Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Aug 13, 2013 2:29:59 GMT
Here's one for the "bits are bits" brigade. There has been a lot of recent discussion about software players that are "bit perfect" having to all sound the same. Archimago and his disciples ( regulars know of one who loves to quote him ) all insist that their tests show that all "bit perfect" players MUST sound the same. Preliminary results seem to indicate that around 78% of respondents can indeed hear differences ! Alex If anybody is interested , the thread in C.A.'s General Forum is "Do you hear a difference...? " P.S. Now 81% at time of editing.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Aug 13, 2013 3:36:41 GMT
Hi Alex,
What does "bit perfect" mean in this context?
Regards
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Aug 13, 2013 4:12:57 GMT
Hi Greg I feel sure that you have already seen this already archimago.blogspot.com.au/2013/05/measurements-bit-perfect-audiophile.htmlA couple of senior members in C.A. didn't agree with his findings, especially after a new commercial Mac software player revision. ( Audirvana 1.5.5 ) This new Poll thread was a follow on from that discussion. Perhaps we should invite J.K. to reply to some of your questions ? Kind Regards Alex
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Aug 13, 2013 12:28:10 GMT
Hi Alex, What does "bit perfect" mean in this context? Regards It means it's "a bit" perfect but not totally ahem, where's my coat...
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Aug 13, 2013 22:45:20 GMT
The C.A. poll is presently running at 82.5% hearing differences. It seems that the "bits are bits" brigade are so desperate to prove that the differences are all in people's imaginations, that they still defend the validity of the reports from Archimago. They just don't get it, that the moment you do an A/D conversion to 24/96 on the assumption that 24/96 is completely transparent,in order to obtain measurements , when others like Barry Diament say that the output of his professional A/D converter doesn't sound like the direct microphone feed until he uses 24/192, that they have damaged their credibility. If that's not enough, more and more people are now enjoying much higher resolution DSD than even 24/192 . It's a whole new world out there with many new DSD compatible DACs from various manufacturers, and extremely high resolution DSD material available from Blue Coast Records, as well as other specialised Recording labels. The moment that Archimago introduced that 24/96 conversion in an attempt to assert his beliefs, it no longer was Scientifically acceptable, no matter what many of his disciples may wish to believe. Those findings are no more Scientifically acceptable proof than anecdotal reports,
Alex
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Aug 14, 2013 0:22:13 GMT
Carrying out this poll at CA is like going into a church and asking "Who believes in God'?
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Aug 14, 2013 0:30:38 GMT
Hi Alex, What does "bit perfect" mean in this context? Regards It means it's "a bit" perfect but not totally ahem, where's my coat... I have been reading all these digital posts with the wrong interpretation. Now it all makes sense. It's a bit like the atom they keep on splitting it into smaller and smaller components. One day those digital engineers will split the bit and find out what is inside.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Aug 14, 2013 0:49:51 GMT
Carrying out this poll at CA is like going into a church and asking "Who believes in God'? Hi Greg Having a similar poll in DIY Audio for example, would likely result in 99.5% saying they hear no differences, because of peer pressure. BTW, the person who organised the poll goes to the same church as you do. The one where they hang on every word from Archimago. This poll is an offshoot she started after those reports about Audirvana 1.5.5 sounding better than previous releases, despite still being "bit perfect", which will hopefully lead to statistically relevant trials, provided that enough people can be persuaded to use their ears for a change. If these guys were right, then our S.C. DACs should sound as uninspiring as the original version from Nicholas Vinen. Haven't you ever wondered why PSU changes mainly to the digital area can give further improvements in SQ ? .Has it altered the binary data ? If so, then the original design, and the manufacturers data sheets that these designs are based on, have major problems ! Kind Regards Alex
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Aug 14, 2013 0:53:33 GMT
It means it's "a bit" perfect but not totally ahem, where's my coat... ..... One day those digital engineers will split the bit and find out what is inside. That would be a marked improvement over "spitting the dummy" when people report hearing things that they can't explain!
