Digital Audio and Amplifier Noise Floor Comparison - Is 16bit/44.1kHz All We Need ???

[Paul R]

10 minutes ago, Rt66indierock said:

 

In my defense, I'm going to keep using my Klipsch Heresy's at home so I need something quiet. John Siau of Benchmark recently wrote and made some graphs about how quiet your stuff has to be to hear certain bit depths. 

 

Those are really efficient aren’t they? The IIIs are like 99db. Generally the more efficient the speaker, the quieter the electronics of course. 

 

P.S. What defense? I don’t think many reasonable people would say those Heresy speakers are not well worth their keep.  If they do, ignore ‘em. 

[Rt66indierock]

Just now, Paul R said:

 

I believe both the Benchmark and maybe the Outlaw monoblocks meet that criteria.

 

We know the Benchmark does, the Outlaw amps don't but they are great for vinyl. Reg Greene of TAS says the Bryston amps meet the criteria. I'm looking at a Canadian design that meets the criteria.

[Rt66indierock]

4 minutes ago, Paul R said:

 

Those are really efficient aren’t they? The IIIs are like 99db. Generally the more efficient the speaker, the quieter the electronics of course. 

 

My Heresy II's are spec'd at 96 dB. 

[Ishmael Slapowitz]

Has anybody personally heard both the Schiit Vidar and Benchmark,? If so, thoughts?

[esldude]

42 minutes ago, Paul R said:

 

Well said. Did you get the typical “bathtub” output? 

 

I just find find it hard to argue that 16bit, even with noise shaping is preferable to 24 bit. Equipment specs are only going to keep getting better and better. 

I don't know if anyone is arguing that 16 bit is preferable to 24 bit.  It saves space, if the difference is near nothing maybe that is important to streaming.  

 

Now go find some recordings that need all of 16 bit and get back to me. 

 

I don't remember if it was here or another thread in the last few days where someone posted a mastering guy showing how 16 bit vs 8 bit had all the music there, and the only difference is noise. 

 

So maybe a K-horn owner needs to listen to dithered silence and let us know which one's are inaudible.  Or we could amplify digitally so that our inefficient speakers are putting out the same level of silence as a K-horn to see which you can hear.  20 db gain should do it for most everyone's speakers (except for mine). 

[esldude]

2 minutes ago, Ishmael Slapowitz said:

Has anybody personally heard both the Schiit Vidar and Benchmark,? If so, thoughts?

My thoughts......why would I buy Schiit when I could have a Benchmark.  Haven't heard either. 

 

Have heard Bryston, those are excellent amps too.  Cost more however. 

[Ishmael Slapowitz]

Just now, esldude said:

My thoughts......why would I buy Schiit when I could have a Benchmark.  Haven't heard either. 

 

Have heard Bryston, those are excellent amps too.  Cost more however. 

Not really helpful...but er, thanks...

 

The Schiit is $2300 cheap. That is why.

[Ralf11]

there are some older threads where the Vidar is mentioned - IIRC, one guy had hum problems

 

 

re 16 bit...  if there is little equipment offering better noise performance, few recordings made > 16 bits that you like, or if very low noise is not a significant factor in improving SQ for you, then...

 

the above are why I got a NuPrime amp instead of the Benchmark; ultimately it may migrate to a 2nd system with a Sonic Frontiers pre-amp, bookshelf speakers, where it would replace a nad amp...

[sandyk]

7 hours ago, Shadders said:

Hi,

The specification is A-weighted, so please correct if wrong, removes noise due to the filter. The ASR review probably does not use A-weighting.

Regards,

Shadders.

Hi Richard

The Benchmark is not alone in giving decent Unweighted and Weighted noise figures these days , with even many kit amplifiers performing very well at a cost, typically including metalwork , transformers etc. of <Au $1,000 

 Even the recent less ambitious Silicon Chip magazine designed 200W SC200 Amplifier specifies Signal-to-Noise Ratio: -116dB unweighted with respect to 135W into 8Ω(20Hz-20kHz)

I have attached the specs for the previous more ambitious SC ULD3 amplifier (123dB Unweighted)

from March 2012, as well as their DAC from way back in Sept.2009 which is capable of further improvement using more recent voltage regulators etc.

