A quick poll on role of DAC and Dynamics

Ravi, the best for you would be to experiment yourself and experience the differences personally.

Cheers
 
Thanks for the advise.

I guess buying the best available sound card just for the spdif out is more of a personal choice rather than a hard and fast rule to get the best out of the computer audio.

I'm wondering how Asus would market their new DAC which is to be launched shortly. Its the Asus Essence One DAC especially targeted at audiophiles using computer as a transport with audiophile grade headphone amplifier.
Would they recommend to buy their Essence STX sound card and use the spdif out to the new Essence One DAC to get the best audio experience?
Or would they be coming out with a stripped down version of STX without the redundant components the DAC and analog output stage especially for the Essence One users?
I would like to wait till its release and see how Asus would be recommending using their DAC. I'm 100% sure the answer would be its designed to use with USB/ an board spdif :D
 
Ravi and Fusioner, the points you have raised have been discussed ad-nauseum many times here at HFV, and on the Net.

Let me see if I can explain this in simple terms.

There are two aspects to a computer's out.

When you are sending data out from a computer it also contains noise.


The advantage data has is two fold - one;the noise can be easily identified and removed, two; data can be verified at the receiving end and retransmitted, if needed. In addition when you are sending data, there is no DAC conversion as the receiving end also understand digital data.

When you are sending sound in digital form - the digital data is a set of discrete representation of the sound's sine wave. Even if you chop off noise above 20,000 Hz and, say, below 50 Hz, noise within the bandwidth of 50Hz to 20,000 will form part of the data that represents the sound. You cannot differentiate between actual data (signal) and noise. This is called signal to noise ratio. The higher the signal to noise ratio, the better your signal is. When you convert a noisy digital signal to analogue, you are adding noise to the sound.

A motherboard, per se, does not bother about this as it expects to transmit and receive only digital 'data'. The parts of a MB is thus optimised for speed and cost. A sound chip in a MB has to work within this restrictive environment.

A sound card, on the other hand, is optimised for the highest possible signal to noise ratio. It has it's own power supply and picks up the sound data from the hard disk at the kernel level skipping the operating system. None of the other environments of the computer has any effect on the signal.

Here is what the Asus STX does, for example.



All the parts used in a sound card such as the Asus STX are optimised to pick up the audio data and make it reach your ears with the least amount of interference.

Cheers

Venkat, where does that noise come from?
 
Ravi and Fusioner, the points you have raised have been discussed ad-nauseum many times here at HFV, and on the Net.

Let me see if I can explain this in simple terms.

There are two aspects to a computer's out.

When you are sending data out from a computer it also contains noise. The advantage data has is two fold - one;the noise can be easily identified and removed, two; data can be verified at the receiving end and retransmitted, if needed. In addition when you are sending data, there is no DAC conversion as the receiving end also understand digital data.

When you are sending sound in digital form - the digital data is a set of discrete representation of the sound's sine wave. Even if you chop off noise above 20,000 Hz and, say, below 50 Hz, noise within the bandwidth of 50Hz to 20,000 will form part of the data that represents the sound. You cannot differentiate between actual data (signal) and noise. This is called signal to noise ratio. The higher the signal to noise ratio, the better your signal is. When you convert a noisy digital signal to analogue, you are adding noise to the sound.

A motherboard, per se, does not bother about this as it expects to transmit and receive only digital 'data'. The parts of a MB is thus optimised for speed and cost. A sound chip in a MB has to work within this restrictive environment.

A sound card, on the other hand, is optimised for the highest possible signal to noise ratio. It has it's own power supply and picks up the sound data from the hard disk at the kernel level skipping the operating system. None of the other environments of the computer has any effect on the signal.

Here is what the Asus STX does, for example.



All the parts used in a sound card such as the Asus STX are optimised to pick up the audio data and make it reach your ears with the least amount of interference.

Cheers

Venkat - I really appreciate your contribution to this forum, however I will have to be candid this time! I believe you should use more of your first hand knowledge rather than just your experience :)

Your analytical mind sometimes take you to places which do not exist... I have seen this in many of your post when you build an incorrect co-relation or inference and share the same in these posts.. Being a moderator, you should avoid speculating things unless we really KNOW they are true.

Here you have DISCOVERED DIGITAL SIGNAL TO NOISE RATIO!!! :yahoo::clapping:

Both Signal and Noise are ANALOG terms and do not have any EXISTENCE in DIGITAL world. Only errors and ARTIFACTS can exist in Digital data.

