[rkt31]

let's be clear. hqplayer is only a processor while dave is much more than a processor. the hardware which converts the processed stream to analog is equally important which in dave is very unique. I do not agree to the claim that hqplayer needs a FPGA for heavy processing capabilities of it. may be the people converting pcm to dsd512 that too in multichannel in hqplayer might need more processing power but that is not applicable for dave as it is 2 channel device and imho will not benefit from pcm to dsd conversion ( I have not experienced but imho people converting pcm to dsd via hqplayer might have penchant for certain flavor of sound) I have lenovo laptop with i7 4th gen processor which only uses 7% processing power while upsampling 16 bit 44.1khz to 24 bit 192khz using sinc filter( sinc uses most processing power) so for a 2 channel pcm to pcm routine even a normal laptop is sufficient indicating that hqplayer is just another resampler with more options. having said that I found hqplayer resampler much better than j river, foobar etc. coming to dave, imho it has much more sophisticated upsampling and noise filtering which can't be matched by hqplayer as there are many factors. hqplayer has the limitation of highest sampling rate, number of taps and accuracy of volume control as compare to dave. dave upsamples and filters at much higher rate and volume control is also much more accurate as it is embedded in the processing of dave ( more can be told by Rob). so the kind of processing dave has , hqplayer currently does not have .as mentioned earlier, before conversion to analog ,the digital stream has much more resolution and data rate in dave which is in mhz and might not be supported by the current pc hardware . also even if hqplayer author redesigns the algorithm, this processing in mhz can't be streamed in real time due to limitations of parallel processing with current processors in pcs .

 

[iDesign]

Rob Watts, I have a question for you. I have read here that you use JRiver for playback. I currently use iTunes as player using macbook and osx and all my music is in apple lossless format. Given that iTunes provides bit perfect playback for my music do you think its possible to improve sound quality by trying another player such as JRiver or Audirvana? I find it puzzling that Audirvana has two bit perfect integer modes that it claims sound different - mode 1 more transparent with better soundstage, and mode 2 that is warmer. How can this be if they are both bit perfect? Thanks.

Audirvanna and Roon will provide improved sound quality. I much prefer Audirvanna for a number of reasons on a Macintosh (specifically its Audio Unit, AFP, iTunes support etc. etc.) 

 

[halloweenman]

Rob Watts, thanks for your reply. Do you agree that if two software players are both playing back bit perfect from the same laptop/setup that both can actually sound different? Or should they sound the same as they are both bit perfect? Thanks.

 

[halloweenman]

idesign, did you do any ab blind testing?

 

[halloweenman]

currawong, thanks, yes those are my settings for itunes osx bit perfect playback. what i am really getting at is - does the things audirvana claims to do actually make a difference to sound quality if bit perfect? for example, will loading entire song into memory actually make any difference to sq? has anyone done ab blind testing itunes vs audurvana? i can appreciate the benefit of the software switching sample rates.

how can audirvana playback two different versions that sound different (integer mode 1 and 2) of the same bit perfect song?

 

[Rob Watts]

Rob Watts, thanks for your reply. Do you agree that if two software players are both playing back bit perfect from the same laptop/setup that both can actually sound different? Or should they sound the same as they are both bit perfect? Thanks.

I think it depends upon the DAC, the OS, and the programs.
 
So if the DAC is not galvanically isolated, then RF noise can leach into the DAC, and hence make it sound harder; the RF noise levels is affected by processor activity, so an inefficient program could sound worse than an efficient one; in principle at least.
 
A second point is that it depends upon the OS - Windows drivers that come with Chord DAC's ensure that it is bit perfect at the DAC with Windows - so if the DAC gets a packet error, data is resent; but driverless OS (iOS, Linux, Andriod) do not have this virtue. So if processor activity is such that time is not available to transmit on the USB, then you will get lost data - but Windows guarantees that this does not happen. That's the reason that with DoP DSD you need a very capable processor otherwise you hear occasional mutes; but with Windows I have never experienced this even at DSD515. So again this is how an efficient program could be better; but not with Windows.
 