 Undoubtedly there are numerous commercially available amplifiers and DACs that easily surpass the performance of these older designs these days.

 

Regards

Alex

[Shadders]

2 minutes ago, sandyk said:

Hi Richard

The Benchmark is not alone in giving decent Unweighted and Weighted noise figures these days , with even many kit amplifiers performing very well at a cost, typically including metalwork , transformers etc. of <Au $1,000 

 Even the recent less ambitious Silicon Chip magazine designed 200W SC200 Amplifier specifies Signal-to-Noise Ratio: -116dB unweighted with respect to 135W into 8Ω(20Hz-20kHz)

I have attached the specs for the previous more ambitious SC ULD3 amplifier from March 2012, as well as their DAC from way back in Sept.2009 which is capable of further improvement using more recent voltage regulators etc.

 Undoubtedly there are numerous commercially available amplifiers and DACs that easily surpass the performance of these older designs these days.

 

Regards

Alex

Hi sandyk,

Yes, there are many excellent DIY designs, which can be built for the fraction of the price. It is a shame that DIY is not more popular.

Regards,

Shadders.

[esldude]

1 hour ago, Ishmael Slapowitz said:

Not really helpful...but er, thanks...

 

The Schiit is $2300 cheap. That is why.

As someone else posted.  Vidar has been a bit iffy with hum issues and shutting down with some loads.  So with that and Schiit's history of problems, I'd go with something else.  I could have suggested some others, but I restricted it to your choices.  I thought it odd you asked to choose between 2 amps differing so much in price.  But hey??

 

 

[psjug]

Delete duplicate

[psjug]

4 hours ago, esldude said:

A-wtd is usually a 4-6 db improvement over no weighting.  So it isn't like unweighted results are suddenly terrible.  More common is fudging on real values or some trickery in gain staging that gives a number you won't see in normal use.  As Rt66indierock said, your noise floor is much closer to the signal at 1 watt levels.  Of course with common speakers you only need mid-80 db SNR to be so low it will never be heard at 1 watt.  Plus your SNR with dithered 16 bit is worse by a half bit to bit.  So rather pointless to engage in petty arguments without putting all this in context. 

 

I've found it interesting no one ever points out that with shaped dither and 16 bit audio there is a real benefit to high sample rates.  Here is where I took silence, and saved it in 16 bit with shaped dither in both 44/16 and 192/16.  I then filtered out everything above 20 khz in the 192 file.  Notice the shaped 44 has a noise level of -84 db while the filtered 192 is -120 db.  This is what would happen if your speakers (and your ears) don't respond above 20 khz.  Of course no one does 192 at 16 bit do they.  Getting back on topic, something like this could have been done with MQA without requiring licensing, messing with undecoded fidelity or causing other problems.  

 

 

This is interesting, but if the goal is increasing SNR then it is a pretty inefficient way of doing it, compared to just going with greater bit depth.

[Ishmael Slapowitz]

2 hours ago, esldude said:

As someone else posted.  Vidar has been a bit iffy with hum issues and shutting down with some loads.  So with that and Schiit's history of problems, I'd go with something else.  I could have suggested some others, but I restricted it to your choices.  I thought it odd you asked to choose between 2 amps differing so much in price.  But hey??

 

 

Actually both amps have similar features...small footprint...can be used in mono or stereo configurations, made in the USA....decent power...

 

Not sure what you mean by Schitt's "history of problems"...and hum and shut down..is that wide spread, or a few squeaky wheels?

 

 

[esldude]

47 minutes ago, psjug said:

 

This is interesting, but if the goal is increasing SNR then it is a pretty inefficient way of doing it, compared to just going with greater bit depth.

I agree.

[esldude]

12 minutes ago, Ishmael Slapowitz said:

Actually both amps have similar features...small footprint...can be used in mono or stereo configurations, made in the USA....decent power...

 

Not sure what you mean by Schitt's "history of problems"...and hum and shut down..is that wide spread, or a few squeaky wheels?

 

 

I believe Chris had the shutdown issue. Caused by cables and speakers it didn't like. A few others reported it too. Supposed to be fixed.  Some reports still. Hum was a reported problem now and again.  Don't know how common it is?