Sound card, even a Rs.500 one will not USE its DAC unless decoding analog audio. Hence, whether you use the "motherboard" sound card or your Asus STX, it would NOT make any difference if the output is thru SPDIF.

A "good" sound card essentially has a very good DSP which is VERY useful when you NEED resampling (ideally we should NEVER RESAMPLE to a NON-INTEGER multiple of the original PCM AUDIO). Like if you HAVE to send output to an AVR or any other device which only accept 48kHz sample rate and take a CD rip, you will have to do a resampling and HOPE that you get somewhere close to the original sound.

EMIs or not, it will not hamper the sound unless they are so strong to change the bit value from 1 to 0 or vice versa. Nevertheless, if that happens, it will be ERROR and not noise.

I agree that Asus STX card must be great for Analog output and must be having one of the best DACs/DSPs available. Also, most importantly, it must be sending SPDIF especially without any resampling and hence you are getting best audio possible through your AVR's DAC. Same can apply to an onboard sound card which bypass resampling...

I give the example of my setup... I am currently using a THX certified Creative Audigy 2 ZS sound card which costed more than half of the cost of my motherboard.. However this card has a serious limitation that its DSP can not be bypassed and can only send SPDIF output in 48kHz or 96kHz. On the other hand, the same card's DAC can handle ANY sampling rate till 192kHz. So all the sound has to be resampled to either of those 2 frequencies before being sent. This significantly reduces the SQ despite the DAC being one of the best with analog output's SNR of more than 100dB and noise floor crossing 110dB (these figures will come to picture ONLY when I take analog audio). Its DSP even use an 8 point interpolation when resampling or processing digital audio by adding "Hollywood" quality sound effects, however it still fails the direct 44.1kHz output thru the onboard sound card's SPDIF. And my onboard card sound MUCH BETTER for 44.1kHz content as it allows unprocessed 44.1kHz (or 88.2kHz or 176.4kHz) PCM through SPDIF.

This is what you need to understand - DAC has NOTHING to do with Digital Audio and it is used ONLY for the purpose as the name suggests - Digital to Analog Converter!!!

Thad, what I was trying to impress here is that a sound card works better than a motherboard in extracting sound as people were talking about an USB output from the PC/Laptop itself. That is all.

BTW the fan noise is an audible noise. We are talking about electrical noises. :)

Cheers
Well, fan noise would certainly add to the signal to noise ratio of your ears which is the ultimate destination of the PCM AUDIO.. Electrical noise can AT-MOST generate ERRORS..
 
Last edited:
That reminds me - I forgot about upsampling to integer values. Must check with Venu if his DIY DAC does that. Didnt know about 176.4khz, thought it was only 88.2. Good to know.

Which of the two - ESI vs Asus does 88.2 and 176.4?

--G
 
Rahul
My god this thread has too many deliberations!
Science overdose!

though many members prefer reviews and tech specs and use them as more than a guide - im quite the opposite

ive used quite a few dacs by now only to deduce the following
This SC or that
This DAC or that
it all just system specfic.Theres just no sureshot way of knowing.


a highly reviewed dac or SC may just not sound right to you.
You'll have to just take the plunge . ( dont worry it will never be a bad descision whichever way you go, just an evolution )
Youre not going to get it right till you try

Though i use a Juli i see your preference for the Asus

I would advice you get it
use the analogue outs
(The digital out is always there later for your use with a dac - and dont worry, that digital out wont matter too much vs other sc's as compared to the dac you will change in the chain)

later you will just have to get a couple of dacs to try when youre ready to buy
Try a tube dac maybe youll like it - all the good ones are not cheap though.
 
Venkat - I really appreciate your contribution to this forum, however I will have to be candid this time! I believe you should use more of your first hand knowledge rather than just your experience :)

Your analytical mind sometimes take you to places which do not exist... I have seen this in many of your post when you build an incorrect co-relation or inference and share the same in these posts.. Being a moderator, you should avoid speculating things unless we really KNOW they are true.

Both Signal and Noise are ANALOG terms and do not have any EXISTENCE in DIGITAL world. Only errors and ARTIFACTS can exist in Digital data.

I think you have not read my posts properly and are coming to conclusions.

I have been talking about three distinct processes and their interrelationship. One is extraction of (music) data from a hard drive, the other is the DAC component, and the third is how important extraction is to DAC. I am not even talking about the transmission of these signals across any medium - S/PDIF, USB, Optical or other wise.