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Aug 14, 2013 6:08:08 GMT
If these guys were right, then our S.C. DACs should sound as uninspiring as the original version from Nicholas Vinen. Haven't you ever wondered why PSU changes mainly to the digital area can give further improvements in SQ ? .Has it altered the binary data ? If so, then the original design, and the manufacturers data sheets that these designs are based on, have major problems ! Hi Alex, I think you might be taking things out of context. regards
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Aug 14, 2013 6:31:45 GMT
If these guys were right, then our S.C. DACs should sound as uninspiring as the original version from Nicholas Vinen. Haven't you ever wondered why PSU changes mainly to the digital area can give further improvements in SQ ? .Has it altered the binary data ? If so, then the original design, and the manufacturers data sheets that these designs are based on, have major problems ! Hi Alex, I think you might be taking things out of context. regards Hi Greg I I am just trying to illustrate the point that there is far more to digital than just the right sequence of 1s and 0s. For example, someone in C.A. just mentioned about running a 2.5": SSD off 5V only from the +12V and +5V molex. I pointed out that if the person was into DIY he could use a 7805 with a small heatsink, a capacitor either side of the regulator, and run it off 12V instead. He is VERY likely to get improved results. Kind Regards Alex
|
|
jkeny
Been here a while!
Posts: 463
|
Post by jkeny on Aug 15, 2013 13:53:51 GMT
Greg, If you are truly live by & believe Archimago's measurements then explain how he gets a -90dB null using his Dell's inbuilt soundcard's ADC which he states has a noise floor of -86dB? Do you not find an anomaly in this?
BTW, here's a good one from the bit-is-bits camp - JRiver, one of the stallwarts of this view have this in heir recent release on memory playback:
So here is a company introducing a feature which they "know" has no benefit but has possible/probable technical issues:
I call that snake-oil selling - selling products with features that the producer "knows" to have no effect.
|
|
|
Post by freddypipsqueek on Aug 17, 2013 19:38:50 GMT
Read this with a bit of interest.
Alex, I note you say, "there is far more to digital than just the right sequence of 1s and 0s"
It might be a question of interpretation, but I think digital is precisely just the right sequence of 1s and 0s or to quote the dictionary, "data that is represented using discrete (discontinuous) values"
Surely anything else is not digital as such - I;m not saying digital is perfect but there has to be some assumption, as some stage, that the raw data is correct - otherwise cars stop, missiles launch themselves and this post will be unreadable !!.
Adrian.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Aug 17, 2013 21:08:19 GMT
Read this with a bit of interest. Alex, I note you say, "there is far more to digital than just the right sequence of 1s and 0s" It might be a question of interpretation, but I think digital is precisely just the right sequence of 1s and 0s or to quote the dictionary, "data that is represented using discrete (discontinuous) values" Surely anything else is not digital as such - I;m not saying digital is perfect but there has to be some assumption, as some stage, that the raw data is correct - otherwise cars stop, missiles launch themselves and this post will be unreadable !!. Adrian. Adrian Timing is not as critical with normal programs as it is with good audio and video . Latest research discussed in a C.A. thread shows that a respected designer (John Swenson) has seen noise riding on top of the waveform going to the DAC. The absolute timing, including the shape of the analogue waveform representing the digital signal, as well as the correct spacing (Jitter) does matter for both Audio and Video. As the resolution goes up, so do these requirements.. Later this morning I will see if I can provide a link to the original discussion. BTW, This is the most recent from Martin Colloms in HFC Forum www.hificritic.com/Forum/yaf_postst1373p5_Problem-when-hocus-pocus-works.aspx?= Most recent.See replies 86 to 91 Regards Alex
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Aug 18, 2013 1:39:21 GMT
|
|
|
Post by freddypipsqueek on Aug 18, 2013 11:43:51 GMT
I think I understand what is being said though I have skim read the thread.
As I understand it (and I agree with this) different software may playback differently. Correct me if I am wrong but the comments all imply that this is due to the characteristics of the computer or software and the various factors which influence the sound. To me software which is "bit perfect" is like a car which is "the ultimate driving machine"; its become sale hype.
