Oversampling /Follow Source Frequency

[john chung]

I have recently converted some songs to 24-bit 48kHz 2.3Mbps WAV format and wonder if I still need to choose between Oversampling /Follow Source Frequency.
Also, Can Crossfeed be set as a default DSP effect? Everytime I open the app or connect to a dap, the Surround Sound icon appears and I have to reset it to Crossfeed. Thanks!

[blaubär]

Everytime I open the app or connect to a dap, the Surround Sound icon appears and I have to reset it to Crossfeed.

Perhaps you have activated "Settings > Connection and Controls > Auto-switch DSP Effect" ? Try disabling it.

[blaubär]

I have recently converted some songs to 24-bit 48kHz 2.3Mbps WAV format and wonder if I still need to choose between Oversampling /Follow Source Frequency.

That depends on the sampling rate that you want Neutron to use for its internal calculations and transfer to the DAC. The default values ( both off, Frequency default ) should work fine. In the case of music with 48 kHz sampling rate and a DAC that supports this ( nearly every DAC will do that ) Neutron then won't change the sampling rate and transfer 48 kHz to the DAC.
---
Another question is why you converted songs to 48 kHz - what sampling rate did they have and why didn't you use that ?

[john chung]

All my songs were downloaded from youtube in either mp3 or opus format. Some experts recommend that 24-bit 48kHz is good in either Flac/Wav format. I am using a Hidizs S9 Pro dap which promises up to 32-bit 768kHz and dsd 64/128. So I comverted most of the songs to 48kHz (Some in 96/192kHz) and played them on bit-perfect. I found it sound better in general. More depth in music background and individual music instrument more distinct. However, the vocal sound sometimes a bit diminished and distorted, may be it was the sound source. I have also seen in the forum that Pre-amp settings can affect the audio quality, Is it true ?
So, what is your advice ? ( I am using Goldwave for convertions )
16/24 bit
44.1/48 kHz
Flac/Wav
I am using a Huawei Media pad.
I also tried PCM to DSD and it came out 1-bit 2.8MHz Stereo DSD64. It sounds ok .
Are there any best settings for using bit-perfect and PCM to DSD ?
By the way, I have disabled the auto-switch on DSP. Thanks.

[blaubär]

Well ... when I download from youtube I use the Firefox browser with the "Easy Youtube Video Downloader Express" addon, that gives me mp4-files. The mp4 is a container format that includes both video and audio. I'm interested in the audio and won't convert that further, so I extract the audio part as follows :

Code: Select all

set -e
export LC_NUMERIC="en_US.UTF-8"

while (( $# > 0 ))
do

    inpNam="$1"
    basNam="${inpNam%\.[^.]*}"
    outNam="${basNam}.m4a"

    if [[ ! -f "${inpNam}" ]]
    then

        echo
        echo "Fehler: \"${inpNam}\" existiert nicht"
        echo

    else

        echo
        echo "processing \"${basNam}\" ..."
        echo

        ffmpeg -hide_banner -i "${inpNam}" -vn -acodec copy "${outNam}"

        echo
        ls -lh "${inpNam}" "${outNam}"
        echo

    fi

    shift 1
done

This yields an m4a file, which is another container, but with an empty video. I add sensible tags to that file using mp3tag ( and sometimes create a cue-file to get the individual songs ) .

Screenshot 2022-11-19 104808.jpg (45.87 KiB) Viewed 1112 times

And that's the file that neutron plays. I generally activate Oversampling with factor 2, so that Neutron uses 88.2 kHz sampling rate.

[rauschkind]

All my songs were downloaded from youtube in either mp3 or opus format. Some experts recommend that 24-bit 48kHz is good in either Flac/Wav format. I am using a Hidizs S9 Pro dap which promises up to 32-bit 768kHz and dsd 64/128. So I comverted most of the songs to 48kHz (Some in 96/192kHz) and played them on bit-perfect. I found it sound better in general. More depth in music background and individual music instrument more distinct. However, the vocal sound sometimes a bit diminished and distorted, may be it was the sound source. I have also seen in the forum that Pre-amp settings can affect the audio quality, Is it true ?
So, what is your advice ? ( I am using Goldwave for convertions )
16/24 bit
44.1/48 kHz
Flac/Wav
I am using a Huawei Media pad.
I also tried PCM to DSD and it came out 1-bit 2.8MHz Stereo DSD64. It sounds ok .
Are there any best settings for using bit-perfect and PCM to DSD ?
By the way, I have disabled the auto-switch on DSP. Thanks.

you can not improve the quality of an audio file by converting it to a higher resolution, information that is not there can not be restored this way and the conversion will further reduce the quality of the tracks.
just leave the files as they are. oversampling option in neutron might gain you theoretical advantages, expecially if you use dsp effects.

[blaubär]

[...]
I have also seen in the forum that Pre-amp settings can affect the audio quality, Is it true ?
So, what is your advice ? ( I am using Goldwave for convertions )
[...]
Are there any best settings for using bit-perfect and PCM to DSD ?

As you see in my last post I don't do conversions ( of the audio data, not the container ) if I can help it. The reason is that every ( lossy ) conversion necessarily decreases the quality. Which doesn't mean that it cannot produce pleasant effects, but I like to keep the quality as good as it is and then, if I think I need to change the audio, do things where I know what I'm doing, e.g. use crossfeed, equalize or frequency response correction.
---
The preamp affects the audio quality : if too high then the eq/frc might produce distortions, if too low you reduce the dynamics. I suggest to use balanced eq and frc presets, where the preamp is set in such a way that the maximum amplification of the bands doesn't exceed 0 dB.
---
The best settings for PCM to DSD would probably be to switch it off. Unless of course it produces some nice effect that you happen to like. Generally the DAC can handle PCM, and it is not plausible to assume that Neutron can do the conversion better than the DAC, which has got an optimized hardware for that ( assuming that a DAC that accepts DSD does this because it will transform the PCM into DSD anyway ).

[john chung]

Thanks for the replies. I used to convert songs using online converter which normally took less than a minute to do the job. I now use Goldwave as it will decompose the song first and resample to the reuqired frequency. It takes a longer time and much more storage spaces. It looke like Goldwave is doing a thorough and professional job inorder to achieve better audio quality. I shall compare the difference.

Balanced EQ: Do you mean the EQ presets on Neutron player, without sliding the knobs up or down?
I am using the 4-band parametric (default) EQ and frc, I do play around with the slidings and preamp voloume but keep the volume below 0db.

As to PCM to DSD, I do find they sound better on most songs, especially on music tracks which were recorded 50-60 years ago, like the Ventures guitar and some chacha music. Again, are there any setting for this function?

One more question. From time to time I need to clear the cache and the data on the app as I found out the songs begin to deteriorate when the cache started to build up, or just that I mess around the EQ too often. Thanks again.

[MotleyG]

Thanks for the replies. I used to convert songs using online converter which normally took less than a minute to do the job. I now use Goldwave as it will decompose the song first and resample to the reuqired frequency. It takes a longer time and much more storage spaces. It looke like Goldwave is doing a thorough and professional job inorder to achieve better audio quality. I shall compare the difference.

It is not possible to restore audio quality once a lossy format has compressed it. It doesn’t matter what software you use.

[blaubär]

Balanced EQ: Do you mean the EQ presets on Neutron player, without sliding the knobs up or down?

For any eq- or frc-preset where any band has any setting different from 0 you might want to adapt its dedicated preamp-level so that the resulting amplification for no frequency execeeds the maximum possible value of 0 dB.

Screenshot 2022-11-19 164828.jpg (53.88 KiB) Viewed 1081 times