data is the same, but, playback sound has a little bit different.
we use prism sound ada-8xr + mm27
at the beginning of co-operation, one of our OEM customers told us local playback is much more clear than network play (file store in UPNP server), different class network cable will have different sound, that explains why Linn's player's playback quality is not as good as their own obsolete CD player.
also, playback route causes different sound, shorter will be better, so we have an invonation to make the shortest playback route and use FPGA as the output timing control. that will suppose to solve all above affection in theory, but who knows the real result cause different sector in SSD will have different sound...
I do have some articles in Chinese that tells why there will be different sound by pc playback... but no time to translate into english, mainly cause by memory latency, that's deep involve hardware, a little hard to explain in some words only.
BTW, our machine will be available soon after we tested a year long.
Last edited:
Hmmm I do struggle with this concept, that data from different sources that is bit identical when passed through all tghe pathways of the system sounds different, it goes against all digital engineering principles. Where does the difference come from, if you present the same (exact bit match) data to the SPDIF out or the DAC from any digital storage source, be it solid state or a hard drive, it is exactly the same data and cannot sound different, it is impossible. You are presenting the same bit pattern to the DAC so will get the same analogue stream out (let us ignore system noise as we are discussing just the integrity of digital data). The same reason why I can store photos on Flickr, my hard drive a SSD, when accessed they are exactly the same, the bit pattern is the same and they look the same.
Agian with the cable, sorry its digital, if the data is getting from point a to point b unchanged there is no difference, if there is then the bit pattern will be different, if the bit pattern is the same the sound will be the same, this is true audiophile myth territory.
Audio digital signals are trated like any other digital data within the system, and the systems are very good at moving data from a to b uncorrupted, I have yet to see any proof otherwise.
Agian with the cable, sorry its digital, if the data is getting from point a to point b unchanged there is no difference, if there is then the bit pattern will be different, if the bit pattern is the same the sound will be the same, this is true audiophile myth territory.
Audio digital signals are trated like any other digital data within the system, and the systems are very good at moving data from a to b uncorrupted, I have yet to see any proof otherwise.
I've seen the proof, with digital images, and heard the proof with digital audio files. (pun intended)
Check-sums the same!
Looks different and sounds different.
Verified by multiple independent people, on various systems.
No person taking part had any commercial interest in the matter.
Sorry Marce, as far as Im concerned, its been verified lock stock and barrel.
Its not a matter of theory. We all know the digital theory. In fact, no one is disputing the theory.
The matter of contention is in fact, that despite the bits being bits, and the check-sums being the same - digital images and audio can look different!
The evidence may be presented by the research group one day. But in the middle of another persons thread is not the time and place.
Depends on how fragmented you disc is, it could be spread over multiple sectors or on contigous ones, because of the speed of transfer even of traditional hard discs this is of no consequence. But the DOS controls all that and fetches the data, with no problems especially when you consider the size of an audio track, it is loaded into the buffer and passed down the line to the next stage in it journey from long term storage to eventually the dac. To the DOS data is data, be it a picture an audio file or a word document.
This would only be true if one cable was corrupting the data, as data transfer uses check sums data packets etc to guarantee data integrity again this is an audiophile myth, I have heard some say it is low level digital distortions, sorry but that is a load of ...the data is either intact or it isn't, you cant get a zero and a tiny bit otherwise we would be back to analogue.
Can be percieved as different but actually are not different, where have I missed the point. We should not postulate such theories. Saying that a picture (or a track) stored on a hard drive will look sound different from a picture (or track) stored on a SSD drive when the bit pattern is exactly the same is wrong.
Martin Colloms and others, have also verified that .bmp images saved from the same frame (00.34.00) of the same .ts file stored at different locations
(Christina Aguilera-David Letterman Show) using Corel ULVS12 can look a little different, and the viewer's preference changes when the location of the side by side images created using Photoshop is reversed.
The difference MAY possibly be due to something like low level wideband noise ?
(Christina Aguilera-David Letterman Show) using Corel ULVS12 can look a little different, and the viewer's preference changes when the location of the side by side images created using Photoshop is reversed.
The difference MAY possibly be due to something like low level wideband noise ?
What is low level wideband noise in relation to a digital file, is it stored with the data.