With Dave, and using a battery powered non grounded Windows lap-top, where all the evidence I have is that it is completely and perfectly isolated from the lap-top RF noise wise, and is guaranteed to be bit perfect at the DAC, I would be very surprised if there were differences on any capable bit perfect program.
 
Rob

 

[Mark UX]

  I think it depends upon the DAC, the OS, and the programs.
 
So if the DAC is not galvanically isolated, then RF noise can leach into the DAC, and hence make it sound harder; the RF noise levels is affected by processor activity, so an inefficient program could sound worse than an efficient one; in principle at least.
 
A second point is that it depends upon the OS - Windows drivers that come with Chord DAC's ensure that it is bit perfect at the DAC with Windows - so if the DAC gets a packet error, data is resent; but driverless OS (iOS, Linux, Andriod) do not have this virtue. So if processor activity is such that time is not available to transmit on the USB, then you will get lost data - but Windows guarantees that this does not happen. That's the reason that with DoP DSD you need a very capable processor otherwise you hear occasional mutes; but with Windows I have never experienced this even at DSD515. So again this is how an efficient program could be better; but not with Windows.
 
With Dave, and using a battery powered non grounded Windows lap-top, where all the evidence I have is that it is completely and perfectly isolated from the lap-top RF noise wise, and is guaranteed to be bit perfect at the DAC, I would be very surprised if there were differences on any capable bit perfect program.
 
Rob

 
"With Dave, and using a battery powered non grounded Windows lap-top, where all the evidence I have is that it is completely and perfectly isolated from the lap-top RF noise wise, and is guaranteed to be bit perfect at the DAC, I would be very surprised if there were differences on any capable bit perfect program."
 
@Rob Watts, above statement also applies to 2 Qute? (which is also galvanically isolated)

 

[Rob Watts]

Yes sure - 2 Qute, Hugo TT and Dave have the same USB decoder and galvanic isolation.
 
Rob

 

[halloweenman]

thank you Rob, much appreciated.

 

[gnomen]

   
If all your music is the same bit depth and sample rate, and that is what is set in Audio Midi Set-up for output to the DAVE, then it will be bit perfect. If you have a mix of high res files, then it wont, because OSX will re-sample to whatever Audio Midi Set-up was set to at the start. 
 
One advantage of the audiophile players is that they match the output to the DAC to the file, so if you go from playing a 16/44.1 to a 24/192 file then the program will switch the output. As I understand it, the sonic advantage they are supposed to have involves taking complete control of the USB output to the DAC, preventing the system using it. That reduces the amount of unnecessary interruptions and processing done by the USB sending and receiving chips, and the amount of noise those circuits generate and dump into the electronics. Audirvana has a white paper on their site about it from memory.

Currawong makes an important point.  On a Mac Audio Midi will resample the files unless you are using an audiophile player.  I don't know about other players, but Pure Music does this, and even has a 'hog' mode to ensure the hifi signal is kept separate from general computer sound output.

 

[christian u]

I concur, those downloads are worth a listen. The sound stage is incredible. This is quality above quantity. And the albums have been mixed on HD800 and AKG 702 wich explains why they sound so good on cans.
 
Quote:

  For those interested in high quality recordings, Dutch outfit Sound Liaison http://www.soundliaison.com/ are having a summer sale.  All albums at all HD levels are reduced to €10.  Worth a look if you do not know them.  Great jazz, blues and world music ensemble performances.  Enjoy!

