaygain and I saw on your website that you call it alternative replay gain and the initial release was also in 2018, I thought it was a recompiled version of Luigi's.
Thanks for warning me.
But why do the Foobar's higherups allow it to stay in the official components if it sucks so much?

Before I open a new thread just to ask this, can you tell me if there's a list of official components that should be avoided?
And, apart for the official components and your website, is there another (reliable, safe) source of components?