DSD file view vs. WAV file view, what am I seeing?

[JonUrban]

I ripped an SACD with the Oppo/SACD Extract and ended up with .dsf files. So rather than converting them with Foobar, I opened them in Sound Forge, which now will read .dsf and .dsd files. Cool.

So I ended up with a file like the one below (I only snipped out one tune). You can see there isn't much definition in the wav graphics:

But when I dropped it down to 24/96, I now could see all of the definition in the wav graphics. Is this just an anomaly with Sound Forge, or is the display seeing something at 352.800/96 that's not there at 24/96? Not a big deal, but to me it's a bit curious.

 

[MrSmithers]

Interesting - I wasn’t aware of a DAW that could import native dsd files?! Can Sound Forge then export .dsf files?

@jimfisheye gave quite a quite a useful titbit on this when converting dsd > wav > flac.

Converting DSD Files to FLAC

I hate to get mired in any of these discussions, but anyone who says that different FLAC encoders produce different sounding files is selling you audiophile snake oil. It's a lossless 1:1 codec, so what you put in is exactly what you get out, bit for bit. It's like suggesting if you compressed...

www.quadraphonicquad.com

Perhaps it's something to do with Normalisation? I've only just started to play with this but initial tests sound pretty good!

 

[steelydave]

The flaw of DSD is that it has massive ultrasonic noise starting around 30-35kHz - that's why a lot of SACD players have low-pass filter options, and programs like foobar allow you to apply various low-pass filters too.

Most likely what you're seeing is that the 24/96 has been filtered thusly (or there's simply less ultrasonic noise to waveform plot since the upper frequency response of 24/96 is 48kHz) and the 24/352 hasn't. I think if you did spectral analysis of the two versions you'd find the 24/352 has way more sonic garbage above 30kHz, which when plotted as a waveform just looks like a big dense log that obscures all the usual peaks and valleys you'd expect to see.

 

[JonUrban]

Interesting - I wasn’t aware of a DAW that could import native dsd files?! Can Sound Forge then export .dsf files? ....

Apparently so, although I have never tried it. I am not a huge .DSD fan. I can appreciate it, I just would prefer PCM. It's what I'm used to, and more comfortable with.

 

[MrSmithers]

Apparently so, although I have never tried it. I am not a huge .DSD fan. I can appreciate it, I just would prefer PCM. It's what I'm used to, and more comfortable with.

View attachment 39060

I was always a sucker to believe the hype from those Sony SACD inserts! Still the nostalgia has me thinking the same thing! But totally on the same page now... if you hear some of the well mastered blu rays it's hard to know how you could tell a difference? And in the digital world PCM is a lot more convenient!

 

[bluelightning]

The flaw of DSD is that it has massive ultrasonic noise starting around 30-35kHz - that's why a lot of SACD players have low-pass filter options, and programs like foobar allow you to apply various low-pass filters too.

Most likely what you're seeing is that the 24/96 has been filtered thusly (or there's simply less ultrasonic noise to waveform plot since the upper frequency response of 24/96 is 48kHz) and the 24/352 hasn't. I think if you did spectral analysis of the two versions you'd find the 24/352 has way more sonic garbage above 30kHz, which when plotted as a waveform just looks like a big dense log that obscures all the usual peaks and valleys you'd expect to see.

I thik you are right on the money. DSD has prodigious amounts untrasonic noise which are more than likely the cause. Makes the waveform look much noisier, thus masking the actual signal.

To me, DSD is, from an engineering perspective a completely useless format.