Recently a guy told me a fact that made me shocked.
"you have to use the internal HD to store your audio files on your PC or MAC before you send to the DAC, you will gain a lot of separation in the instruments, better sound floor noise..."
But for me, this is so nonsense that I want to share with you this audiophoolery.
There is a lot of people with this thought, and I wanna know how can a digital information 00110110 01010100 11101010 gets better or worse if it's from a stream, a usb external HD or an internal HD if everything goes first to the volatile memory (RAM) and only after will go to the DAC via coaxial or USB.
I use JRIVER on a MacbookPro and a Schiit Audio Gungnir via USB and 4 HD, 1 in with an airport extreme, and the 3 other in my Imac. All are connected to my macbook by wireless, the software plays files from the memory instead from the disc, is there any audio difference in adopting this system or copying every file to the mac internal HD?
"you have to use the internal HD to store your audio files on your PC or MAC before you send to the DAC, you will gain a lot of separation in the instruments, better sound floor noise..."
But for me, this is so nonsense that I want to share with you this audiophoolery.
There is a lot of people with this thought, and I wanna know how can a digital information 00110110 01010100 11101010 gets better or worse if it's from a stream, a usb external HD or an internal HD if everything goes first to the volatile memory (RAM) and only after will go to the DAC via coaxial or USB.
I use JRIVER on a MacbookPro and a Schiit Audio Gungnir via USB and 4 HD, 1 in with an airport extreme, and the 3 other in my Imac. All are connected to my macbook by wireless, the software plays files from the memory instead from the disc, is there any audio difference in adopting this system or copying every file to the mac internal HD?
Last edited:
unless you have a real problem with bandwidth over your wifi there should be no difference. Ask the 'guy' to see if he can come over and hear the difference in a blind test.
Good idea, but these guys runs from blind test as the devil runs when see the cross.
There isn't, but sometimes a difference can be heard like with the SqueezBox just because the layout which manage the datas after the reception buffer ! some difference between the wifi and the wire input layout !
So this is not the difference between air and wire transportation which can heard with tcpi/ip but sometimes the layout between the reception of the packets and the DAC chip !
As said, but a problem the buffer can not manage at the reception like something who avoid to receive the information : it would be the same.
If a difference is heard : could be e.g.: bad internal powersupply management for a chip which is not present in the other way of transport, different layout after the reception as the hard involved is not the same for the two way of transport on the reception pcb to the dac chip, etc, etc ! So between the IP reception and the DAC chip, jitter could happen and be different and generated after the reception on few cm on the pcb ! That's why with the SB Touch many seems to prefer the RJ45 instead the Wifi !
So this is not the difference between air and wire transportation which can heard with tcpi/ip but sometimes the layout between the reception of the packets and the DAC chip !
As said, but a problem the buffer can not manage at the reception like something who avoid to receive the information : it would be the same.
If a difference is heard : could be e.g.: bad internal powersupply management for a chip which is not present in the other way of transport, different layout after the reception as the hard involved is not the same for the two way of transport on the reception pcb to the dac chip, etc, etc ! So between the IP reception and the DAC chip, jitter could happen and be different and generated after the reception on few cm on the pcb ! That's why with the SB Touch many seems to prefer the RJ45 instead the Wifi !
Last edited:
except the link from laptop to dac is constant in this case and JRiver has a memory buffer. completely different case.
My understanding is we talk wifi and wire with IP packets to compare apple and apple.
I know myself any transporter which transport pcm or DSD via wifi ! Doest it exist. This one linked above ?
I know myself any transporter which transport pcm or DSD via wifi ! Doest it exist. This one linked above ?
(Data 10010101 01010101 01001001) External HD > Usb > Imac > Wireless 802.11 > Macbookpro > Jriver > RAM > Usb DAC (Data 10010101 01010101 01001001)
Ok. How can
(Data 10010101 01010101 01001001) Macbookpro internal HD > Jriver > RAM > Usb DAC (Data 10010101 01010101 01001001)
get's better or worse?
Nonsense.
Ok. How can
(Data 10010101 01010101 01001001) Macbookpro internal HD > Jriver > RAM > Usb DAC (Data 10010101 01010101 01001001)
get's better or worse?
Nonsense.
The same ! No difference before the Wireless as you seems to talk about TCP/IP.
Sometimes with massive copy via an USB port it can be not bit perfect, but with a song and the HDD buffer ?! I don't believe a difference can occur !
After Jriver : jitter became possible , but as they are the same in the two experiments and assuming an asynchronous USB, I will say again The Same !
Difference heard ! Wall AC fluctuation to the second PC between the two listening tests (try the internal battery of the laptop) ? Fluctuation of the DC PS of the USB streamer if it is an external device ! Ears not rested during five minutes between two listenings tests ?
Sometimes with massive copy via an USB port it can be not bit perfect, but with a song and the HDD buffer ?! I don't believe a difference can occur !
After Jriver : jitter became possible , but as they are the same in the two experiments and assuming an asynchronous USB, I will say again The Same !
Difference heard ! Wall AC fluctuation to the second PC between the two listening tests (try the internal battery of the laptop) ? Fluctuation of the DC PS of the USB streamer if it is an external device ! Ears not rested during five minutes between two listenings tests ?
Last edited:
If the guy runs from blind listening tests, can he at least provide some numbers to back up such a claim?
His idea suggests that an SSD would be still better than a spinning drive. And if you really want the best, you'll need to buy one of those RAM drives that are the same as your main system memory, just backed by a battery.