[Ishmael Slapowitz]

1 minute ago, esldude said:

I believe Chris had the shutdown issue. Caused by cables and speakers it didn't like. A few others reported it too. Supposed to be fixed.  Some reports still. Hum was a reported problem now and again.  Don't know how common it is?

I guess it is hard to say how common...I know Stoddard seemed to be not the least bit concerned, and said the amp should work flawlessly. I am not sure if this was damage control, or the few claims were overblown.

 

I did hear from many users the Vidar runs rather warm.

[firedog]

7 hours ago, Ishmael Slapowitz said:

Actually both amps have similar features...small footprint...can be used in mono or stereo configurations, made in the USA....decent power...

 

Not sure what you mean by Schitt's "history of problems"...and hum and shut down..is that wide spread, or a few squeaky wheels?

 

 

The Vidar has had a fair amount of negative posts about it having problems of various types in various setups. Would definitely make me wary of it.
The rest of their stuff? Pretty much positive reports. The preamps and the integrated don't seem to have problems, and get very good reports about SQ, especially for the price. The DACs  and headamps have lots of fans, although I seem to remember a couple didn't measure to well at Amir's site. 

[Ishmael Slapowitz]

7 hours ago, firedog said:

The Vidar has had a fair amount of negative posts about it having problems of various types in various setups. Would definitely make me wary of it.
The rest of their stuff? Pretty much positive reports. The preamps and the integrated don't seem to have problems, and get very good reports about SQ, especially for the price. The DACs  and headamps have lots of fans, although I seem to remember a couple didn't measure to well at Amir's site. 

I agree on the other stuff. The Freya (and Saga)  is a little Miracle. The DACs are never on my radar due to lack of DSD support.

 

I had reservations about the Vidar when I followed the development on Head Fi. It was kind alike watching a car crash in slow motion.

 

However, I have also read many posts where users seem to have zero issues. 

[psjug]

18 hours ago, sandyk said:

Hi Richard

The Benchmark is not alone in giving decent Unweighted and Weighted noise figures these days , with even many kit amplifiers performing very well at a cost, typically including metalwork , transformers etc. of <Au $1,000 

 Even the recent less ambitious Silicon Chip magazine designed 200W SC200 Amplifier specifies Signal-to-Noise Ratio: -116dB unweighted with respect to 135W into 8Ω(20Hz-20kHz)

I have attached the specs for the previous more ambitious SC ULD3 amplifier (123dB Unweighted)

from March 2012, as well as their DAC from way back in Sept.2009 which is capable of further improvement using more recent voltage regulators etc.

 Undoubtedly there are numerous commercially available amplifiers and DACs that easily surpass the performance of these older designs these days.

 

Regards

Alex

Just out of curiosity, have you or @Shadders seen measurements of finished amplifiers made from these kinds of kits?  I'm not sure I've seen measurements of any amplifier other than the AHB2 that meets the 96dB THD+N at 1W.  Maybe nCore is right on the edge of meeting this with a 22KHz filter.  Maybe there are a few others, but I bet they are rare.

[Shadders]

1 minute ago, psjug said:

Just out of curiosity, have you or @Shadders seen measurements of finished amplifiers made from these kinds of kits?  I'm not sure I've seen measurements of any amplifier other than the AHB2 that meets the 96dB SNR+N at 1W.  Maybe nCore is right on the edge of meeting this with a 22KHz filter.  Maybe there are a few others, but I bet they are rare.

Hi,

Sandyk can confirm this, but the specification for the SC200 was measured for the completed amplifier - not simulation. The ULD was also the measured performance. If you purchase the kits/PCB, and follow the article instructions, then you will achieve the measured performance.

Regards,

Shadders.

[psjug]

1 hour ago, Shadders said:

Hi,

Sandyk can confirm this, but the specification for the SC200 was measured for the completed amplifier - not simulation. The ULD was also the measured performance. If you purchase the kits/PCB, and follow the article instructions, then you will achieve the measured performance.

Regards,

Shadders.

I can't find any numbers or curves showing THD+N at 1W for that amp.  The neuorochrome modulus stuff also looks like it may meet the 96dB at 1W, but the unit reviewed on ASR is not quite there.

