Jump to content
IGNORED

Would you expect DSD to be quieter than PCM?


Norton

Recommended Posts

Having used various DSD sources and media with various DACs over the years, my impression is that DAC analogue output when fed by DSD is lower than with PCM in certain cases.  I'm not talking about the somewhat "softer" sound of DSD, but marked differences in volume.

 

This  phenomenon varies according to DAC and media but is consistent within them.  Thus there is no apparent difference in output between SACD and CD with either of my SACD players.  On the other hand, using my Resonessence DAC via its built in SD card transport with SACD rips, output seems markedly lower than with RBCD rips or PCM downloads and has me reaching for my preamp's volume control. 

 

Is this to be expected and if so, why?

Link to comment
2 hours ago, Norton said:

Having used various DSD sources and media with various DACs over the years, my impression is that DAC analogue output when fed by DSD is lower than with PCM in certain cases.  I'm not talking about the somewhat "softer" sound of DSD, but marked differences in volume.

 

This  phenomenon varies according to DAC and media but is consistent within them.  Thus there is no apparent difference in output between SACD and CD with either of my SACD players.  On the other hand, using my Resonessence DAC via its built in SD card transport with SACD rips, output seems markedly lower than with RBCD rips or PCM downloads and has me reaching for my preamp's volume control. 

 

Is this to be expected and if so, why?

 

I have noticed that too sometimes...i usually notice the difference in volume when i switch drivers...not sure why.

Link to comment

There are multiple factors at play here. Firstly, the PCM and DSD versions of the same track might inherently have different levels. This explains why the effect depends on the content. Secondly, the conversion to analogue in the DAC might not be level matched.

 

Here are a couple of test files, one PCM and one DSD. Both have a 1 kHz sine tone at -3 dBFS. In theory, they should play at the same level. If they don't, it's not necessarily a bad thing. It just means your DAC hasn't been tuned for this.

 

For example, on the iFi Nano, the output is 1.5 dB lower with DSD than with PCM.

dsd-test.zip

Link to comment
3 minutes ago, kumakuma said:

DSD is generally mastered 6 db lower than PCM.

That's a bit of a misconception. DSD defines the 0 dB level such that if a full-scale PCM signal is converted to DSD and low-pass filtered back to PCM, the resulting signal will be 6 dB lower. The reason for this is that the sigma-delta modulator isn't stable for large signals.

 

This does not mean that DSD is mastered lower. All it means is that the analogue voltage to digital value mapping uses a different scale. In fact, PCM uses many different scales. For instance, with 16-bit integer PCM the 0 dB level corresponds to a numerical value of 32767 while floating-point PCM (typically) encodes the same level as 1.0.

Link to comment
5 minutes ago, Samuel T Cogley said:

 

I find this to be generally true (EDIT: @mansr points out the reasons for this perception above), but I always disable the "+6dB boost for DSD" option.  Depending on the input material, +6dB may be excessive.  This is what volume knobs are for.  :)

 

I do the same thing. 

Sometimes it's like someone took a knife, baby
Edgy and dull and cut a six inch valley
Through the middle of my skull

Link to comment
8 hours ago, Norton said:

On the other hand, using my Resonessence DAC via its built in SD card transport with SACD rips, output seems markedly lower than with RBCD rips or PCM downloads and has me reaching for my preamp's volume control. 

 

It's been the same for other DACs I've owned as well.  Nothing out of the normal with the Resonessence.

 

I've been upconverting my PCM with HQPlayer and find it a great way to level out the playing field a little.  Headroom applied to PCM and send DSD direct.

Roon Rock->Auralic Aria G2->Schiit Yggdrasil A2->McIntosh C47->McIntosh MC301 Monos->Wilson Audio Sabrinas

Link to comment
36 minutes ago, Dr Tone said:

 

It's been the same for other DACs I've owned as well.  Nothing out of the normal with the Resonessence.

 

I've been upconverting my PCM with HQPlayer and find it a great way to level out the playing field a little.  Headroom applied to PCM and send DSD direct.

 

Dr T  you keep the Mirus at -2db with DSD too?

 

Reminds me I must try out the Mirus with HQP .

 

 

Link to comment
1 hour ago, Norton said:

 

Dr T  you keep the Mirus at -2db with DSD too?

 

Reminds me I must try out the Mirus with HQP .

 

 

 

-2db at all times it would be a pain to go up and down.  Nothing wrong with having to turn the preamp volume up and get it a little closer to unity gain.  I think that's the term.

 

The Linear Apodizing is real good.  It's the extra special filters in HQPlayer or I just stick with the aforementioned.  I go back and forth.  Right now I'm rocking closed-form PCM->PCM upsampling.

Roon Rock->Auralic Aria G2->Schiit Yggdrasil A2->McIntosh C47->McIntosh MC301 Monos->Wilson Audio Sabrinas

Link to comment
12 hours ago, Norton said:

Having used various DSD sources and media with various DACs over the years, my impression is that DAC analogue output when fed by DSD is lower than with PCM in certain cases.  I'm not talking about the somewhat "softer" sound of DSD, but marked differences in volume.

 

6 dB difference is completely normal. This is becauase DSD 0 dB level is defined to be 50% of the theoretical maximum. And the specification allow temporary peaks up to max +3.15 dB DSD level. While PCM has hard ceiling at 0 dBFS.

 

Some DAC chips like ESS Sabre attempt to normalize the output volume by applying 6 dB gain on DSD. Then PCM and DSD have same 0 dB reference level. But that has a risk when some material actually exceeds 0 dB DSD level and goes to +3 dB or so. Then it may hit maximum voltage span of the analog section (for example if it's running from 5V supply)...

 

TI chips have varying DSD output level depending on selected DSD analog filter. But typically have about -3 dB level compared to PCM 0 dBFS level. AKM chips Direct DSD output level is -3.5 dB compared to 0 dBFS PCM so they have enough headroom to accommodate the allowed DSD peaks.

 

Because of uncertainty of behavior above 0 dB DSD level, for example HQPlayer is trying to enforce output to stay within 0 dB and don't allow output to exceed the +3.15 dB level.

 

Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Link to comment
52 minutes ago, Miska said:

Some DAC chips like ESS Sabre attempt to normalize the output volume by applying 6 dB gain on DSD

 

That must be configurable?  @Norton was referencing his current Resonessence as an example.

Roon Rock->Auralic Aria G2->Schiit Yggdrasil A2->McIntosh C47->McIntosh MC301 Monos->Wilson Audio Sabrinas

Link to comment
2 minutes ago, Miska said:

 

AFAIK it is not. I cannot see any mention of it in the datasheets either...

 

 

Well Resonessence is doing some mojo then, the DSD playback volume is definitely at least 6db lower.

Roon Rock->Auralic Aria G2->Schiit Yggdrasil A2->McIntosh C47->McIntosh MC301 Monos->Wilson Audio Sabrinas

Link to comment
11 minutes ago, Dr Tone said:

Well Resonessence is doing some mojo then, the DSD playback volume is definitely at least 6db lower.

 

It is of course possible that they are just setting the Sabre volume control to 6 dB lower value to deal with the things I stated earlier...

 

(or they know some secret undocumented feature of the Sabre chip... wouldn't be first one)

Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Link to comment

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now



×
×
  • Create New...