Nope checksums are 100%, not 99.99% or 99.9999999%
A digital file copied and verified is an exact clone, there are no differences, sorry, but that's how all this wonderful digital stuff works, there is no room for even 0.1% error, the data would likely be corrupted.
So no you are wrong, if the file copies correctly (which the operating system will ensure) then the two copies are identical, this works with CAD files, photos, documents and every other thing stored as digital data.
A digital file copied and verified is an exact clone, there are no differences, sorry, but that's how all this wonderful digital stuff works, there is no room for even 0.1% error, the data would likely be corrupted.
So no you are wrong, if the file copies correctly (which the operating system will ensure) then the two copies are identical, this works with CAD files, photos, documents and every other thing stored as digital data.
Wrong
Last edited:
Sorry to be rather harsh in my reply, but we are talking about identical bit map files sounding different. I know that no hardware is 100% reliable, but I have done both electronic CAD and mechanical CAD for 25+ years and these days it is a shock when a file gets corrupted or damaged during copying etc, whereas 15+ years ago it was much more common ( but even then not an everyday occurrence). Mechanical means of playback, CD DVD are flawed and can give rise to errors, though even these means these days are much more reliable than they use to be.
I believe a lot of peoples perceptions that digital is bad is a hangover from the early days, and probably extra noise from poorly integrated interfaces.
I believe a lot of peoples perceptions that digital is bad is a hangover from the early days, and probably extra noise from poorly integrated interfaces.
Hi
Just echoing what Pano said about SSD and HDD... datafiles/checksum maybe the same but playback may be different.
Has anyone here considered between distance of your hard drive and your speakers.
True that many people here have done some form of checksum experiment. Maybe we should repeat the experiment with the HDD placed on top of a subwoofer.
I remember someone in the hard drive industry commented, the relationship between the magnetic head and the platter is like flying a supersonic jet 1 meter above ground. Now of course if the ground is having an earthquake (maybe a REL, Pun intended)
I am not on either camps, I am just offering a different possibility on the cause the vibration caused by sound (come on guys, we are in Diyaudio, we are not going to listen to music at 60dB
) causes error in the reading of databits from the platter. We would not have observed it normally because subwoofers are genarally not allowed in office (at least in mine..)
Seriously though it might not be the sub that would tickle the HDD, it could higher frequency stuff as well as long as they hit the resonant frequency of the arm/platter/GMR head... who knows...
Oon
Just echoing what Pano said about SSD and HDD... datafiles/checksum maybe the same but playback may be different.
Has anyone here considered between distance of your hard drive and your speakers.
True that many people here have done some form of checksum experiment. Maybe we should repeat the experiment with the HDD placed on top of a subwoofer.
I remember someone in the hard drive industry commented, the relationship between the magnetic head and the platter is like flying a supersonic jet 1 meter above ground. Now of course if the ground is having an earthquake (maybe a REL, Pun intended)
I am not on either camps, I am just offering a different possibility on the cause the vibration caused by sound (come on guys, we are in Diyaudio, we are not going to listen to music at 60dB
) causes error in the reading of databits from the platter. We would not have observed it normally because subwoofers are genarally not allowed in office (at least in mine..)Seriously though it might not be the sub that would tickle the HDD, it could higher frequency stuff as well as long as they hit the resonant frequency of the arm/platter/GMR head... who knows...
Oon
My laptop works fine in moving cars, trains, and airplanes.
I'd be FAR more worried about HDD corruption by the woofer's magnetic field!
When someone shows actual EVIDENCE of any sonic differences between media, then it's worth a look. Otherwise, it's all handwaving (a very popular occupation among fashion audio aficionados).
There was a story a couple of years back about an IT installation's HDDs being upset by vibration. Not so much as to cause any data errors rather the drives' throughput slowed down noticeably due to suffering many more retries on seeks.
Inability to read the hard drive (in time to get the data through the buffering and to the DAC) will result in stuttering and other obvious problems, not the subtle differences that I recall were claimed earlier in the thread. Hard disk data that's been modified or changed in any way doesn't get through, very much unlike what a CD player does.
You SHOULD assume it's 100% perfect - right now - but your drive may fail tomorrow. THAT'S why you want a backup.
G²
Something would have to be VERY SERIOUSLY messed up for that to happen. I routinely run 2 HDTV recording events each of which needs 2.5 MBytes / second. Redbook audio needs under 180 KBytes.
Next question. Why do you all assume only the LSB might be damaged ? Statistically any bit _could_ get messed up. I can tell you that an incorrect MSB will be noticed by all. The bottom line is that corrupted data is very rare in a properly operating system.
Your CD player can reconstruct a limited number of errors per block of data and the corrections are perfect. If there are more errors it will interpolate and that will be very subtle. Too many errors and it will be not so good. That same disc used as a 'data carrier' can store 700 MBytes but as an 80 minute audio disc you will listen to 846 MByte of audio data. The 'lost' 146 MBytes are used for error correction. 'Interpolation' isn't terrible for audio but 1 bad bit in the right opcode can crash the program.
G²
Part2: Main difference is sound conducted through the power circuit V+ and 0V too.