Digital extraction is through simple electrical circuits, and is susceptible to interference. What you want to call it is your choice - noise, interference, errors, or whatever. If you tell me there is no interference at all, then I really have nothing to say.

My point was simple - if you send near perfect data to the DAC process using any transmission method, the resultant analogue sound will be as close to the original sine wave as possible. This data could be, as you call it, with the least amount of errors. The only way you can remove these errors is by re-reading the original data. Once a data reaches out of the source, it cannot be re-read or re-sampled if you find errors. If you have used EAC, you will understand what I am talking about. As a program, EAC is not sure of what it is getting is correct, unless it is matched to a set of parameters it sets for itself. It rereads the data up to 20 times, and get what it thinks is the best, by comparing to a set standard database or to a calculated average from the data that is extracted. Mind you, all this is in the extraction process from the CD to the hard disk. A sound card simple takes this to the next step - from the hard disk to the memory.

BTW, I never say anything without experiencing it myself. I have used multiple sound cards, multiple DACs, and multiple PCs to experience the same song connected to multiple speakers, pre-amps, and power amplifiers.

In terms of analysis, I do believe in organisations such as Stereophile who do a lot of research and testing. I, for one, have no issues with quoting such articles IF it is backed by scientific and analysis. If I run my life disbelieving everyone, my life would be completely chaotic.

Cheers
 
Just wanted to share an observation with you guys about the same DAC performing differently on different machines.

I am using this HRT Streamer II with my netbook, which is my main music source. Yesterday I hooked the DAC to my main PC (the amp, cables and speakers were all same) and realized that the whole setup was sounding a lot better and more authoritative. Couldn't really understand why would that happen. Probably the DAC was being fed the right voltage on my main PC, than it was on the netbook.

Got me a lot more confused since I am now listening more on my PC and the idea of using the netbook as my source has more or less gone into drains.
 
That reminds me - I forgot about upsampling to integer values. Must check with Venu if his DIY DAC does that. Didnt know about 176.4khz, thought it was only 88.2. Good to know.

Which of the two - ESI vs Asus does 88.2 and 176.4?

--G
Most of the sound cards including the onboard ones which supports upto 192kHz sampling rate can do it unless their DSP locks these rates like my sound card.

I think you have not read my posts properly and are coming to conclusions.

I have been talking about three distinct processes and their interrelationship. One is extraction of (music) data from a hard drive, the other is the DAC component, and the third is how important extraction is to DAC. I am not even talking about the transmission of these signals across any medium - S/PDIF, USB, Optical or other wise.

Digital extraction is through simple electrical circuits, and is susceptible to interference. What you want to call it is your choice - noise, interference, errors, or whatever. If you tell me there is no interference at all, then I really have nothing to say.

My point was simple - if you send near perfect data to the DAC process using any transmission method, the resultant analogue sound will be as close to the original sine wave as possible. This data could be, as you call it, with the least amount of errors. The only way you can remove these errors is by re-reading the original data. Once a data reaches out of the source, it cannot be re-read or re-sampled if you find errors. If you have used EAC, you will understand what I am talking about. As a program, EAC is not sure of what it is getting is correct, unless it is matched to a set of parameters it sets for itself. It rereads the data up to 20 times, and get what it thinks is the best, by comparing to a set standard database or to a calculated average from the data that is extracted. Mind you, all this is in the extraction process from the CD to the hard disk. A sound card simple takes this to the next step - from the hard disk to the memory.

BTW, I never say anything without experiencing it myself. I have used multiple sound cards, multiple DACs, and multiple PCs to experience the same song connected to multiple speakers, pre-amps, and power amplifiers.

In terms of analysis, I do believe in organisations such as Stereophile who do a lot of research and testing. I, for one, have no issues with quoting such articles IF it is backed by scientific and analysis. If I run my life disbelieving everyone, my life would be completely chaotic.

Cheers

Sir, like you have also done a lot of research and used multiple products. And I am sure many have done the same as well particularly on this forum and that is the reason we are so tightly integrated to each other and so passionate about discussing these things. However, your analysis at times speak of pure analysis and experience like I used to do before I did my engineering... I hope you would not take it otherwise, however unless we have a proof in hand and know the REASON behind something, we should hold our horses in recommending things...