A digital file is a digital file and if they match bitwise they are exactly the same, no difference what so ever. If just one bit is different then they are different, but if bitwise the files match they are the same.
A digital file is a digital file and if they match bitwise they are exactly the same, no difference what so ever. If just one bit is different then they are different, but if bitwise the files match they are the same.
You mean the guy who rates wire to two hundred different incremental levels?
How exactly was this "verified"?
Stuart
In the case of the comparison .wav files it was verified over 6 separate listening sessions with different participants, over a period of several months using Blind A/B/A 3 minute sessions . The track used mainly was "Dire Straits_ Private Investigations, " which M.C. was thoroughly familiar with, having heard the original master.The results were published in HiFi Critic Vol.6 No.1, but the various HFC forum threads were more in depth than the article.
M.C. only sighted the .bmp images using a Samsung 23" monitor.
Alex
let me try to explained more, audio playback system is kinda complex, hard to explain in just softwar or hardware. every digital signal will be turned to analog to distinguish it's 0 or 1, so, when you use oscilloscope to monitor the eye diagram, you will find the difference between playback, that's why cable causes the difference.
different pcb layout will have different sound, we must accept then we could think then we could setup the model.
let me tell more about why different sector cause the difference, because different data path(route) will have different access time, and MMU in the system will enlarge the latency, then affects the jitter...
it took us 5 years to setup the model, now we do find a way out to eliminate those differenc. but it takes time to make it a finished product.
current model we proved digital output is better than CD-PRO, don't know how the latest model sounds, looking forward to listening that in the future. but latest model could playback no limited for the sample rate(DAC has), support DSD output, and so many functions....
different pcb layout will have different sound, we must accept then we could think then we could setup the model.
let me tell more about why different sector cause the difference, because different data path(route) will have different access time, and MMU in the system will enlarge the latency, then affects the jitter...
it took us 5 years to setup the model, now we do find a way out to eliminate those differenc. but it takes time to make it a finished product.
current model we proved digital output is better than CD-PRO, don't know how the latest model sounds, looking forward to listening that in the future. but latest model could playback no limited for the sample rate(DAC has), support DSD output, and so many functions....
Last edited:
we are discussing data route, that's the very low level issue, none of the business of the theory.
playback is not hard in theory but wide corss field, including semiconductor, IC design, PCB layout, schematic design and software control, no audio theory will discuss MMU in CPU, right? different compiler will make the program has different playback sound, wierd?
5 years ago, we use S/PDIF to record the data from PC sound card to verify the bit accuracy, no matter coax or fiber will have the same result, but fiber sounds bad, cause physical signal carrier like 48M were added into the output laser, so the PLL will work in a worse envirornment, this will not be explained in audio theory.
Interesting Topic!
I'm on a quest to build my first class D amp. I've scoured the site to figure out the best amp board. I don't want to get more involved than I can complete in a few days, so no board level soldering for my first. First people were in love with SI, then modded them, then went to other Tripath models like Hifimediy. Now I see people like Other chips by Philips and TI. I don't know what to do.
My home system has SET monoblocks that sound transcendent if the moon is right and you sacrifice a small animal. To oscilloscopes they look broken, which makes them happy in a malevolent way.
Now much of my work is designing computer networks and increasingly, servicing them, ferreting out niggly problems that can't happen per the manufacturers, but are.
Yes, I can't believe that bit for bit files are any different wherever they're stored. It's possible there's a transfer timing, jitter, missing data, or emi/rfi leaking into the analog circuit issue, but I do come down on the side of motivated cognition. Believing is hearing. My DAC is toslink fed, BTW, to eliminate the electrical noise possibility.
So, I'll pay my money and take my chance with a class D board, and if you think some digital cleaner gear is worth it, you remind me of the people that thought edging CDs with green magic marker improved the sound.
I'm on a quest to build my first class D amp. I've scoured the site to figure out the best amp board. I don't want to get more involved than I can complete in a few days, so no board level soldering for my first. First people were in love with SI, then modded them, then went to other Tripath models like Hifimediy. Now I see people like Other chips by Philips and TI. I don't know what to do.
My home system has SET monoblocks that sound transcendent if the moon is right and you sacrifice a small animal. To oscilloscopes they look broken, which makes them happy in a malevolent way.
Now much of my work is designing computer networks and increasingly, servicing them, ferreting out niggly problems that can't happen per the manufacturers, but are.
