Hi guys,
I've recently built a 6ch amp for my HTPC based on 3X Tripath TA2020.
I'm trying to improve the build by optimizing the signal path and remove unnecessary components such as the "volume control" pot (which should be for 6ch or 3X Stereo units...)
I currently feed the amp. with analog line level directly from the o/b sound based on ADI AD1986a aka. SoundMAX codec (motherboard is ASUS M2NPV-VM).
I might add an external sound card or just a DAC for better Stereo quality and feed it digitally from the PC (this M/B has digital outputs as well or I might go for a USB solution).
My current concern is regarding the quality of the Windows 7 volume control.
It's very convenient and I can easily balance between different channels, some DRC etc. with the remote control I already have...
I know that MS invested significant resources to improve the sound quality of Vista/7 and the digital volume control algorithms are pretty sophisticated.
So what do you think - do you still recommend some "pre amp" as separated volume control (i.e. some basic ALPS pots or go for one of the DIY kits) or will it just interfere within the signal path so I'll better stay with the PC volume controls ?
Any insights/suggestions/links will be appreciated,
Gal
P.S. - The PC is the only sound source and I don't intend to use the amp. with other setups...
I've recently built a 6ch amp for my HTPC based on 3X Tripath TA2020.
I'm trying to improve the build by optimizing the signal path and remove unnecessary components such as the "volume control" pot (which should be for 6ch or 3X Stereo units...)
I currently feed the amp. with analog line level directly from the o/b sound based on ADI AD1986a aka. SoundMAX codec (motherboard is ASUS M2NPV-VM).
I might add an external sound card or just a DAC for better Stereo quality and feed it digitally from the PC (this M/B has digital outputs as well or I might go for a USB solution).
My current concern is regarding the quality of the Windows 7 volume control.
It's very convenient and I can easily balance between different channels, some DRC etc. with the remote control I already have...
I know that MS invested significant resources to improve the sound quality of Vista/7 and the digital volume control algorithms are pretty sophisticated.
So what do you think - do you still recommend some "pre amp" as separated volume control (i.e. some basic ALPS pots or go for one of the DIY kits) or will it just interfere within the signal path so I'll better stay with the PC volume controls ?
Any insights/suggestions/links will be appreciated,
Gal
P.S. - The PC is the only sound source and I don't intend to use the amp. with other setups...
Hi!
As far as I know, digital volume control in the pc is always achieved by reducing the bit-depth of the signal. This is quite ok, if you adjust the volume from the pc in the upper one-third of the volume range. But the problem is, with no preamp (volume pot) at all, you have to stay in the lowest adjustment range most of the time for "normal" listening levels, and that means you will get a heavily quality-reduced signal out of the pc most of the time.
I use a passive pre (blue Alps pot) between my external dac and my power amp to make a pre adjustment, then I can adjust the volume from the pc (via wireless mouse) in a comfortable range without such a great loss of quality. This works perfectly for me.
By the way, in case you've never heard of it, here's a small and very useful application for the pc to adjust volume with the mouse wheel:
Volumouse - control the sound volume with a wheel mouse
Regards!
martin
As far as I know, digital volume control in the pc is always achieved by reducing the bit-depth of the signal. This is quite ok, if you adjust the volume from the pc in the upper one-third of the volume range. But the problem is, with no preamp (volume pot) at all, you have to stay in the lowest adjustment range most of the time for "normal" listening levels, and that means you will get a heavily quality-reduced signal out of the pc most of the time.
I use a passive pre (blue Alps pot) between my external dac and my power amp to make a pre adjustment, then I can adjust the volume from the pc (via wireless mouse) in a comfortable range without such a great loss of quality. This works perfectly for me.
By the way, in case you've never heard of it, here's a small and very useful application for the pc to adjust volume with the mouse wheel:
Volumouse - control the sound volume with a wheel mouse
Regards!
martin
Last edited:
Thank you martin for the informatin and the link 
but I already have almost total control on the HTPC with MCE remote kit (and other equipment using the IR blasters) so no need for mouse interface...
