DSD Conversion: dBpoweramp or JRiver?

This is ultrasonic noise, correct? (So Far Away from MFSL's Dire Strait's Brother's in Arms SACD, converted to 24/96 by dBpoweramp)



Should I be using a filter for these SACDs that isn't being applied by default?


Also, moving up to 24/96 caused my usual method (scan with foobar2000, then apply amplification rounded to the closest dB without clipping, in this case +6 dB) to clip with this on So Far Away, while 24/88.2 didn't cause this. Hm.

Actually, according to audacity it doesn't clip, just hits right at 0 dB. Huh. Guess I need to convert, then scan with foobar2000, then apply the amplification.

 

Brothers In Arms on SACD is a special case. The original was recorded with early digital in 16-bit PCM at 44.1. There isn't going to be any musical high frequency content above 20 kHz from the digital master. However, I believe the MFSL SACD was crated from an analog tape made from the original digital master. Then the SACD mastering was done from that analog tape. So anything above 20 kHz in that spectrogram could be considered ultrasonic noise that's not in the original. The reason for doing a higher sampling rate capture of the analog tape is to better capture the transients and timing and other elusive sonic elements present on the analog tape. And that's the real reason for doing high-res captures of analog tape in the first place. The reason for high-res isn't to capture sounds that only bats can hear. The reason is to better capture the transients and timing and other elusive sonic elements below 20 kHz that we do hear. High-res does make a difference.

So don't get aggressive and filter away that ultrasonic noise and end up degrading the transients and timing and elusive sonic elements present on the SACD.

 

Ah! You're right. Should've thought about that. What is that section of noise above everything, though?

 

Tape noise? Mastering chain electronics noise?

 

I use an app called DSD Master to covert my SACDs to 176KHz/24-bit FLAC files.

 

I used JRiver for some time, converting DSD to PCM 24/176 (my convertor at the time could not handle DSD). It sounded good.
Now, converting SACD image ->DSF files was poor, as it is involves intermediate conversion to PCM and back.
Now my Auralic Aries Femto converts on the fly (and does really good job at that) and I no longer use any PC convertor...

 

Can your DAC play DSD? Is the Auralic passing PCM or DSD to the DAC and at what resolution?
I'm always curious about other ways of processing or converting DSD. Especially if that method may end up sounding better.

 

I sold recently my DAC with DSD support, and now using AYON CD-5 as DAC, whihc is a fantastic DAC, but no DSD and tops at 24/192.
I set my Auralic Aries to convert DSD64 (and up) to PCM 24/176. There are also some filter choice to be applied when resampling. All resampling BTW is done by Aries, I think it is way more advanced then 15 year old resampling in AYON.
It also does "restoring" MQA into corresponing 24/96 or 24/192, had parametric EQ (new update), etc. Nice little device, not without some quirks, but works fine as server/streamer for several years... And sounds as a source for DAC better then my OPPO-105.

 

AYON is nice. I got to hear a Skylla DAC in a really good headphone system (tubes in the DAC, full tube amp with 300B). Sonic holography bliss.

 

What software did you use to generate this graph?

 

That is Spek. A free and open source cross-platform acoustic spectrum analyser.

 

I'm curious about this. I've understood DSF to be the two channel container of DSD that also allows tagging. Where DFF can be the multi or two channel container but does not store tagging. I'd like to understand more about how a DSF was being converted from ISO to PCM to DSF.

Edit, a little more digging, DFF includes data compression and DSF does not. So DFF can have smaller file storage space.

 

I communicated with JRiver on this issue some time ago, and they say that all conversions invloved PCM, and so we have ISO->PCM->DSD(DSF).
Loss of quality was significant, and I used sacd_extract.exe instead to split into dsf (sacd_extract.exe -2 -s -P -i"file.iso").
It worked fine, and no quality loss - but due to some block size issues can result in tiny click at the beginning of the song.
All it because oppo 105 refused to play ISO images.

 

Oh, it's a JRiver thing when playing ISOs as DSFs. That makes sense. I use ISO2DSD software to convert them to DSFs so JRiver isn't involved.

I started out ripping to ISOs but this year have been ripping right to DSF and not saving the ISOs. Wish I would have been doing that all along.

 

DSF stereo file I believe are compressed as result (but with some metadata), but multi-channel tracks are indeed huge.

 

Both DSF and DFF can support multichannel as well as stereo.

Not sure what you are asking. JRiver, for example, uses a PCM intermediary when converting among the DSF/DFF formats/sampling rates and that is to be avoided, if possible.

Nope. DST includes data compression while DSF/DFF should not.

Actually, it is a JRiver thing when converting ISOs to DSFs but not when playing ISOs as DSFs.

Nope. See above. But, yes, multichannel DSF/DFF files can be huge. My largest ones are in excess of 10Gb.

 

I have no problem with that and I routinely convert DSD to PCM (on the fly) so that I can use room EQ.

 

DSF is not, in and of itself, compressed. DST is but it is lossless. Number of channels is not relevant but, obviously, compression is more likely to be needed with more channels.

 

I'm puzzled by this -- if the original recording was 16/44.1 digital, and then there was an analog copy made of it, how could that analog copy of the digital contain any information outside the scope of what can be captured, recreated, etc. with Redbook digital? (I get why it's useful to make a hi-res digital master from analog tape when analog tape was what was used to make the recording in the first place.)

Second question: Why can't transients and timing and other elusive sonic elements whose frequency is less than 22kHz be captured and recreated by recordings with a sampling rate of 44.1kHz?

Final question, for now, for everybody: What is everybody using to rip SACDs? Is the old, not-upgraded PS3 still the only option?

 

I'm using a Pioneer BDP-80FD Blu-ray player that is able to rip SACD if you plug in a properly configured USB thumb drive with the right script files. As described in this Audiophile Style thread: SACD Ripping using an Oppo or Pioneer? Yes, it's true!
That thread mentions the other Blu-ray players that are also able to rip SACD.
You no longer need the special old PS3 to rip SACD. Some special Blu-ray players are able to do it too.