Yes, I can't believe that bit for bit files are any different wherever they're stored. It's possible there's a transfer timing, jitter, missing data, or emi/rfi leaking into the analog circuit issue, but I do come down on the side of motivated cognition. Believing is hearing. My DAC is toslink fed, BTW, to eliminate the electrical noise possibility.
So, I'll pay my money and take my chance with a class D board, and if you think some digital cleaner gear is worth it, you remind me of the people that thought edging CDs with green magic marker improved the sound.
This nonsense seems to crop up every so often. If the data is the same, and the timing is the same, then if the sound is genuinely different there are only two possible explanations:
1. something else is different (e.g. PSU noise, mains voltage)
2. there really are little gremlins who live in data files who can not only corrupt data but do it in a way which preserves checksums
Which explanation do you think is most likely?
When someone argues that 2=3 then I have to ask whether they understand what '2', '=' and '3' mean!
1. something else is different (e.g. PSU noise, mains voltage)
2. there really are little gremlins who live in data files who can not only corrupt data but do it in a way which preserves checksums
Which explanation do you think is most likely?
When someone argues that 2=3 then I have to ask whether they understand what '2', '=' and '3' mean!
I have to agree with Marce.
If we apply the same principles to a binary file that's a critical part of the operating system and doing something like copying the file and renaming it (or just running disk defragmenting and having this result in a re-arranging of the data chunks in the harddisk) could then this result in a completely broken computer since the executable binary could have changed and not be doing what it is supposed to do (like screw up the memory mapping or context switching or disk drive driver)?
Does this mean that once I rip my cds and back them up or copy them (from one location to another, or another partition or over the network) that the sound of my ripped files will be changing RANDOMLY and at worst case they will never be the same? This is without even playing them yet, just as a result of changing data media or sectors in the media?
I could understand this in analogue media, like a record's groove's being deformed with each playback as the needle traces the groove. But 20+ years of working with computers make it very difficult for me to accept that copying binary files around changes them. My computers would have all been broken years ago.
It's even harder to believe if this inconsistency is selective and only affects sound and image files! Marce is correct in that to the operating system, these are just chunks of bits in memory or in media, they are all the same.
If we apply the same principles to a binary file that's a critical part of the operating system and doing something like copying the file and renaming it (or just running disk defragmenting and having this result in a re-arranging of the data chunks in the harddisk) could then this result in a completely broken computer since the executable binary could have changed and not be doing what it is supposed to do (like screw up the memory mapping or context switching or disk drive driver)?
Does this mean that once I rip my cds and back them up or copy them (from one location to another, or another partition or over the network) that the sound of my ripped files will be changing RANDOMLY and at worst case they will never be the same? This is without even playing them yet, just as a result of changing data media or sectors in the media?
I could understand this in analogue media, like a record's groove's being deformed with each playback as the needle traces the groove. But 20+ years of working with computers make it very difficult for me to accept that copying binary files around changes them. My computers would have all been broken years ago.
It's even harder to believe if this inconsistency is selective and only affects sound and image files! Marce is correct in that to the operating system, these are just chunks of bits in memory or in media, they are all the same.
Last edited:
********, I worked in pre-press/imaging for nearly a decade. I ran a large Hell-Linotron postscript rip (raster image procesor), imagesetter and drum scanner for about 5 years of that. if the bmp is created using the same source file exactly and created using the same colour profile, compression settings (or uncompressed) with the same colour gamma profile (not generated corrected for 2 different monitors, or printers) then it will be the same, every pixel will be the same.
I used to work on such images at >4800dpi and generate deep etched masks at the pixel level. these were stored on a network drive, sometimes I would work on a lower resolution proxy, sometimes I would work on the file directly over the network, yet it would be RIP'd/spooled from the server, which was remote from both locations.
if what you say was true, it would have been a total pita... I would spend hours on a high res composite, only to find errent pixels change colour and thus the masks, which would sometimes be generated on the file in a non-destructive way, would produce a different result from the 2 locations. Let me tell you, that never happened...ever...
its possible that the video file format they were pulling the frames from was mpeg with slightly different settings or colour targets, or the freeze frames generated from say a different frame rate proxy did not turn out the same. this is different from 2 copies of the same file looking different depending on the 'location' of the data.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.