Quote:Vista will adjust the volume using internal floating point values rather than integer. So it does not reduce resolution as XP did.
Another good read: An audiophile’s look at the audio stack in Windows Vista and 7
In any case - If a "pre-amp" is unavoidable - what will be the best way to do it? (in order to get significantly superior quality than just using the Win7 controls...)
- Integrate the pot in the amp chassis ? (exposed to internal "noise" but close to the amp. input)
- Building a separate "passive" component with the pot shielded and wired on top of the 3 stereo line signals?
- Using an "active" pre-amp kit - i.e.
Crystal PGA2311 Volume remote control preamplifier kit
(Or any other recommended kit - on budget please...)
- shall I use one 6 Channel pot like:
alps-100kx6
Or better to go with 3 separated 2ch controls such as:
ALPS 09 Potentiometer
Cost is pretty much comparable...
Any other insights, suggestions, etc. ?
Gal
It's a really cool application
but I already have almost total control on the HTPC with MCE remote kit (and other equipment using the IR blasters) so no need for mouse interface...Is this still true in Vista/ Win7 !?
Quote:Vista will adjust the volume using internal floating point values rather than integer. So it does not reduce resolution as XP did.
Another good read: An audiophile’s look at the audio stack in Windows Vista and 7
In any case - If a "pre-amp" is unavoidable - what will be the best way to do it? (in order to get significantly superior quality than just using the Win7 controls...)
- Integrate the pot in the amp chassis ? (exposed to internal "noise" but close to the amp. input)
- Building a separate "passive" component with the pot shielded and wired on top of the 3 stereo line signals?
- Using an "active" pre-amp kit - i.e.
Crystal PGA2311 Volume remote control preamplifier kit
(Or any other recommended kit - on budget please...
)- shall I use one 6 Channel pot like:
alps-100kx6
Or better to go with 3 separated 2ch controls such as:
ALPS 09 Potentiometer
Cost is pretty much comparable...
Any other insights, suggestions, etc. ?
Gal
Last edited:
If I'm getting it correctly (from the links I've posted above) - they solve that exact issue by up-converting all audio streams to 32bit before processing:
"...
The new audio stack automatically upconverts all streams to a 32-bit floating-point sample depth (the same that is used in professional studios) and mixes them with the same precision. Because of the amount of headroom that comes with using 32-bit floats, there is no more clipping when playing two samples at the same time. There is also no loss of resolution when you lower the volume of a stream...
..."
So !?!?!
"...
The new audio stack automatically upconverts all streams to a 32-bit floating-point sample depth (the same that is used in professional studios) and mixes them with the same precision. Because of the amount of headroom that comes with using 32-bit floats, there is no more clipping when playing two samples at the same time. There is also no loss of resolution when you lower the volume of a stream...
..."
So !?!?!
Hi Andrew,
I'm not an expert (far from it to be honest)-It was a quote from the link I previously posted:
An audiophile’s look at the audio stack in Windows Vista and 7
It's probably explained better there (easy and recommended read)- but in my laymen words, since they process the audio stream in the unconverted 32bit domain, there is enough headroom to maintain all the original sample resolution (24bit) even if you lower the volume level...
Can anyone verify?
Gal
I'm not an expert (far from it to be honest)-It was a quote from the link I previously posted:
An audiophile’s look at the audio stack in Windows Vista and 7
It's probably explained better there (easy and recommended read)- but in my laymen words, since they process the audio stream in the unconverted 32bit domain, there is enough headroom to maintain all the original sample resolution (24bit) even if you lower the volume level...
Can anyone verify?
Gal
Hmmmmm... I missed that part that ASIO bypass all this... How about using WASAPI instead? maybe good enough and still gives the benefit of that Upconvert "madness" (btw - I think it's completely different than up-sampling and shouldn't change the raw data...)
My main concern is that I'll need to control multiple (6ch) audio streams and not only 2CH stereo music...
I might add a DAC in the near future to improve the 2ch stereo music - but I'll still need to control the volume level.