My point though was to pick up on your statement that "there is far more to digital than just the right sequence of 1s and 0s"
I might have misunderstood it but digital is 1s & 0s. I don't think there is any more to it than that. There may be far more to the playback of digital than 1s & 0s but at that point the variance cease to be in the digital domain.
I think the problem is that by using the label 'digital' incorrectly an unnecessary difference of opinion is created.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Aug 18, 2013 12:09:13 GMT
Adrian In this discussion , they are talking about how different "bit perfect" player software affects the DAC. What is going into the DAC is an ANALOGUE representation of 1s and 0s with less than precise shape and spacing between elements due to :"Jitter." Riding along with it also is noise, as John Swenson mentioned. How it sounds will depend mainly on how good the DAC is at interpreting the waveform correctly, and rejecting rubbish.. Devices like Schmidt triggers can help restore the waveform, such as in coax SPDIF inputs , but "Jitter" in the signal, as well as less than perfect PSUs will see slightly different trigger points in the Schmidt trigger.. (See Hysterisis) Have you ever looked at the stored waveform on a HDD, for example ? ( try Google Images) There is a lot of cleaning up required. This is done in the HDD itself. Perhaps that's part of the reason why some HDDs are reported to sound better than others, just as some SSDs can sound a little different to HDDs and even each other. ? Regards Alex
P.S. How can this waveform be a perfect sequence of text book looking 1s and 0s ,when it more than likely goes through a transformer (coax SPDIF) in better than average gear, or just an electrolytic capacitor and a s resistor perhaps, as well as another transformer or capacitor at the DAC end ? Add to that transmission lines with less than optimum impedance,(reflections) RCA sockets in most cases that certainly aren't 75 ohm impedance etc. Other transmission methods will have their own set of problems too.
|
|
jkeny
Been here a while!
Posts: 463
|
Post by jkeny on Aug 20, 2013 17:10:54 GMT
Yes, Alex, the problem with using the word "digital" is that people don't really understand what "digital" is.
It's just an agreed protocol nothing magical about it. The protocol is simply that signals over a certain voltage level are considered to be "1" & below a certain voltage level to be a "0". So there is a range of voltages that all represent a "0" - let's say any voltage from 0V to 1.2V is interpreted as "0" & a range that represents a "1" say any voltage from 2.1 to 3.3V. The advantage of this protocol is that it is relatively immune to noise - in other words it would take a lot of noise (0.9V or more) to flip a "0" into a "1". So noise is generally not of much concern in digital communication when that communication remains in the digital domain.
However, digital audio has one foot in the digital domain & the other foot in the analogue domain (digital to analogue conversion). Noise is of concern in analogue signals. If there is noise riding on the digital signal (which has no effect on the correct interpretation of the digital bits) - it can still have an effect on the analogue side of the DAC. This noise is very difficult to completely & effectively block.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Aug 21, 2013 0:33:52 GMT
Add to that the word digital was very much a buzz-word for years, digital headphones ( does that mean we had a DAC in our ears? ), digital this or that, digital ready. No wonder many folk get confused about what digital actually is, though folk qualified in the field should know better.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Aug 21, 2013 3:16:29 GMT
Hi jkeny,
A pretty good description of binary digital. I would have to agree with every word. Most people seem to think digital means binary digital now days. Strictly speaking digital can have any number of discrete states.
I think the main confusion is a lot of people are applying the issues with audio digital to areas where it doesn't apply, like computer memory, cache, storage media, ethernet etc.
Regards
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Aug 21, 2013 4:20:01 GMT
|
|
jkeny
Been here a while!
Posts: 463
|
Post by jkeny on Aug 21, 2013 12:23:19 GMT
Yes, ground is probably a term that is even more confused than digital. It has the unfortunate name which leads one to assume that it is always at 0V. It also has the dual role mentioned in the article - that of being a reference against which all the signals are compared & also it is a return path for analogue signal currents/digital signal currents/PS currents. This mixed role is not such a happy one & careful consideration needs to be given to grounding. For me it has been one of the most rewarding areas to delve into.
|
|
jkeny
Been here a while!
