Holy Crap, Why use a pre-amp at all ?

[joessportster]

8 minutes ago, Dave A said:

I read about all the devices people use to hook their systems up with and I guess I am a simple guy. I like the idea of using my workstation for mixing music for the day and feeding directly to a Crown xli and then to the MCM 1900's. Outside of the speakers the single biggest fidelity boost I have found was when I discovered the Realtec sound card had a high definition program download of some 250+ MB. So I thought what is there to lose by trying and this made a huge difference. My speaker customers are amazed at how simple my setup is for the sound produced.

Dave am i understanding you correctly in that your workstation is a pc / mac, and you use the onboard DAC (which is realtec that you upgraded via download) feed it directly to crown amp ? You control volume using the onboard gain on the crown or ?

[joessportster]

17 minutes ago, dtel said:

Exactly, but I used the first version, it's also great to have a remote. 

 

I got it cheap when many were switching to the 2 version. I think many people get stuck on the brand for whatever reason but I would guess it sounds as good as many of the DAC's out there and built very well.  Myself I love things that easily outperforms their price range, again like the A-100 amp, better than it should be for the price.

Emotiva DAC / pre Interesting.....................

[Marvel]

You can get asio drivers for some of the Realtek audio chipsets, too.

 

Bruce

[PrestonTom]

It gets confusing since the marketing verbiage is not a substitute for technical specs. The issue of floating point (as opposed to integer) representation is restricted to any filtering, interpolation, sampling or decimation etc that the processor is doing. During the D - A conversion it is a different story, in spite of the marketing verbiage. The conversion is performed on integer values, not floating point. The higher bit values, say 16 bit or 24 bit, will lead to higher resolution (potentially). In turn this will decrease the quantization noise (potentially).

 

 Two things to remember. If the integers are scaled down ("digital attenuation"), the quantization noise is impacted. If the original signal was recorded at 16 bit, playing them at 24bit will not improve (lower) the quantization noise. Likewise throwing away a bit (halving the voltage or decreasing the output by 6 dB) will similarly impact the quantization noise.

 

My feeling is that the best attenuation (volume control) is done on the analog signal and not on the digital "bits". 

[Dave A]

15 minutes ago, joessportster said:

Dave am i understanding you correctly in that your workstation is a pc / mac, and you use the onboard DAC (which is realtec that you upgraded via download) feed it directly to crown amp ? You control volume using the onboard gain on the crown or ?

It is a Dell 3620 Workstation running Win10 and I suppose it works like a DAC but is it technically the same as a dedicated DAc in what it does I don't know. Windows Media has a volume  control as does the system sound settings and then the Crown also does. I turn the system volume up most of the way and the media player becomes my fine adjustment over the Crown volume control which goes in steps. You can use either the speaker out or headphones out plugs and then RCA's into the back of the Crown ( has two voltage inputs selectable) or my Integra D50.4. 

 

  I might be missing a lot here I just don't know. Later this year I will build some Super MWM bins and get some 402's to go on top of them and then I am told I will need to get more complicated to get the best sound. What I have today is good enough that I honestly could be happy right here and not spend another dime. But if this last step up is as large as the MCM 1900 has been over other speakers I have had I want to hear that too. 

[joessportster]

23 minutes ago, PrestonTom said:

It gets confusing since the marketing verbiage is not a substitute for technical specs. The issue of floating point (as opposed to integer) representation is restricted to any filtering, interpolation, sampling or decimation etc that the processor is doing. During the D - A conversion it is a different story, in spite of the marketing verbiage. The conversion is performed on integer values, not floating point. The higher bit values, say 16 bit or 24 bit, will lead to higher resolution (potentially). In turn this will decrease the quantization noise (potentially).

 

 Two things to remember. If the integers are scaled down ("digital attenuation"), the quantization noise is impacted. If the original signal was recorded at 16 bit, playing them at 24bit will not improve (lower) the quantization noise. Likewise throwing away a bit (halving the voltage or decreasing the output by 6 dB) will similarly impact the quantization noise.

 

My feeling is that the best attenuation (volume control) is done on the analog signal and not on the digital "bits". 

