Reducing DAC volume doesn't necessarily decrease bit-depth of audio data.

[Jimbob54]

Now I have my attenuators .

Using -20 dB before the preamp there is nothing worse in the sound . Maybe a little bit of an improvement with Yamaha wxc50 as digital volumeregulation.

With -20 dB attenuating I can also use my Yamaha as a streamer turning the digital preamp off , with only one rather loud volume . But its a rather interesting comparison.

Its clear that the sound this way is better , and that the digital volume control with the Yamaha wxc50 as preamp, is not very good. I can hear the sample rate conversion to 48 KHz .

Conclusion : I will buy a WiiM pro but maybe ” only ” the pro version, and use it as a digital preamp with volume regulation avoiding sample rate conversion.

Are you sure the sound that way isnt just louder?

 

[antcollinet]

Its clear that the sound this way is better , and that the digital volume control with the Yamaha wxc50 as preamp, is not very good. I can hear the sample rate conversion to 48 KHz .

Say what now?

Lets (for the time being) give you the benefit of the doubt, and say you are hearing something better in the sound waves, and not (like most would be in this circusntance) juse being hit by confirmation bias etc. What on earth makes you think that is coming from sample rate conversion, and not some other factor. Like (for example) a small (or large - as you have described it) volume change?

What evidence do you have as to the reason for what you are hearing?

The reason I ask - sample rate conversion (if done correctly and we've no reason to assume yamaha do it wrong) is *inaudible*

 

[Tangband]

Say what now?

Lets (for the time being) give you the benefit of the doubt, and say you are hearing something better in the sound waves, and not (like most would be in this circusntance) juse being hit by confirmation bias etc. What on earth makes you think that is coming from sample rate conversion, and not some other factor. Like (for example) a small (or large - as you have described it) volume change?

What evidence do you have as to the reason for what you are hearing?

The reason I ask - sample rate conversion (if done correctly and we've no reason to assume yamaha do it wrong) is *inaudible*

Ok this is a valid doubt and it might be something else than the sample rate conversion but the fact is that wxc50 in ”player mode” , - the spdif output follows the sampling frequency input and this is not the case with wxc50 in ”preamp mode ” where every material is SRC to 48 kHz on the spdif out.

Its easy to draw the conclusion that the 48 KHz SRC in the Yamaha does something to the signal . The chip Yamaha uses has a SRC sinad of 108 dB, where other chip often have 130 dB or better.

My Mac with DDC does SRC entirely inaudible and better than the Yamaha in my opinion.

 

[antcollinet]

Ok this is a valid doubt and it might be something else than the sample rate conversion but the fact is that wxc50 in ”player mode” , - the spdif output follows the sampling frequency input and this is not the case with wxc50 in ”preamp mode ” where every material is SRC to 48 kHz on the spdif out.

Its easy to draw the conclusion that the 48 KHz SRC in the Yamaha does something to the signal . The chip Yamaha uses has a SRC sinad of 108 dB, where other chip often have 130 dB or better.

My Mac with DDC does SRC entirely inaudible and better than the Yamaha in my opinion.

But also - the volume is different. And even small volume changes that you don't even perceive as a volume change, change the perception of music quality. It is why comparing sound is meaningless until you level match at the speaker with a voltmeter.

So what is the most likely reason for your perception of the change. A volume difference (even a very small one) where we know that different quality perception is a very likely result? Or a sample rate change - where almost universally that is inaudible? We can include in that any Sinad difference - which again once you get beyond 80 db (probably much lower) differences are not audible in most real world listening conditions.

 

[neutronsix]

I'll try 0,5 : 64 = 0,0078125 dB
then I remember that dB are logarithmic values, then I reflect that we are talking about sound pressure and the values are quadratic... help. But how infinitesimal is the value we obtain. Do AKM chips also have the same type of digital volume?

