Print

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

[rotho]

Hi,

Thanks to last build (see thread http://yabb.jriver.com/interact/index.php?topic=92637.0), I am now able to play DSD256 files on my DSD64-only DAC (I even tried DSD512 : it works  ).
To do that, I must have the following settings :
- Bitstreaming : None
- Output encoding : None
- Output format : Input > 384,000 Hz --> Output = 176,400 Hz

The drawback of this method is that my DSD64 files are converted to PCM 

Now, if I set Bitstreaming to Yes (DSD), that means (in theory) that all DSD files are treated "as is" : DSD64 --> DSD64, DSD128 --> DSD128, DSD256 --> DSD256, and so on...
So it can't work for a DSD64-only DAC like mine. (In practice, I noticed that DSD256 and DSD512 files are rather converted to PCM, according to the Output format setting, even with the Bitstreaming setting to DSD, but that's not the case for DSD128 files : that seems like an inconsistency bug  )

On the other hand, if I set Output encoding to 1xDSD, then all files are converted to DSD64, but the process of conversion for DSD input files requires a step in the PCM domain, and I found that the DSD256 --> PCM --> DSD64 path makes a lot of parasite noises, so it is not a good solution either...

So, here is my suggestion  :

--> Extend the sample rate values of the Output Format settings (to reach DSD64, 128, 256, 512, ... frequencies), and make these settings the priority over the Bitstreaming setting.

For example, in the case of my DSD64-only DAC, I would have the following settings :
- Bitstreaming : Yes (DSD)
- Output encoding : None
- Output format :
Input <= 192,000 Hz --> Output : No change
Input = 352,800 Hz --> Output : 176,400 Hz
Input = 384,000 Hz --> Output : 192,000 Hz
Input = 2 822 400 Hz --> Output : No change [DSD64 would be taken into account by the Bitstreaming setting]
Input > 2 822 400 Hz --> Output : 176,400 Hz [DSD128 and above would be converted to PCM]

For DSD128-only DACs, the output format settings would be :
...
Input = 384,000 Hz --> Output : 192,000 Hz
Input = 2 822 400 Hz --> Output : No change [DSD64 would be taken into account by the Bitstreaming setting]
Input = 5 644 800 Hz --> Output : No change [DSD128 would be taken into account by the Bitstreaming setting]
Input > 5 644 800 Hz --> Output : 176,400 Hz [DSD256 and above would be converted to PCM]

And so on...

What do you think about this suggestion ?

[InflatableMouse]

If I'm reading correctly you would only need to be able to specify which DSD files you want to bitstream.

For DAC's that can only do 1x or 2x for instance, you might want to bitstream those and resample the others. Bitstreaming takes precedence over output format resampling rules.

[rotho]

If I'm reading correctly you would only need to be able to specify which DSD files you want to bitstream.

For DAC's that can only do 1x or 2x for instance, you might want to bitstream those and resample the others. Bitstreaming takes precedence over output format resampling rules.

Yes, that would be another solution : in the Bitstreaming : Yes (DSD) setting, to be able to specify the concerned DSD rates (64, 128, ...) and then the other rates would be treated in accordance with the Output format settings...

[tyler69]

Earlier I posted an idea of mine which implied your suggestions: http://yabb.jriver.com/interact/index.php?topic=90247.msg619780#msg619780

[6233638]

It's not impossible if you use Zone Switch, but it is messy.
I agree that it could be better.
 
One solution might be to add DSD rates to the Output Format DSP, but after some consideration, I'm not sure that is the best place for it any more.
I do think that the 352.8kHz setting should indicate that it also applies to 1xDSD playback, as the program never tells the user this, and DSD used to be covered by "Greater than 192,000Hz".
And while we're at it, I'd like to see the height of the Output Format DSP increased to avoid scrolling in that selection.
 
What would be best is if enabling DSD bitstreaming would auto-detect the rates which your DAC supports, similar to the way that bit-depth is now automatically configured.
Expanded options in the custom bitstreaming configuration window would also be useful:
 


A lot of people with DSD DACs seem to prefer bitstreaming content at the highest supported rate, and upsampling everything else to that rate—or at least everything below 2xDSD.

[rotho]

A lot of people with DSD DACs seem to prefer bitstreaming content at the highest supported rate, and upsampling everything else to that rate—or at least everything below 2xDSD.

I think treating the problem via the Bitstreaming settings would indeed be the simplest way to do it (from a user point of view, at least).