I still dont get how lowering output VOLUME affects the bits, My understanding and the argument I have heard hundreds of times is bits are bits  and cables etc... cannot improve upon which would also mean cannot degrade. I will have to take you word for it but what I can say is I do not perceive a loss of SQ when I lower the volume. Although I do find there is a SWEET spot to all volume controls. I will try and read more maybe I can get it clear in my head.

 

26 minutes ago, Dave A said:

It is a Dell 3620 Workstation running Win10 and I suppose it works like a DAC but is it technically the same as a dedicated DAc in what it does I don't know. Windows Media has a volume  control as does the system sound settings and then the Crown also does. I turn the system volume up most of the way and the media player becomes my fine adjustment over the Crown volume control which goes in steps. You can use either the speaker out or headphones out plugs and then RCA's into the back of the Crown ( has two voltage inputs selectable) or my Integra D50.4. 

 

  I might be missing a lot here I just don't know. Later this year I will build some Super MWM bins and get some 402's to go on top of them and then I am told I will need to get more complicated to get the best sound. What I have today is good enough that I honestly could be happy right here and not spend another dime. But if this last step up is as large as the MCM 1900 has been over other speakers I have had I want to hear that too. 

If we are not moving forward we are standing still   Sounds like the new setup (if when you get around to it) will require some type Bi-Amping and yes a whole new can of worms although not as complicated as it sounds. There are easier ways to approach it.  The sound cards Real-Tec Included do have an onboard dac

 

We are basically running the same set up except my DAC is external and I use J-river instead of windows media player and my amp is at max volume all the time so I must use Jriver to control it.

 

Do you have a link for the real tec download (My pc also uses real tec sound card)

[Dave A]

"realtek high definition audio driver" in your search bar should get you there if the link does not. There might be on from the pc makers driver page to.

 

 

 

 

 

http://www.realtek.com.tw/downloads/downloadsView.aspx?Langid=1&PNid=14&PFid=24&Level=4&Conn=3&DownTypeID=3&GetDown=false

[Dave A]

53 minutes ago, joessportster said:

Although I do find there is a SWEET spot to all volume controls. I will try and read more maybe I can get it clear in my head.

I seem to remember reading that you do not want to max out the PC volume controls

[dtel]

2 hours ago, joessportster said:

I still dont get how lowering output VOLUME affects the bits, My understanding and the argument I have heard hundreds of times is bits are bits 

It's true bits are bits but there are different kinds of volume controls, so loose bits at lower volumes.

 

For example,  part of the reason some people switched from the Emotiva version 1 DAC to the 2 version when it came out was on the first version it would lose some bits at lower volumes.

 

I got the first version because I didn't care, I only used the top 10% of the volume control which was not affected. I did this by using the amps volume controls at 50% or less to keep the DAC volume at 90%, it was a simple way to be able to use the first version which was available very cheap at the time. 

[pzannucci]

6 hours ago, dtel said:

It's true bits are bits but there are different kinds of volume controls, so loose bits at lower volumes.

 

For example,  part of the reason some people switched from the Emotiva version 1 DAC to the 2 version when it came out was on the first version it would lose some bits at lower volumes.

 

I got the first version because I didn't care, I only used the top 10% of the volume control which was not affected. I did this by using the amps volume controls at 50% or less to keep the DAC volume at 90%, it was a simple way to be able to use the first version which was available very cheap at the time. 

Yes there is a thing called precision.  A simple example is as you do certain math, you loose stuff like remainders.  An over simplified example is 13/3=4.333333.  In lower bit width math this could be 4 or 4.3 vs. 4.33333333333333333333.  All of those multiplied by some number is going to give you far different results.  This also works in diffusion / dithers that you hear about in dacs.

[joessportster]

OK but in my case I am lowering out put at the pc which goes to my dac and is upsampled to 24/192, my dac puts out what it puts out no volume control no way to raise or lower its output so does it become an issue in that situation ?  How would you know if it is an issue in a particular situation / setup ?  By Ear ? I cant recall having ever read a spec that says anything about fixed or floating points. also if a manufacturer Knew that there product lost fidelity when output was lowered why not simply limit the amount output could be lowered ?  Poor Fidelity will kill a product so it would be in there best interest to maintain a certain amount of fidelity

 

All builds are a compromise. 