Posts: 463
|
Post by jkeny on Aug 21, 2013 12:39:01 GMT
Hi jkeny, A pretty good description of binary digital. I would have to agree with every word. Most people seem to think digital means binary digital now days. Strictly speaking digital can have any number of discrete states. No, Greg that is incorrect or at least reads incorrectly. Digital is binary & can only have two "interpreted" states. In other words, as per what I said & what John S said, any voltage below the agreed threshold is "interpreted" as a "0" bit. This means that there are a range of voltages all below the threshold that represent a "0". It's the difference between the physical representation of the "digital" signal & it's interpretation as digital. Your statement above is confusing! Again, I'm not sure I'm with you here - you seem to be talking about Alex's experiments with stored digital? The simplistic explanation I gave of digital is only the beginning of what needs to be considered when we are talking about digital audio. For instance, both I & John Swenson have built DACs which isolate the PC from the analogue side of the DAC. This isolation was done using some of the best devices currently known NVE GMR isolators. In theory the designs completely isolated all signals, PS & ground planes between the digital side of the DAC & the analogue side. So according to all of the above explanations it should therefore be immune to anything on the PC side, right? Well wrong! The PC's PS, the playback software still influenced the sound. How did this happen if it's just digital signalling & all possible influences on the analogue side of the DAC have been removed? Maybe the isolators didn't isolate "fully" & some influences crept through below what can be measured? You tell me because neither John S nor I know why (at the moment)
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Aug 22, 2013 1:16:31 GMT
Hi jkeny, A pretty good description of binary digital. I would have to agree with every word. Most people seem to think digital means binary digital now days. Strictly speaking digital can have any number of discrete states. No, Greg that is incorrect or at least reads incorrectly. Digital is binary & can only have two "interpreted" states. In other words, as per what I said & what John S said, any voltage below the agreed threshold is "interpreted" as a "0" bit. This means that there are a range of voltages all below the threshold that represent a "0". It's the difference between the physical representation of the "digital" signal & it's interpretation as digital. Your statement above is confusing! Hi jkeny, Sorry for confusing you. Communication is not one of my strong points. Here's a reference that might help explain things. en.wikipedia.org/wiki/Digital_electronicsDigital electronics, or digital (electronic) circuits, represent signals by discrete bands of analog levels, rather than by a continuous range. All levels within a band represent the same signal state. Relatively small changes to the analog signal levels due to manufacturing tolerance, signal attenuation or parasitic noise do not leave the discrete envelope, and as a result are ignored by signal state sensing circuitry. This is the general definition of Digital (or it was when I was a kid before the proliferation of computers) In most cases the number of these states is two, and they are represented by two voltage bands: one near a reference value (typically termed as "ground" or zero volts) and a value near the supply voltage, corresponding to the "false" ("0") and "true" ("1") values of the Boolean domain respectively. This example of Binary or Boolean Digital is the definition most people now understand as Digital. No I was not referencing Alex's stuff. Alex and I usually discuss this stuff via email. Thanks for the your information. The issue I see with PCs is the almost infinite combinations of hardware available, then there is the software. Your might end up resolving an issue on your PC but that may or may not help everyone. What projects are you and John Swenson doing? I thought you mainly did USB DACs and I know John is doing a DAC daughter board for the Wandboard computer. So possibly completely different causes resulting in the same problem. At least your USB DAC isn't sitting on top of a 1GHz CPU. I supposed you have stuck your "device" into a Faraday cage to eliminate the possibility of an external source of noise. I have seen a case where a local power substation affected every computer monitor in a 2 storey building. Also, my microwave effects my wireless network. What does the signal look like immediately before and after the isolators using an oscilloscope? Anyway, you and John have far more experience in electronics design so the chances of me helping are very slim. regards
|
|