 

[neutronsix]

Always from the link shared by @antcollinet. I honestly don't know the reliability of the source. Why should Sabre chips reduce spikes? If I wish I perform a post-equalization of the signal at the output of the DAC or preamplifier interface. I am aware that each electronic component bears its signature but the thought that a DAC is not transparent precisely because of the main component...

 

[mdsimon2]

Always from the link shared by @antcollinet. I honestly don't know the reliability of the source. Why should Sabre chips reduce spikes? If I wish I perform a post-equalization of the signal at the output of the DAC or preamplifier interface. I am aware that each electronic component bears its signature but the thought that a DAC is not transparent precisely because of the main component...

I don't think there is anything particularly controversial in that statement. However, the level of distortion being discussed is likely insignificant for most DACs.

Many DACs have their highest distortion near 0 dBFS. If you are not attenuating in the DAC but rather using a downstream analog preamp, this distortion will always be present in the DAC output. Now, if you have a DAC that does -120 dB THD+N at 0 dBFS does it really matter? I say no.

The other potential issue is intersample over clipping. I am not sure if this is true for all ESS DACs, but on the two that I own (MOTU Ultralite Mk5 and Okto dac8 pro) using the DAC volume control will eliminate intersample over clipping. Again, this is an issue that occurs at high signal levels and can only be eliminated by digital attenuation. This might be slightly more audible but I am not fully convinced. See here -> https://www.audiosciencereview.com/...nts-and-rising-noise-floor.42383/post-1503597 for some more info on this issue.

Michael

 

[NTK]

I believe there is too much emphasis on how DAC's handle inter-sample overs, and IMHO bordering onto the territory of snake oil.

If you have inter-sample overs, that means your signal is clipped (technically you can only get inter-sample overs if your original analog signal is > full scale). So we are really talking about how 'best' to reproduce (or guestimate to reproduce) clipped signals, which are already irreversibly impaired at the source.

 

[Peterinvan]

My dongle DACs behave differently:

iPhone > FiiO KA3. This DAC depends on the source to adjust the volume.
iPad Pro > Cayin RU6. The iPad only need a tiny amount of volume (just above mute). Albeit the RU6 is an R-2R DAC with Resistor Array Volume Control.

My question: is there any benefit to raising the volume on the iPad Pro to 100 when using the RU6?

 

[signalpath]

the resolution of the output of any DAC is limited by physics to 22 or 23 bits, no matter how many bits it is using internally.

In a single-path design, yes.
In a multi-path design, the number of conversion bits is, effectively, unlimited.

 

[antcollinet]

In a single-path design, yes.
In a multi-path design, the number of conversion bits is, effectively, unlimited.

That is just another form of internal bits.

Once you output those billion bits into an actual analogue circuit, then all but the top 23 are down in the noise.

 

[AnalogSteph]

Once you output those billion bits into an actual analogue circuit, then all but the top 23 are down in the noise.

Well, it's kind of a matter of analog supplies. The maximum you can realistically hope for on +/-15 V is about 25 bits, though that circuit isn't doing an awful lot then. If you were to pull an SPL and go for +/-60 V, you could hope for 2 more bits.

 

[Sokel]

The strangest thing I often see with my Khadas is that the whole distortion pattern changes big time both with changing input levels and changing levels by the DAC alone.
The ancient EMU's AKM chips for example don't do that,the pattern (higher distortion but much cleaner spectrum after H4) doesn't do that,H2 and H3 goes down proportionally with the volume until after -15-18db are at absolutely insignificant levels (less that -120db).

There's a lot at play for a finished product obviously.
(I can show you with measurements what 0.10db level changing does to the whole pattern if you like )

 

[signalpath]

Once you output those billion bits into an actual analogue circuit, then all but the top 23 are down in the noise.

Multi-path topology can be realized in analog or digital circuits.

Remember, the output of a DAC is analog. And a multi-path DAC can achieve true conversion of any number of bits. Right now we're settled at 27-bits, but there's no reason we couldn't do all 32-bits. It's just not practical.

