Print

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

[Dr.Ames]

Hello all,

just installed the latest version under Windows 11 and selected my external USB connected DAC "Project S2" as default via the ASIO interface. I have DSD bitstreaming enabled and DSD files are played as such without any problems.

Now to my question, which has also been addressed in the past here in the forum. Is it still true for you that it makes no sense from a sonic point of view to upsample PCM files to DSD x8? My AMD 12 core processor is only used to 2% capacity even then - so that wouldn't be a problem.

many thanks and Greetings from Germany

[Awesome Donkey]

In my opinion there's no point, I've done it before many times in the past and I cannot hear a difference at all. On top of that it's a lossy conversion (PCM to DSD, or DSD to PCM) AND the conversion on-the-fly is very computer resource intensive so if you ask me it's best to leave the source material as-is whether it be PCM audio or DSD audio, especially if you care about bit-perfect output.

[Dr.Ames]

Thanks for your assessment - with Foobar I used this a lot. In fact, the interference-free playback of native DSD works better here in JRiver.

[Dr.Ames]

I would have liked to ask one additional question. Does it make sense to load music files into RAM to play them from there, especially because I play some very large DSD albums. Some tracks have a maximum size of 8.8 GB (Bruckner 9th Symphony - DSD512).

My music library is not on an SSD and therefore it might make sense? Would the decrypted single track setting make sense then?

[JimH]

There should be no difference.  Either should work.

[Awesome Donkey]

I personally find memory playback to be a hindrance when playing large files, versus smaller compressed FLAC files so I mainly keep the feature off.

[kr4]

I would have liked to ask one additional question. Does it make sense to load music files into RAM to play them from there, especially because I play some very large DSD albums. Some tracks have a maximum size of 8.8 GB (Bruckner 9th Symphony - DSD512).

My music library is not on an SSD and therefore it might make sense? Would the decrypted single track setting make sense then?

I have many similarly large tracks, keep my files on a NAS and, after much experimentation, have settled on not using memory playback.  MC manages the buffering without issues.

[eve]

Hello all,

just installed the latest version under Windows 11 and selected my external USB connected DAC "Project S2" as default via the ASIO interface. I have DSD bitstreaming enabled and DSD files are played as such without any problems.

Now to my question, which has also been addressed in the past here in the forum. Is it still true for you that it makes no sense from a sonic point of view to upsample PCM files to DSD x8? My AMD 12 core processor is only used to 2% capacity even then - so that wouldn't be a problem.

many thanks and Greetings from Germany

*In Theory* Yeah, depending on what D/A you have, specifically which chip it utilizes and how things are implemented. Up sampling to DSD at the source can be arguably beneficial. See internally, in D/A chips like ESS's high end offerings even PCM streams technically get converted to multibit SD. If you feed them the signal they end up converting to, in theory we can tailor how that process happens at the source rather than relying on the D/A. In practice? No. There's no major difference. We're talking about things that are probably not audible, or at the very least impactful.

[Dr.Ames]

Thanks for all your replies, in fact I am pleasantly surprised how smoothly DSD playback works with JRiver. With Foobar, which I still appreciate very much, I had repeatedly dropouts during playback, which could only be solved by quitting the software and deleting the temp files. So far, JRiver runs absolutely error-free here.

But some small things really drive me crazy here, how do I get JRiver, for example, which are in the same directory as the music files "cover.jpg" files to display as a cover without re-tagging. 

[JimH]

https://wiki.jriver.com/index.php/Cover_Art

[blgentry]

"Upsampling to DSD" is a destructive process.  This means that once it is converted to DSD, it can not be converted back into the exact same PCM signal again.  Some information is lost (discarded) in the process.  So from a pure information theory point of view there's no benefit to ever converting PCM to DSD, regardless of what device or software does it.

But if we step away from scientific analysis and go to human experience, there might be a reason.  Euphonics is a way of describing audio effects that sound pleasant.  An example of this is the harmonics that are added to music by a tube amplifier.  These harmonics are "distortion".  But they sound good because they emulate the action of of real musical instruments, which add harmonics to the base tone when you play the instrument.

Turntables have related euphonic properties that humans like the sound of.  Are they as accurate or more accurate than PCM files?  No.  Do they sound pleasant?  Yes.

DSD playback in general has some euphonic properties.  I credit these effects with the popularity of this unusual format.  It sounds good so people like it.  I've been told that conversion from PCM to DSD can also produce some eurphonics.  I have not experienced this myself, but some number of people seem to agree that it sometimes sounds good, depending upon the exact DAC and conversion method.

As always, like with food or wine, it's up to your individual tastes.  If you are a purist, then of course conversion to DSD is "wrong" and definitely unnecessary.  But if you are the type that wants to pour hot sauce on your eggs, or put butter in your pan sauce to punch it up... then maybe you should experiment with converting PCM to DSD.

Brian.

[Dr.Ames]

The main reason why I always used the upsamling in Foobar was an article that the processing power of a modern CPU can do a better conversion to DSD format than any DAC chip. However, I re-read the article and it probably only referred to 1 BIT DACS and my DAC contains two ESS Sabre ESS9038s, which are probably multibit converters and therefore do not convert PCM to DSD internally. Thus, the argument with the possible technical superiority of the CPU would then also be dropped.

[blgentry]

Almost all modern DACs are Delta-Sigma.  Unless it's prominently advertised as being multi-bit it almost certainly is not.  I'm not aware of any ESS DACs that are multi-bit.

To be clear, when I say "multi-bit" I mean a DAC that internally uses a 16 bit or greater DAC and does not convert to a Delta Sigma signal with a "1 bit" or "low bit" converter.

Anyway, it probably doesn't matter much at all.  Try running PCM to your DAC.  Try converting to DSD.  Compare the sound. You might not be able to tell a difference at all. Most of these effects are pretty subtle.  MC is a high fidelity player that uses the best practices in digital audio.  It's going to sound good in its default configuration and might sound a little different when converting to DSD.  Or it might sound exactly the same.

Brian.

[Dr.Ames]

Oh, I actually thought that the ESS chips contained multibit converters. In fact, I don't hear any significant difference when listening to PCM vs. upsampled DSD.

[dtc]

Most DAC chips use a technique called delta sigma modulation which basically oversamples the signal to a very high sample rate (Mhz) and reduces the number of bits. If it produces 1 bit it is basically a PCM to DSD conversion. Many chips produce 2-6 bit data (DSM) and they are known as multi-bit chips. The multi-bit format allows for better noise handling in the chip. ESS uses a proprietary delta sigma like process.

Some DACs (Berkeley I believe) maintain the data as PCM, often using more bits than the Redbook 16 bits. It does not convert to the high frequency of the delta sigma process.  This is the technique used by the original Pacific Microsonics DACs, which many still consider to some of the best sounding DACs ever made.

So, you have to be careful to specify multi-bit DSD (DSM) or multi-bit PCM.

Which is better really depends on the actual implementation and what the listener enjoys.

R2R Ladder DACs use a resistor network with alternating resistors of R and twice that (2R). It allows the direct processing of PCM data without any need to convert it to anything else.  They are not that common, but they have a loyal following. The design predates the widespread use of delta sigma processing, but was large stopped because it required very high precision for the resistors. Modern techniques have made it viable again.