I have: MSB Reference + DD and Renderer and I2s HDMI inputs, Antipodes K50 streamer.
Reading about many people here on the MSB Forum who appreciate PGGB files, I installed the program to do some tests.
A problem emerged: the oversampled files at 706/768 all sound distorted on I2S and perfect on Renderer. Even with Roon’s oversampling it’s the same: if I send it to the Renderer sound OK, if I send it to I2S it distorts sound…
The same problem also occurs with MinimServer + MPD + JPLAY. I haven’t tried Squeeze + Squeeze
So it’s not a problem of PGGB (which I didn’t like anyway!!) …
Since the Renderer is connected to the Antipodes server and I2S is connected to the digital outputs of the Antipodes Player, could it be a technical limitation of the K50’s digital outputs?
The Renderer is the only MSB input module that is compatible with 32 Bit, every other module is 24 Bit, so if your PGGB files are 32 Bit this is the reason.
Hi Mark: I’m aware of the 24/32 bit argument so with PGGB I created both 24 and 32 bit files: I can guarantee that they are correct (verified with JRiver).
Roon has two different outputs: one directly to the Renderer with a 32-bit setting, the second to the I2S with a 24-bit setting: it is precisely the frequency of 706/768 that is not good !!
I asked the same question to MSB: according to Jonathan it is the K50 that on the I2S digital output does not send the 706/768 frequencies correctly… unfortunately on the DAC I don’t have a USB input to be able to check another output (the SPDIF is limited to 192).
It is an attempt to add the information lost in the original analog to digital conversion by the use of special filters and upsampling. iirc. To my ears there is an improvement in both transients and timbre. I especially enjoy the added air. I use an mscaler for streaming and pggb (16fs pcm) for local files. ymmv.
The goal is to recreate the original analog signal just before it was converted to digital. Sampling theories like Whittaker–Shannon tell us how to achieve perfect reconstruction - though in the real world perfection isn’t possible but some methods can get us closer than others.
Most DACs due some form of reconstruction. Software methods aim to improve upon what our DACs can achieve, but the benefits can be hindered if the DAC’s own processing gets in the way.
My DAC, the T+A DAC 200, has a pure path for both PCM and DSD so it allows PGGB’s improved reconstruction accuracy to be clearly heard. To my ears the improvements aren’t subtle - especially so when scaling to DSD512.
I think it depends on the DAC: on the internal processing and the filters that are implemented in it. Some time ago I did tests with HQPlayer but I have always preferred the native format of the files (my dac is an MSB Reference with Digital Director)… now I’m trying with FGGB (since many have expressed positive opinions)… I’ll see how the tests go !!
It will be interesting to see how the soon to be released new Chord Quartet scaler sounds because this can process the music in real time albeit at a hefty price (rumoured to be £25k UKP). Also of course it depends on being connected to the Chord Dave via dual bnc cables. I did hear it last weekend but it was in a show setting and it was not possible to form any meaningful assessment of the sound.