Print

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

[tyler69]

Hi all,

as far as I understand it is not possible to use convolution when bitreaming audio. Would it be possible to implement this functionality in any way? I have read about one player (HQPlayer), which seems to offer this.

[adamt]

As I understand it, bitstreaming is sending the data exactly how it is with no modifications.  It isn't run through the JRiver audio engine and convolution can't be applied. 

[tyler69]

Thanks for replying. Yes, that's my understanding too. Unfortunately the developer of HQPlayer doesn't share his approach to this..

[Hendrik]

It is theoretically possible to do processing directly on the DSD data, however we do not support this, and I'm not aware of any plans in this regard either.
All of the MC DSP engine is designed around PCM, and duplicating everything for DSD would be quite a lot of extra stuff to look after.

[tyler69]

Thanks for the clarification.

Maybe this will be considered in JRiver "if" DSD will become more widespread as a music format. I admit that only a small fraction of my music is DSD, but it's unconvenient nonetheless, being able to use convolution on "only" ~90% of my music (this is for bitstreaming DSD).

[Hendrik]

You could just play DSD as high-samplerate PCM and enjoy convolution all the same.

[tyler69]

Yes, I know. That's why I put the brackets at the end of my post  
I recently got a (native) DSD capable DAC, and I really enjoy it..Thanks for the suggestion though. In fact: my common sense tells me, that the gain in audio quality (assumed the music sounds better in DSD than in PCM) when using convolution with DSD played as HiRes PCM ist higher than when listening to bitstreamed DSD without convolution.. it's just knowing that it seems to be possible to do both things what bothers me + the fact, that my DAC supports native DSD

EDIT: Related threads http://yabb.jriver.com/interact/index.php?topic=81541.0; http://yabb.jriver.com/interact/index.php?topic=81541.0; http://yabb.jriver.com/interact/index.php?topic=84785.0

[InflatableMouse]

Isn't it possible to set output format as DSD when using convolution?

[glynor]

Isn't it possible to set output format as DSD when using convolution?

Yes.  In the Output Format DSP.

Which is, I'm reasonably sure, what his alternative application is doing.  That's not bitstreaming if it can modify the signal.  It cannot be, as the definition of bitstreaming means that the sending application cannot mess with the signal.  At least, as it is commonly understood and implemented (vocabulary is fiddly).

It is possible that someone out there has built a native DSD audio processing DSP engine, but without seeing it, I'd bet you $20 that it is instead doing the exact same thing as enabling DSD Output in MC's Output Settings DSP.

[6233638]

It is theoretically possible to do processing directly on the DSD data, however we do not support this, and I'm not aware of any plans in this regard either.
All of the MC DSP engine is designed around PCM, and duplicating everything for DSD would be quite a lot of extra stuff to look after.

Even if it were possible (I didn't think it was possible to edit 1-bit audio) and even if all the processing happens on a native DSD stream without converting to PCM, it would still not be bitstreaming.
 
 
DSD Encoding will take a DSD input, process in PCM, and then encode the output back to DSD.
I don't see anything on the HQPlayer website to indicate that it's doing anything different from this.
 
Considering that the professional tools don't offer native 1-bit audio processing, it seems unlikely.

[glynor]

even if all the processing happens on a native DSD stream without converting to PCM, it would still not be bitstreaming

Exactly.

I don't see anything on the HQPlayer website to indicate that it's doing anything different from this.
 
Considering that the professional tools don't offer native 1-bit audio processing, it seems unlikely.

Exactly.  And the fact that it is "vaguely explained" makes me think he doesn't want his customers to know what is really happening.  If he had, indeed, built some kind of magical DSD native processing engine, you'd think he'd be shouting it from the rooftops, and it would be discussed all over the place in Pro Audio circles.

[6233638]

And the fact that it is "vaguely explained" makes me think he doesn't want his customers to know what is really happening.  If he had, indeed, built some kind of magical DSD native processing engine, you'd think he'd be shouting it from the rooftops, and it would be discussed all over the place in Pro Audio circles.

Well it would be theoretically possible to process it in DSD-wide, which is usually 8-bit 2.8MHz, and then convert that output to 1-bit DSD, so all your processing is happening natively as DSD audio without downsampling - but that's still not bitstreaming, and I agree that they would be making it very clear on their website if they were doing this, since basically nothing else does.

[tyler69]

Thank you guys for the input. I was not aware of the fact, that JRiver's DSP can output DSD. So the process steps would be DSD-to-PCM -> DSP (convolution) -> PCM-to-DSD? If so, does JRiver automatically detect the input source and converts sample rates (to what?)? Does JRiver then automatically output DSD?

[6233638]

Thank you guys for the input. I was not aware of the fact, that JRiver's DSP can output DSD. So the process steps would be DSD-to-PCM -> DSP (convolution) -> PCM-to-DSD? If so, does JRiver automatically detect the input source and converts sample rates (to what?)? Does JRiver then automatically output DSD?

It's processed as 64-bit floating-point PCM at 1/8th the DSD sample rate.

[tyler69]

OK, thanks.
So let's assume the following scenario:
A playlist consists of PCM and DSD files aswell, both in various sample rates/resolutions.

Is it possible to use convolution with the files being output in their native sample rate/resolution? I can't seem to find the right settings in JRiver..

[Matt]

OK, thanks.
So let's assume the following scenario:
A playlist consists of PCM and DSD files aswell, both in various sample rates.

Is it possible to use convolution with the files being output in their native resolution? I can't seem to find the right settings in JRiver..

DSD will be output as high sample rate PCM.  The transformation is effectively lossless.

[tyler69]

Thanks Matt. InflatableMouse mentioned that DSD can be set as an output format (after convolution), hence my question.

[glynor]

Well it would be theoretically possible to process it in DSD-wide, which is usually 8-bit 2.8MHz, and then convert that output to 1-bit DSD, so all your processing is happening natively as DSD audio without downsampling - but that's still not bitstreaming, and I agree that they would be making it very clear on their website if they were doing this, since basically nothing else does.

Yeah, I didn't mean it was impossible, just that it would be something they'd want to trumpet loudly if they'd done the required and substantial engineering work required.