Taking of processing compressed audio and electrical interference causing issues - firstly the sound card is NOT in picture here (which I believe you would agree as well) and secondly we both know that a computer data has some very tight integrity checks which can not allow a CORRUPT data stream to be processed even when it it audio or video data. Exceptions are with A/V sources like Audio or Video CDs which do not have many integrity checks integrated to the media.. However in these media as well (particularly Audio CDs), there are "C2" algorithms on the current drives to detect and to an extent correct read errors... However when we talk of processing something like mp3 or FLAC... I don't see any reason to believe that EMI will cause these streams to alter as till they get passed to the sound card, they will be treated as a regular digital data with all the CRC integrity checks in place..

You talk more in terms of analog electronics when evaluating these things. However, despite your experience (which might be more associated with other variables), I can assure you that its almost impossible for these streams to be altered by external influence and definitely it does not amount to added noise.. What it CAN do is to introduce jitters, spikes and static to the PCM stream and unless you do not notice this, it should not be of ANY concern..
 
Venkat, where does that noise come from?

Venkatcr, You missed my query. From a subsequent post, I am guessing that you imply that noise is coming from interference/EMI interacting with the electric circuits carrying the digital signals. Is that right?
So the question is, can EMI alter digital signals/ or add noise to it?

Most of the sound cards including the onboard ones which supports upto 192kHz sampling rate can do it unless their DSP locks these rates like my sound card.



Sir, like you have also done a lot of research and used multiple products. And I am sure many have done the same as well particularly on this forum and that is the reason we are so tightly integrated to each other and so passionate about discussing these things. However, your analysis at times speak of pure analysis and experience like I used to do before I did my engineering... I hope you would not take it otherwise, however unless we have a proof in hand and know the REASON behind something, we should hold our horses in recommending things...

Taking of processing compressed audio and electrical interference causing issues - firstly the sound card is NOT in picture here (which I believe you would agree as well) and secondly we both know that a computer data has some very tight integrity checks which can not allow a CORRUPT data stream to be processed even when it it audio or video data. Exceptions are with A/V sources like Audio or Video CDs which do not have many integrity checks integrated to the media.. However in these media as well (particularly Audio CDs), there are "C2" algorithms on the current drives to detect and to an extent correct read errors... However when we talk of processing something like mp3 or FLAC... I don't see any reason to believe that EMI will cause these streams to alter as till they get passed to the sound card, they will be treated as a regular digital data with all the CRC integrity checks in place..

You talk more in terms of analog electronics when evaluating these things. However, despite your experience (which might be more associated with other variables), I can assure you that its almost impossible for these streams to be altered by external influence and definitely it does not amount to added noise.. What it CAN do is to introduce jitters, spikes and static to the PCM stream and unless you do not notice this, it should not be of ANY concern..
 
So the question is, can EMI alter digital signals/ or add noise to it?
Back to basics.

My Word document is full of spelling mistakes: must be the electrical noise inside the machine.

No. It doesn't happen --- or when it does, we speak of data corruption and look for something that is broken or breaking. Recently I found one of my tracks playing like the lady was screeching, not singing. I cannot explain how that sort of corruption can happen (the human operator is always suspected! :o) but luckily I had another copy, and the problem was fixed.
 
Well, neither am I, to be honest.

I am clear about digital data being digital data and that's it --- but there is a lot that, not being a hardware guy (or a mathematician) that I do not understand about the pathway between bits on disk and sound coming from a speaker, and it seems to me that the more we too and fro the whole thing, disagreements and all, the more we'll learn about it.

One thing about bits on a disk: I think getting them off a HDD is a lot more reliable than getting them off a CD that might not be in prime condition (and some CDs seem to come brand-new complete with faults already built in.) Something like EAC will (I think) read, and re-read, until it gets a sensible result, and will even try to access data from elsewhere, as Vencat has mentioned.

Remember the Floppy days? brrrrrrrrrrrr brrrr brrrrrrrrrrrrrrr TIK.... TIK..... TIK < Abort? Retry? Cancel? > (and what the hell was the difference between abort and cancel?)
 
I believe EMI/RFI do have a huge impact on the DACs

SPDIF: from what know works in the 2-3 Mhz range (Can anyone confirm ? )
EMI/RFI works in the 1KHZ to 10Ghz range. so there is a chance for interference and distortion..although it is random and unpredictable.