 

Edit:  The ASR test of the neurochrome amp  is only slightly worse than the manufacturer curve (apparently due to intentional enclosure compromise), and this is into 4 ohms.  THD+N into 8 ohms is better for this amp - I think it may meet the 96dB at 1W mark.

[esldude]

You do realize with most speakers being about 86 db for one watt you only need -86 db to get everything down to 0 db SPL when measuring at 1 watt.  When taking into account Fletcher-Munson you can probably have complete perceptual silence with even less.  I'm all for good performance, but the idea you need THD+N this low is a little over the top for 1 watt.  And for most good amps you really are going to being talking about N(oise).  

[Paul R]

1 hour ago, psjug said:

Just out of curiosity, have you or @Shadders seen measurements of finished amplifiers made from these kinds of kits?  I'm not sure I've seen measurements of any amplifier other than the AHB2 that meets the 96dB THD+N at 1W.  Maybe nCore is right on the edge of meeting this with a 22KHz filter.  Maybe there are a few others, but I bet they are rare.

 

It is also interesting that one would pick THD+N when choosing to compare against the noise level of a digital signal.  Perhaps it would be good to very briefly make sure we are all on the same page here regarding what we are talking about. 

 

As far as I know, most ADCs have a voltage range of plus/minus 10V. A 16bit device has 2^16 (65536) possible values, which means a resolution of 20/65536, or 0.3mV.   That means the dynamic range in decibels would be  20 LOG₁₀  (65536)  which equals 96dB. In other words, the absolute noise floor of a 16bit device can never exceed -96dB.  It will never reach -96dB either, as there is electronics noise to measure. Commonly, 16bit devices like CD players have a much lower noise floor.  (I didn't go look them up, but if you do, be aware, most oversample before they hit the onboard DAC. You have to measure off their SPDIF outputs, and that doesn't always work either.) 

 

Now a 24 bit system never really hits 24bit resolution of course, but by the same maths as above, it's absolute noise floor would be -144dB. It's actual noise floor will be even lower than that due to electronics noise and such. 

 

And using just a little bit of that common sense, we can say the noise floor of any system will be limited but the highest noise floor of any electronic component in the chain. All of which says that THD+N is a good measure. Right? 

 

So, in comparing a 16 bit system - to compare apples to apples - you really need to look at components that exceed the noise floor, as measured by THD+N of the digital system. Not as may be happening here, by comparing the THD+N of an amp to the theoretical maximum noise floor of the digital system. 

 

Everyone agree?  If so, then getting the THD+N of the digital systems is important, and honestly, I don't have those. I would expect 20dB to be a reasonable figure, as suggested, so for 16 bit systems, you are looking for electronics quieter than about -76dB. For 24bit systems, perhaps a bit larger difference, say, 35dB, or around -109db. You can of course, play with those figures, but I think that hits the ballpark. And that may change your thinking on 16 bit systems a bit. Or not, as you will.  But amps that exceed the actual noise floor - how so ever you may measure it - for 16 bit systems seem to be pretty common. 

 

Also, so far as I know, most companies measure distortion over the entire Nyquist range using 2.3V shorted. At least, respectable manufacturers of high end audio gear do these days. That was so uncommon before, it is is good to be suspicious of it.  

 

The upshot of all that is that I firmly believe that 16bit audio covers most of the dynamic range of a recording, but that 24bit covers it all, with a bit of room to spare. Does that mean CDs don't sound great? Of course not, they do. But if I have the choice, I will also choose a 24bit high sample rate file for my archive. Always given of course, that the file is actually that, and not a upsampled 16/44.1k source - or worse.

 

-Paul 

 

[Rt66indierock]

4 minutes ago, esldude said:

You do realize with most speakers being about 86 db for one watt you only need -86 db to get everything down to 0 db SPL when measuring at 1 watt.  When taking into account Fletcher-Munson you can probably have complete perceptual silence with even less.  I'm all for good performance, but the idea you need THD+N this low is a little over the top for 1 watt.  And for most good amps you really are going to being talking about N(oise).  

 

My Heresy II speakers are 96 dB so I need a little more.