Every link in the audio signal path can be realized in multi-path (mics, ADC, DAW, DAC, power amp), which could maintain 27-bit-equivalent performance from capture to delivery. Keep in mind, almost all of this additional dynamic range is achieved on the low side, the noise side. I'm convinced this is future of all audio. But it's very hard to do, and will require tight integrations.

 

[antcollinet]

Multi-path topology can be realized in analog or digital circuits.

Remember, the output of a DAC is analog. And a multi-path DAC can achieve true conversion of any number of bits. Right now we're settled at 27-bits, but there's no reason we couldn't do all 32-bits. It's just not practical.

Every link in the audio signal path can be realized in multi-path (mics, ADC, DAW, DAC, power amp), which could maintain 27-bit-equivalent performance from capture to delivery. Keep in mind, almost all of this additional dynamic range is achieved on the low side, the noise side. I'm convinced this is future of all audio. But it's very hard to do, and will require tight integrations.

It\s utterly pointless for Audio, since the noise floor at currently achievable 22-23 bits (or so) is way below audibility. What we have now is good enough for ever.

I'd like to say that means it will never be the future of audio - but there seems to be no limits to what audiophoolery will pay for - no matter how useless.

 

[signalpath]

It\s utterly pointless for Audio, since the noise floor at currently achievable 22-23 bits (or so) is way below audibility. What we have now is good enough for ever.

I'd like to say that means it will never be the future of audio - but there seems to be no limits to what audiophoolery will pay for - no matter how useless.

Let's do a refresher in "audibility"

At the low side, we can hear -8dB SPL at 4kHz. Room noise typically masks this, but the recent popularity of headphones -- which can attenuate room noise beyond 30dB -- brings -8dB SPL well into "audibility."

On the high side, snare drums, trumpet blasts (etc) achieve +155dB SPL. We (recording engineers) stick microphones on these things every day. Boom bass cars routinely achieve +155dB SPL (the world record is something like +184dB). These are real world numbers. -8dB at the low side and +155 at the high side is 162dB of real world dynamic range, or 27-bit equivalent.

Will everyone need 162dB? Of course not. But this is my world. Moreover, in the realm of mixing consoles, with upwards of 100 channels (or often far more), the sum of noise can be 20-40dB higher than the individual channels. Mixing console micamps, ADCs, and DACs will benefit greatly from multi-path topology.

Ultimately, if a recording engineer is given the choice of a 20-bit process vs. a 27-bit process, at similar cost, the 20-bit process goes obsolete. QED.

 

[antcollinet]

Let's do a refresher in "audibility"

At the low side, we can hear -8dB SPL at 4kHz. Room noise typically masks this, but the recent popularity of headphones -- which can attenuate room noise beyond 30dB -- brings -8dB SPL well into "audibility."

On the high side, snare drums, trumpet blasts (etc) achieve +155dB SPL. We (recording engineers) stick microphones on these things every day. Boom bass cars routinely achieve +155dB SPL (the world record is something like +184dB). These are real world numbers. -8dB at the low side and +155 at the high side is 162dB of real world dynamic range, or 27-bit equivalent.

Will everyone need 162dB? Of course not. But this is my world. Moreover, in the realm of mixing consoles, with upwards of 100 channels (or often far more), the sum of noise can be 20-40dB higher than the individual channels. Mixing console micamps, ADCs, and DACs will benefit greatly from multi-path topology.

Ultimately, if a recording engineer is given the choice of a 20-bit process vs. a 27-bit process, at similar cost, the 20-bit process goes obsolete. QED.

You are talking about the range of sounds that can exist in air. And I'd be prepared to bet that -8dB is not achieveable by many at all under any circumstances, and none at all under real world listening conditions.

Then if people are listening at levels where the snares are hitting 155dB, they are going to be doing the same to their ears that a 1982 Motorhead concert did to one of mine. There is a reason why drummers (should) use ear defenders.

For people listening to reproduced music, 22 bits is already significantly more than needed.