I believe it can not just interfere with the SPDIF signal , but also get carried along with the signal and affect the DACs components

Thats the reason why both Pro and home audio manufacturers build in designs for keeping the same out. Also this is true not just for Audio but also medical devices.

in fact this is one reason why Toslink/optical is favoured in certain situations and entry level DVDps as well, as they are immune to EMI/RFI
 
So we are talking about something that does not interfere with the bits of a digital signal, but that can get carried along the same wire, basically, because it's a wire? So the external DAC or Audio interface, which we think is isolated from the PC, might not be at all isolated?

I was looking at that paper The art of building Computer Transports earlier --- and the writer there says a big no to metalic digital signal carriers. On the other hand, I think I saw someone say (on this forum?) that the Toslink type of optical connector we have for domestic kit is inferior.
 
Last edited:
I was looking at that paper The art of building Computer Transports earlier --- and the writer there says a big no to metalic digital signal carriers. On the other hand, I think I saw someone say (on this forum?) that the Toslink type of optical connector we have for domestic kit is inferior.

The Wikipedia article on SPDIF the Limitations heading says "S/PDIF lacks flow control and retransmission facilities, which limits its usefulness in data communications applications" :lol:
Also it says "TOSLINK cables are not susceptible to ground loops and RF interference, as coaxial cables are".

If that is true then SPDIF/Co-axial is not a foolproof technology than USB for computer audio.
 
Last edited:
The Wikipedia article on SPDIF the Limitations heading says "S/PDIF lacks flow control and retransmission facilities, which limits its usefulness in data communications applications" :lol:
Also it says "TOSLINK cables are not susceptible to ground loops and RF interference, as coaxial cables are".

If that is true then SPDIF/Co-axial is not a foolproof technology than USB for computer audio.

Actually the SPDIF is a compromised format, but we are stuck with it. I2S is better but for very short runs (less than 1/2 feet) above which it distorts too.

USB/Firewire are definitely better but have been begun to be understood only now and hence being used so much better. hence the future is definitely in that path.(although with USB3 Firewire may just get extinct)
 
Yes, but that luxury is not always present. To put it in fanciful terms -
I am hoping for that "standing on the rocks with dangerous waves and surf crashing" experience. Its wonderful when snap and slam catches you unawares and you find yourself reacting belatedly and find you heart racing unpredictably. :licklips:

Wokay maybe an yexaggerration ... but wy yam lige thad wonly!! :ohyeah:

--G

For a 'standing on the rocks, with a tsunami heading straight for you' experience, one needs to view the system AND the room in totality. Focusing on one component, like the source, would come into play, when it is the only weak link in a set up. But before that, all the other variables in the mix have to be fixed.

I acheived the quickest and most dramatic improvement in dynamics, by fixing the power supply, and installing an online UPS. Out of all my hifi adventures, this was the greatest victory. It was like the sun bursting out of a cloud coated sky!

The next essential tweak, IMO, is to 'fix' the room. To understand it's acoustic strength's and weaknesses. To find the right place for the audio rack, speakers and all the furniture in the room.

The next step is debatable, but I believe choosing the right speaker, depending on the room size and listening preferences comes next.

Once the speaker has been 'fixed', the next step would be to find the right amp to drive it. Based more on what I have read, and less on what I have actually heard, I believe that if dynamics are a primary concern, then solid state is the only way to go. The decision to opt for a pre/power or an integrated should also be made after some deliberation and auditions.

Once all these variables become constant, a few home auditons of cdp v/s pc/dac would settle the issue. My preference for a cdp is based primarily on aesthetic reasons. I do not want a desk top/ laptop in my audio rack. If I had to go in this direction, I would keep the pc in a different room, and buy a Squeezebox or Sonos for the audio rack. I have briefly auditioned CD Players from Denon, Nad, Creek, CA and ((entry level ) Marantz with dealers and did not find them very impressive. I found the SQ in all these auditions to be 'muddy', 'noisy' and unconvincing. In most of these auditions, the other components in the mix were also not up to the mark. With better amp/speakers/cables and placement, some of these sources may have fared better. In the midfi segment, Arcam, Naim and Rega cdp's are good options. And if one can afford an Esoteric,Accuphase,Meridian or Ayre, then there would be no looking back!

A final step would be, to buy interconnects and speaker cables, which would do justice to the rest of the components. About power cords, I am not sure. Whether they are a one meter highway leading to audio nirvana or an unpaved road drawing one into a dark forest of confusion.
 
Last edited:
A beautiful, well-constructed speaker with class-leading soundstage, imaging and bass that is fast, deep, and precise.
Back
Top