I won't mind doing some more DIY or adding a PCB or two if you can recommend something decent (Preamp / DAC combination?).
I already opened a related discussion on that matter so will appreciate if you could reply there...
http://www.diyaudio.com/forums/digital-line-level/167448-upgrading-mocha-jy-m2-multi-channel-dac.html
Please consider that I'm on a tight budget and would like to keep my setup as compact and minimal as possible - so would like to avoid purchasing additional devices such as A/V receivers (to decode multi channel audio streams) or an expensive commercial Preamp / DAC.
Thank you,
Gal
As I understand it - The idea is that by adding the extra "bit" depth BEFORE the processing (before dividing by 10) you can preserve the "fractional" data (these small numbers behind the decimal point ) that was previously been truncate and ignored, hence, lost...
Afterwards they can stream the best bit-depth (24bit!?) the HW support by taking the most significant bits of the unconverted sample.
It's very similar to processing images (my area of expertise) derived as high bit depth data - usually from HW that can output 32/48bits or more - and eventually reduce the bit depth by extracting only the significant data bits out of it, hence, preserving almost the whole dynamic range (same as in audio) of the original sample.
I hope I manage to present this property - but as you might guess - I'm not a native English speaker nor an engineering professor ...
Gal
) that was previously been truncate and ignored, hence, lost...Afterwards they can stream the best bit-depth (24bit!?) the HW support by taking the most significant bits of the unconverted sample.
It's very similar to processing images (my area of expertise) derived as high bit depth data - usually from HW that can output 32/48bits or more - and eventually reduce the bit depth by extracting only the significant data bits out of it, hence, preserving almost the whole dynamic range (same as in audio) of the original sample.
I hope I manage to present this property - but as you might guess - I'm not a native English speaker nor an engineering professor ...
Gal
OK guys,maybe here comes something decent:
-theres a ,"the" ,firewire audio controller, has jitter around 50ps or so. Pro audio companies use it in their soundcard stuff, from entry to high end.The deal is, you can buy an "alesis io 14" for 150 usd and i2s its all there, 6 out no problem.
Ofcourse no more stinkin PC smps supply either cause you can jsut swap that out for good. I2s is right there. This soundcard is discontinued so no warranty. You could put some better dac into that 5.1 enclosure of yours that looks cool. Remind you dont want spdif anymore , solely i2s because of the firewire controller . That thing fits into the 5.1 box fine, too. You sure can pick up some fine sigma delta DA for cheap, go for pcm1792-8. That 5.1 power supply looks good enough already. Then you have your 120 SNR, that is , 20 bit accurate.
-theres a ,"the" ,firewire audio controller, has jitter around 50ps or so. Pro audio companies use it in their soundcard stuff, from entry to high end.The deal is, you can buy an "alesis io 14" for 150 usd and i2s its all there, 6 out no problem.
Ofcourse no more stinkin PC smps supply either cause you can jsut swap that out for good. I2s is right there. This soundcard is discontinued so no warranty. You could put some better dac into that 5.1 enclosure of yours that looks cool. Remind you dont want spdif anymore , solely i2s because of the firewire controller . That thing fits into the 5.1 box fine, too. You sure can pick up some fine sigma delta DA for cheap, go for pcm1792-8. That 5.1 power supply looks good enough already. Then you have your 120 SNR, that is , 20 bit accurate.
I am always amused about the floating point high resolution 64bit no-resolution-loss volume control in state-of-the-art whatever.
No matter what source data are, what "sophisticated volume control DSP" is used, there is always at most 24 bit integer DAC with up to 20bit real-world resolution at the end of the digital chain. Any volume control is basically just a division of all samples by a single fixed number, no matter whether in int or float. In the end all samples get converted to 24bit int at best.
Volume control in the digital domain is ALWAYS about losing resolution. Another question is whether it is noticeable on your gear.
No matter what source data are, what "sophisticated volume control DSP" is used, there is always at most 24 bit integer DAC with up to 20bit real-world resolution at the end of the digital chain. Any volume control is basically just a division of all samples by a single fixed number, no matter whether in int or float. In the end all samples get converted to 24bit int at best.