 
 

 
Am I doing something wrong ?
Or are there only 15 titles to choose from?
From  eClassical  I downloaded three  excellently played and realistically recorded Sibelius symphonies, all in all 80 minutes of glorious music, for 12 dollars and 38 cents,and it is one of my favourite go to sites for well recorded classical music.
The 3rd, 6th and 7th symphonies  by Sibelius as recorded by BIS in both 24/96 stereo and mch for those who can play mch files, are imho exactly the kind of music and recordings needed to sort "the wheat from the chaff",when it comes to both DACs and the rest of a good HIFI system.
This BIS album  is so far  my best bargain of 2016 closely followed by the 20 dollars I paid for Shostakovich´s 5th, 8th and 9th symphonies from the BSO.
I also see that there are some DAVEs arriving on the second hand market. But still NOT at prices that would make me bite just yet.
For me DAVE is still a very  overpriced product.
By the way is there anybody here who knows when the Brits are formally leaving the EU?
British pricing and 20% vat return on the price paid, might be just the thing that could tip it for me.
All provided my next  DAVE audition via the new ML 15 A electrostatic speakers too and not only headphones, rings my bell, on really well recorded classical albums.
Personally I couldn´t care less how EDM or any  other Electronica is reproduced via ANY DAC.

 

[Jawed]

When you are doing a doing a FIR filter (a tap) you need to read from memory the audio data; read from memory the coefficient data; multiply the numbers together;then read the accumulated data and add that to the previous multiplication; then save the result. Lots of things to do in sequence. With an FPGA you can do all of these things in parallel at once, so a single FIR tap can be accomplished within a single clock cycle (obviously pipelined) - you are not forced to do things in sequence.

This discussion has sent me rummaging and I came across: Efficient Convolution without Input/Output Delay

A motivation for this technique is room simulation which can involve many seconds worth of impulse response, resulting in a convolution using 10s or 100s of thousands taps.

Direct multiply-accumulate convolution is extremely expensive in terms of mathematical operations. It is very low in latency.

Ordinary FFT/inverse-FFT based techniques result in a huge delay between the input and output. The delay is reduced by working in blocks. More and smaller blocks result in more mathematical operations.

The paper presents a hybrid of multiply-accumulate with block-based FFT processing to substantially reduce the mathematical cost, while at the same time keeping a reasonable delay. Fig 6 is quite an eye-opener.

Is this the technique you've implemented for convolution, Rob? Or is this technique too heavy on internal bandwidth/memory to be applicable in an FPGA?

 

[iDesign]

idesign, did you do any ab blind testing?

Yes, the improvement over CoreAudio is profound. Damien Plisson has been great and very responsive in helping me configure Audirvana as the hub for my audio library which has tens of thousands of albums that were initially encoded several years ago using ALAC (I have one of the largest private collections of rare and out of print classical music in the US). Roon is a fine choice as well but it does not support AudioUnit plugins for an EQ and its options to access music files from a network server was not satisfactory for my needs. 

 

[halloweenman]

thanks idesign. i downloaded and tried audirvana and did a little testing using a Naim Dac V1 (my main system uses hugo tt) i managed to get hold of as this has very useful bit perfect testing technology licensed from Audiophilleo. i did a quick listening test and thought audirvana had better separation and soundstaging than itunes. i then did the bit perfect test. in a nutshell itunes passed when os x dac v1 midi setup sample and bit rate were the same as the file being played back in itunes. however it failed when os x dac v1 midi setup sample and bitbrate were different. so verifies comments made. audirvana passed all bit perfect tests regardless of os x dac v1 midi setup. what was interesting was that audirvana was bit perfect for both integer mode 1 and 2 even though they are supposed to sound different. mode 1 more transparent with better soundstaging, mode 2 warmer. how can this be?!

this testing has given me a problem in using itunes with my hugo tt playing back apple lossless as it will never be bit perfect to the tt. the os x tt dac midi setup bit rate for tt is 32 and cannot be changed. the bit rate for my apple lossless files is 16. itunes/os x converts these somehow to 32 bit before sending to dac - thus tt dac no longer gets a bit perfect file. i dont believe audirvava does this and sends 16 bit file to tt dac.

audirvana is excellent. stable, runs like clockwork, easy to use, and sounds superb and does what it says.

 

[Mavwong]

Thanks, romaz, for that very welcome reassurance. Now to get my hands one. Cheers

 
Not sure this has been covered as I just read till page 243. You can actually get a polarity corrected UK to US converter if that matters.