But when you say "bitstreaming content at the highest supported rate, and upsampling everything else to that rate", if you mean a DSD rate, I would rather avoid that, because for example the DSD128 --> DSD64 path implies an intermediary step in PCM (and the DSD256 --> DSD64 and DSD512 --> DSD64 conversions seem to be completely messy, at least in my setup).

So, in my situation, I'd rather have DSD64 --> DSD64, and DSD128 and above --> PCM 176,4
Owners of DSD128 DACs would for their part choose DSD64 --> DSD64, DSD128 --> DSD128, and DSD256 and above --> PCM 352,8
But on the other hand I would avoid doing DSD64 --> DSD128, because it goes through a PCM step...

Anyway, I vote for your suggestion in the Bistreaming Formats window (with the possibility of choosing one or several DSD rates, which would be treated "as is", without conversion).

[6233638]

Perhaps my image wasn't showing up, that's what I was suggesting.

[6233638]

Thinking about it some more, I realize that this does not account for people who want to only upsample some formats to DSD while leaving other PCM rates alone. E.g. 44.1kHz PCM.
 
Combining the DSD Bitstreaming, DSD Encoding, and Output Format section into a single control may be the easiest way to do this from a user's point of view.
 
 
The proposed options do not solve the issues of bitstreaming multichannel DSD either.
 
I wonder if that should just have its own separate option in the bitstreaming configuration, since so few devices can bitstream anything other than stereo DSD.
 
 
But frankly, I find mixing DSD and PCM to get rather messy, and I personally prefer to play back everything as PCM anyway, since it is required for Volume Leveling to work anyway.

[tyler69]

... , and I personally prefer to play back everything as PCM anyway, since it is required for Volume Leveling to work anyway.

Not if you use an exaSound DAC  

[6233638]

Not if you use an exaSound DAC  

Well it isn't native DSD if the DAC is able to adjust the volume anyway.
 
Here's a mockup of what it might be like if the options were consolidated in DSP Studio:

[tyler69]

Looks very nice!

[mojave]

Well it isn't native DSD if the DAC is able to adjust the volume anyway.

I tend to agree. Below is what George Klissarov of Exasound has to say. It seems to be a difference in semantics, though, since the attenuation is still taking in the DAC chip before analog output.
ES9018 Product Brief

I can switch my Lynx Aurora 16 DAC's output from +4dBu to -10dBv. This is an 11.78 dB difference which is done digitally, but only changes the output voltage so it seems to work the same way. If I have it set to +4dBu and then use the Lynx mixer to attenuate by 11.78 dB, is the signal identical to when the output is set to -10dBv? Attenuating in the mixer causes dither to occur when going from the 48 bit internal processing (I think) to 24 bit for output. What is the relationship between dither and changing from +4dBu to -10dBv?

1) Let me explain about the volume control – it is not analogue nor digital… it is right in-between the two worlds. The signal over the USB  from the computer to the DAC travels always at the 0 dB level. So there is no bit-loss like in digital volume controls.  When you request a volume level, we send a command to the DAC – a couple of bytes to instruct the DAC chip ES9018 to produce lower voltage. So the 0dB sound stream is converted to lower voltage analogue signal. 

From the analogue side of the equation,  the signal is not attenuated, it is just born smaller. So you don’t have the signal to noise difficulties of conventional analogue volume controls.

[6233638]

I believe the ESS DACs are PCM internally anyway, operating at DSD sample rates.
 
Unfortunately they don't accept that as an input, so you have to choose between low sample-rate high bit-depth PCM, or low bit-depth high sample-rate DSD, as an input.
 
With my DAC, I prefer to send it PCM.

[rotho]

I believe the ESS DACs are PCM internally anyway, operating at DSD sample rates.
 
Unfortunately they don't accept that as an input, so you have to choose between low sample-rate high bit-depth PCM, or low bit-depth high sample-rate DSD, as an input.
 
With my DAC, I prefer to send it PCM.

It is my understanding that the ESS DACs use a kind of "DSD-wide" format, which enables to avoid a PCM step for the volume setting. I am also quite positive that this is the case for high-end DACs like the PS Audio Direct Stream (based on a FPGA). And there is a certain consensus that the conversion from DSD can be more "Analog like" than from PCM...

That's why I would like to leave my DSD64 files "as-is", and convert my higher resolution files to PCM, and the current JRiver settings do not allow that...