Do DSD Philips (.dff) Sony (.dsf) Sound different?

[Noodeling]

I'm sitting in front of my Hifi listening to DSD512 offline upsampled Music. By error I checked the "wrong" Button in the Civero/Xiscr Sotware and hit the "Philips .dff" Format Output instead of the usual "Sony .dsf".

 

And guess what, the Formats Do Sound different!

 

Philips .Dff is the "better", cleaner, more Natural sounding Format When playing.

 

So my questions are:

 

1. Do I Dream? Is this possible? Can someone comfirm/disproof this?

 

2. What is the General difference of the two Formats? And can this impact in any way the upsampling/digital-analog-conversation?

 

3. Maybe it's Not a General question of upsampling/format but a Special way my Dac (T&A Dac8Dsd) is handling the different file types?

 

Any comments are welcome!

[firedog]

Extremely unlikely they actually sound different.

 

The only difference in the 2 formats is the file shell, not the data itself. DSF allows for tags, .dff doesn't.

 

So there isn't any real difference in the way the DAC handles the files.

 

Have you tried listening to them when you don't know which format is playing? If you can't tell which one is which "blind" a vast majority of the time, then they aren't really different.

[mansr]

2. What is the General difference of the two Formats? And can this impact in any way the upsampling/digital-analog-conversation?

 

The headers are different.

The bit order within bytes is different.

The channel interleaving is different.

 

None of this affects the actual data contained in the files. The formats can be converted between losslessly.

 

3. Maybe it's Not a General question of upsampling/format but a Special way my Dac (T&A Dac8Dsd) is handling the different file types?

 

The DAC never sees either file format. The player software rearranges the bitstream to the format required by the DAC.

[yamamoto2002]

Also DFF supports DST lossless compression and DSF is not.

 

DoP requires MSB to LSB bit ordering, so (non DST compressed) DFF to DoP conversion process (that is performed on music playback apps) is very slightly more efficient than typical DSF to DoP conversion process which requires bit-reversal.

 

If app composes complete DoP data stream for playback on main memory before playback starts, computing task on playback becomes identical and those DSF/DFF difference disappears completely.

 

Fig.1 Two dsd streams (Left channel and Right channel) interleaved to create SoundData chunk of DSDIFF DFF file

 

Fig.2 Two dsd streams (Left channel and Right channel) interleaved to create Data chunk of typical DSF file

 

My DFF read program. DST compression is not supported on this code.

https://sourceforge.net/p/playpcmwin/code/HEAD/tree/PlayPcmWin/PlayPcmWin/DsdiffReader.cs

 

My DSF read program. There is bit-reversal table there on the last part of the code.

https://sourceforge.net/p/playpcmwin/code/HEAD/tree/PlayPcmWin/PlayPcmWin/DsfReader.cs

[Noodeling]

Problem solved!

 

First, I did blind Tests between dsf and dff.

 

Result: indifferent.

 

With Some upsampled Songs I hit 10/10 (!) with others only 4/10.

 

Second.

Some thoughts.

 

Rating 4/10 is like guessing = there is no audible difference !

 

With the 10/10 blindtest Rating there Must be a difference.

 

Third.

Conclusion.

Software used "XiSRC".

 

I additionaly did hit the "ISP Reduction" Button while upsampling to 512 .dsf!

 

That's what leads to the (Bad) different Sound!

 

DSD .dsf and .dff do not sound different!

 

Trank You all for your comments and helping/explaining!

[Noodeling]

[Dan Gravell]

Well done working it through!