Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Bebop0502]

In Sweden there is a discussion on a forum claiming that JRiver v25 creates an unacceptable noise floor when converting from pcm to dsd. The computer used is an HP with W10 Intel Core i5-8250U CPU 1.60GHz 1.80GH, RAM 8GB, x64 based processor.

A 1kHz 16/44,1 square wave was used:

https://www.dropbox.com/s/ip06gmh8qkazff8/1kHz-square-16-44-flac-FFT.thumb.jpg.969cb6799f26bf0581b454afdb7dc639.jpg?raw=1

Then the wave was converted to dsd64 giving a result showing a noise floor between 36 - 48 dB as seen in this link.

https://www.dropbox.com/s/njj89l3ytbqtld2/1kHz-square-16-44-dsd64-FFT.thumb.jpg.1ac87e04caee1b00de5ea4d60e8d9728.jpg?raw=1

When going back converting from dsd to pcm wav-file it looks similar.

https://www.dropbox.com/s/lnkj15e4lb4o0ip/1kHz-square-16-44-dsd64-pcm1644-FFT.thumb.jpg.a8b1447008ac53be9d232e5feccd0463.jpg?raw=1

I know you recommend to play direct without conversion but the feature is there an should perform better than this, in my view. Another question is if the result is as poor doing conversion from 16/44 to 24/176? Or is there something hidden in these graphs, do you think?

[tij]

it’s nearly impossible to generate perfect square waveform

it’s funny how they show theoretical square wave form ... they should display actual square waveform that was fed to JRiver ... that “noise” level might have been present in their actual square waveform from the start