[pzannucci]

23 hours ago, joessportster said:

OK but in my case I am lowering out put at the pc which goes to my dac and is upsampled to 24/192, my dac puts out what it puts out no volume control no way to raise or lower its output so does it become an issue in that situation ?  How would you know if it is an issue in a particular situation / setup ?  By Ear ? I cant recall having ever read a spec that says anything about fixed or floating points. also if a manufacturer Knew that there product lost fidelity when output was lowered why not simply limit the amount output could be lowered ?  Poor Fidelity will kill a product so it would be in there best interest to maintain a certain amount of fidelity

 

All builds are a compromise. 

You could only know by understanding the software that went into the volume control.  I believe 24 or 32bit should be adequate to maintain resolution as you drop the volume.  This is all done way before it gets to the DAC and is upsampled.

 

As far as poor product, math in real time, when playing music, is processing intensive.  Not long ago it was hard to do right without a dedicated processor.  This is another reason why you see high quality digital volume controls in a number of products vs. analog.  Or case in point, Emotiva XDA-1 vs XDA-2 which lost resolution at lower volumes and switched to the XDA-2 which I believe did volume in the analog domain.

[joessportster]

3 hours ago, pzannucci said:

You could only know by understanding the software that went into the volume control.  I believe 24 or 32bit should be adequate to maintain resolution as you drop the volume.  This is all done way before it gets to the DAC and is upsampled.

 

As far as poor product, math in real time, when playing music, is processing intensive.  Not long ago it was hard to do right without a dedicated processor.  This is another reason why you see high quality digital volume controls in a number of products vs. analog.  Or case in point, Emotiva XDA-1 vs XDA-2 which lost resolution at lower volumes and switched to the XDA-2 which I believe did volume in the analog domain.

I found a way to keep the Digital Volume near 100%, J-River also has an EQ on board which has a pre-amp adjustment I lower the output there and turn the volume near max, Sound wise I really dont perceive much difference. But what I read / gather analog volume control is the way to avoid the issue all together correct ? (at least within the PC / DACs abilities). Since my source is basically 100% digital now it would no doubt benefit me to Keep the signal as clean as Possible. (even if I can not hear it  )

[adam2434]

JRiver's "internal" volume control is 64 bit and is purported to be a very good implementation of digital volume control.

 

I believe it is best to set the JRiver output bitdepth to the highest bitdepth the DAC can handle (24 bit for most DACs).

 

https://wiki.jriver.com/index.php/Volume

[dtel]

On 7/19/2018 at 5:54 AM, pzannucci said:

Yes there is a thing called precision.  A simple example is as you do certain math, you loose stuff like remainders.  An over simplified example is 13/3=4.333333.  In lower bit width math this could be 4 or 4.3 vs. 4.33333333333333333333.  All of those multiplied by some number is going to give you far different results.  This also works in diffusion / dithers that you hear about in dacs.

It's more than math, it's a different volume pot, I remember something called stepped control volume pot.  I don't remember if it was the better one or the problem one, it was almost the only change between the Emotiva DAC version 1 and 2, plus the addition of a headphone jack. The realized to late they used a volume control that was actually losing bits at low volume, apparently after many were made and sold. 

 

You may know more about the different volume controls than me but I remember  reading about the problem because I had one and figure a way to work around it since I really liked the DAC. 

[pzannucci]

7 hours ago, dtel said:

It's more than math, it's a different volume pot, I remember something called stepped control volume pot.  I don't remember if it was the better one or the problem one, it was almost the only change between the Emotiva DAC version 1 and 2, plus the addition of a headphone jack. The realized to late they used a volume control that was actually losing bits at low volume, apparently after many were made and sold. 

 

You may know more about the different volume controls than me but I remember  reading about the problem because I had one and figure a way to work around it since I really liked the DAC. 

It is math, no pot.  Nothing is analog until after the conversion at the DAC that can use a pot.  It's all bits.

If it had a pot, there wouldn't be a problem (kind of).  Analog cab still has impedance issues but.

 

Since most computers today have significant processing power, I would hope math is done in a way as to not affect resolution.  Without being told or seeing the source code one would not know for sure.  

 

I believe anything upsampled to 24/192 before the math is done is immune to the problem.  Math precision should be high enough