I do believe it is possible that a really lossy network could have audible effects when streaming music data. But that would be obvious "music keeps cutting out" issues, not nuanced sound quality effects. But the same holds true for a failing or overloaded local drive (overloaded in terms of IO requests, not capacity).
By his thought process, do videos played from a remote file store look worse than those played back from local disk?
They can do (stuttering, gaps, buffering), but it's more likely to be as a result of slow serving by the file store than lack of bandwidth on the Wi-Fi. You can overload the system if you try and stream to too many sinks though.
There should be no problem with music in most circumstances. If there are 6 users on YouTube it could be a different matter though. You can have distinct networks if you need to.
As I have reported many times on this forum over the years- no difference. It is easy to transfer many gigabits over wifi with out a single error. Not 1 error in a billion. Not one.
And it's also trivially easy to get a bit perfect copy of an audio file from a remote server, thru a music player via SPDIF out an in, and still be bit perfect. The system simply works. As for what happens once your DAC has the bits, that's another story. But getting them to the DAC is NOT a problem.
I've done the tests. It's not difficult. And it's not analog! The digital format is what makes it easy for the end user to get a perfect transfer. Not one part in a billion wrong. Kinda cool, really. 🙂
And it's also trivially easy to get a bit perfect copy of an audio file from a remote server, thru a music player via SPDIF out an in, and still be bit perfect. The system simply works. As for what happens once your DAC has the bits, that's another story. But getting them to the DAC is NOT a problem.
I've done the tests. It's not difficult. And it's not analog! The digital format is what makes it easy for the end user to get a perfect transfer. Not one part in a billion wrong. Kinda cool, really. 🙂
Mike, just to mention some rather uncomfortable issues, one of which is that we've burned two bit-perfect copies from one source that sounded different, actually quite a bit different. There is a lot more than ones and zeroes involved and when manufacturers started getting concerned why digital wasn't really all that good they finally started bringing in EE's to solve the problems that the Computer Engineers couldn't (which, BTW, were pretty much old hat to the Audio EE.)!
Best Regards,
TerryO
Tell us more - how did you compare the two copies? Any chance of sharing those two files with us?
so what was different between the 2 burns if the source was the same? different CDRs, different drives?
The title says it all: Data quality, there is no such thing - data is either intact or corrupted.
However since Wifi uses resources on less capable parts of the PC subsystem (such as the USB PHY or an internal one) it is possible for an active Wifi device to introduce glitches in the playback stream. Units with poor drivers will fare worse. The CPU core does not usually get involved in I/O work unless it needs to, all of that is handled by the PCI controllers.
This is easily verified using something like the DPC Latency checker from Thesycon, you may see big glitches when wifi is enabled. It is recommended to turn off wifi for that specific reason (not just refraining from using it for streaming data) while playing back audio (or video, but that isn't very practical). Physical network is almost immune to glitches.
In practical application if laptops are used, these usually have very bad power supplies and motherboards, and wifi signals being generally omnidirectional there is a lot of interference and EMI generated within the system (as wifi apparatus is usually placed right in the center of the laptop, under the service hood). Laptops are probably the single worst device choice for audio playback, with a combination of small harddrives and generally poor build quality.
In theory, there should be absolutely no difference. In real life though, there usually is. Mostly, quite easy to overcome.
One good way offered by Foobar is to cache the entire file to hard drive before playback starts. This usually is terrible for live albums, but works fine for shorter songs. A very large network buffer also helps.
I'm with the original statement that playback should be from local drive as far as is practically possible.
Burning is an issue as there is no guarantee that each burn will restore the data exactly - unless you rip and burn your audio CDs in data mode to disable error correction.
However since Wifi uses resources on less capable parts of the PC subsystem (such as the USB PHY or an internal one) it is possible for an active Wifi device to introduce glitches in the playback stream. Units with poor drivers will fare worse. The CPU core does not usually get involved in I/O work unless it needs to, all of that is handled by the PCI controllers.
This is easily verified using something like the DPC Latency checker from Thesycon, you may see big glitches when wifi is enabled. It is recommended to turn off wifi for that specific reason (not just refraining from using it for streaming data) while playing back audio (or video, but that isn't very practical). Physical network is almost immune to glitches.
In practical application if laptops are used, these usually have very bad power supplies and motherboards, and wifi signals being generally omnidirectional there is a lot of interference and EMI generated within the system (as wifi apparatus is usually placed right in the center of the laptop, under the service hood). Laptops are probably the single worst device choice for audio playback, with a combination of small harddrives and generally poor build quality.
In theory, there should be absolutely no difference. In real life though, there usually is. Mostly, quite easy to overcome.
One good way offered by Foobar is to cache the entire file to hard drive before playback starts. This usually is terrible for live albums, but works fine for shorter songs. A very large network buffer also helps.
I'm with the original statement that playback should be from local drive as far as is practically possible.
Burning is an issue as there is no guarantee that each burn will restore the data exactly - unless you rip and burn your audio CDs in data mode to disable error correction.
If your theory is correct then the local playback would also suffer as wifi is not turned off. That is easy enough to test.
But glitches are clearly audible and noticeable - they don't cause any subtle degradation of sound quality.
A 100uS interruption may not be noticed as an audible glitch but still cause bits to be dropped from the stream.
What is audible or not depends on many things and it is not possible to factor in every single permutation of systems, ears and beliefs.
My understanding is that how things should be and how they actually are, are different.
What is audible or not depends on many things and it is not possible to factor in every single permutation of systems, ears and beliefs.
My understanding is that how things should be and how they actually are, are different.
- Status
- Not open for further replies.