Volume control in the digital domain is ALWAYS about losing resolution. Another question is whether it is noticeable on your gear.
Correct - and it's stated that eventually the data will be at most 24bit since that's what the HW (DAC!?) support...
However, while some resolution will indeed be lost in the digital domain - AFAIK the main benefit of the upscaling is reducing the losses while processing (i.e. decreasing the volume!?)...
In the linked example the resolution "loss" is within 0.04% of the “ideal” samples which is a range I personally can live with...
Especially considering the fact that my alternative is adding few DIY level (basic ALPS pot !?) components to the signal path with the extra wire connections etc. Now - I'm pretty skilled but still not sure I'll achieve significantly better results (probably worse !?)...
I'm not talking about HiEnd , Hi$$$ source/DAC/Preamp/amp - just a humble "budding audiophile" trying to upgrade his HTPC based system to a decent listening machine with best cost/benefit ratio and is not afraid from DIY...
So - any suggestions?
Gal
The example study is comparing 16 bit int processing chain to 32bit floating point one. Of course you will get better results for volume control, since 16bit int itself is causing resolution loss within the chain (vast majority of DACs are 24bits these days). Any system using 16bit-only processing is pretty outdated. That still does not refute the fact that digital volume control, where the output value is inherently lower than the input one, leads inevitably to resolution loss. Unless your input source has a higher real resolution (not just by plain overtyping to another numerical type, even with dither) than the DAC, in which case the chain has a few bits of headroom for the division operation in volume control. I do not know of any generally available recordings better than 24 bits as the dynamic range of the best existing ADCs does not even reach the 24 bits of their technical resolution. True, electronic artifically-made music could be generated at any resolution, but I have not seen any tracks like that yet.
However, a similar resolution loss occurs in analog domain volume control too, if noise level of the audio chain is not changed by the volume control. I like the idea used in some digital amplifiers which combine digital domain attenuation with changing output bridge power supply voltage. Though I assume proper perfection of all details in such a complicated technology is no easy task and the actual results may be worse than top quality volume attenuation combined with classical linear amplifiers.
However, a similar resolution loss occurs in analog domain volume control too, if noise level of the audio chain is not changed by the volume control. I like the idea used in some digital amplifiers which combine digital domain attenuation with changing output bridge power supply voltage. Though I assume proper perfection of all details in such a complicated technology is no easy task and the actual results may be worse than top quality volume attenuation combined with classical linear amplifiers.
The best way will be to keep the soundcard close to clipping point (testing that with a maximum amplitude wave file) and attunate the volume external.
I think you can compare it with digital zoom or optical zoom on a camera (digital zoom is just a algorithm with loss of detail).
You can upscale to high resolution then do the math(volume attunate) and downscale afterward. It will have almost no effect, in the downgrade the details will be lost also.
I think you can compare it with digital zoom or optical zoom on a camera (digital zoom is just a algorithm with loss of detail).
You can upscale to high resolution then do the math(volume attunate) and downscale afterward. It will have almost no effect, in the downgrade the details will be lost also.
Isn't that exactly what the "new" Win7/Vista volume control intend to do ?
I understand there are some fine details lost when downgrading - but it should be the "trunked" data beyond the 24bit range -
Won't it be negligible or perhaps even superior Vs. the losses/degradations I'll encounter adding an extra "DIY" grade analog volume control with all these extra wires, connections, soldering points and a basic ALPS pot to the audio path?
I'm really trying to get the best solution - but considering the whole picture - I'm still not convinced the additional analog solution will indeed gives a significantly superior solution - perhaps even inferior to using the win7 digital volume control and get the rest of the setup much simpler with less components within the audio path...
I guess i'll just have to experiment a little bit - w/o the analog pot I need to set the win7 volume to around 25-30% for reasonable listening level - but my amp setup is still "under construction" so I still can't evaluate the differences between the alternatives...
Gal
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.