There's definitely power circuit differences especially with amplifiers either running on same power circuit as the HDD and/or Flash or connected via the 0v line anyway, which is nothing other than a backwards way to say the same thing (except for highly regulated balanced drive and/or optic systems). USB is not an exception and is affected by and by amplifying power circuit differences, much the contrary to most marketing information saying that USB is external, which it is, but only in appearance and not really "external" for signal. Anything different in the power circuit will make a difference.
Power + Signal = Amplifier, and generally the power portion is larger and gets amplified more. This is significant grounds to claim that the differences in power circuit between HDD and Flash may be audible.
For example, I took 3 identical RockBoxed Iriver H10 players, changed one to old style flash, changed one to new fast flash and refreshed the hard drive on the third via low level; and, very thorough data integrity checks were performed on all three. Settings and operating system and all contents were identical contents and in identical load order.
They all sound different, and dramatically so. All three burst to ram cache and play from Ram. The one and only significant difference between all three is the power circuit located at the media and the different load on the power circuit as well.
I suppose that the power circuit accounts for the majority of the differences that you may hear.
There's definitely power circuit differences especially with amplifiers either running on same power circuit as the HDD and/or Flash or connected via the 0v line anyway, which is nothing other than a backwards way to say the same thing (except for highly regulated balanced drive and/or optic systems). USB is not an exception and is affected by and by amplifying power circuit differences, much the contrary to most marketing information saying that USB is external, which it is, but only in appearance and not really "external" for signal. Anything different in the power circuit will make a difference.
Power + Signal = Amplifier, and generally the power portion is larger and gets amplified more. This is significant grounds to claim that the differences in power circuit between HDD and Flash may be audible.
For example, I took 3 identical RockBoxed Iriver H10 players, changed one to old style flash, changed one to new fast flash and refreshed the hard drive on the third via low level; and, very thorough data integrity checks were performed on all three. Settings and operating system and all contents were identical contents and in identical load order.
They all sound different, and dramatically so. All three burst to ram cache and play from Ram. The one and only significant difference between all three is the power circuit located at the media and the different load on the power circuit as well.
I suppose that the power circuit accounts for the majority of the differences that you may hear.
Last edited:
Daniel, this was discussed in either this thread or another thread along similar lines, I will find links later. This is not the issue, the differences were said to be down to recordings. Power system noise is a known problem and if you suffer from it sort it out by galvonicly isolating the DAC and analogue from the digital playback system.
The original claim is that low level distortion is recorded with the digital data when the recording is made.
To bring power supplies into the equation is doubling the variables and thus renders the investigation null, as we know power supply quality is paramount to any analogue system, more than for digital systems that can tolerate much higher levels of noise.
The original claim is that low level distortion is recorded with the digital data when the recording is made.
To bring power supplies into the equation is doubling the variables and thus renders the investigation null, as we know power supply quality is paramount to any analogue system, more than for digital systems that can tolerate much higher levels of noise.
I thought digital systems were much worse in that regard, since power noise throws them into error correction mode reliance and then. . . it is a bit closer to a mistake.
Digital signals are much more tolerant of noise than analogue signals, any deviation of an analogue supply is going to effect the signal, whereas digital can take a lot before the bit pattern becomes corrupted. This is the I/O supply for digital signals, core supplies have quite strict requirements so local supplies and ferrites abound to keep processor (FPGA etc) cores clean and within tolerance. The I/O supplies are not primarily made dirty by the power supplies but by the actual physical switching of the gates, which often has instantaneous requirements into the 100 of amps, simultaneous switching noise. If the core voltages are effected and become to dirty and meander from there often very tight voltage range then your system will become unstable, hence a lot of care is taken when laying out PCB's to ensure the critical supplies are as clean as possible, have the correct impedance for the required switching (bypass/decoupling capacitors, size, placement and planes).
This is why you will never clean a PC supply up enough for really critical audio, and why I promote like many others (and use) galvanic isolation (wireless in my case). it dose not matter whether you use a linear or a SMPS as the main supply you will not really affect the noise signature, as a lot of the noise is generated by the system itself.
I often do boards with upwards of 20 separate supplies, for the various devices and usually a minimum of 8 layers and more often 12 to 14 to get enough ground planes to minimise the noise, compare that with an average PC where 6 layers is often the norm due to market pressure and cost restraints.
Just recently I have seen a few inards and boards of so called high end digital audio gear and in some cases I was shocked to see primitive layout with the maximum cost cutting restraints applied, and mid range stuff makes me weep.
This is why you will never clean a PC supply up enough for really critical audio, and why I promote like many others (and use) galvanic isolation (wireless in my case). it dose not matter whether you use a linear or a SMPS as the main supply you will not really affect the noise signature, as a lot of the noise is generated by the system itself.
I often do boards with upwards of 20 separate supplies, for the various devices and usually a minimum of 8 layers and more often 12 to 14 to get enough ground planes to minimise the noise, compare that with an average PC where 6 layers is often the norm due to market pressure and cost restraints.
Just recently I have seen a few inards and boards of so called high end digital audio gear and in some cases I was shocked to see primitive layout with the maximum cost cutting restraints applied, and mid range stuff makes me weep.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.