|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
76.122.133.81
In Reply to: RE: Yes: EXACTLY the same way posted by hfavandepas on January 31, 2011 at 06:54:38
Thanks Mark ... now I will spend rest of the day figuring out what a NAS is.
Follow Ups:
fully optimized cMP2 PC -> ESI Juli@ -> Van den Hul Optocoupler MkII-> Lavry Black DA10 -> XLR Mogami Gold -> Klein & Hummel O300
Some interesting points are being raised here. Here's some more for the mix.* Following a tip from Theo, I experimented yesterday with playing from data stored not on a "watched" folder on the 2.5" HDD attached to my cMP^2 box but on a folder on a USB flash drive.
The data are identical, the HDD is still spinning and, obviously, a USB port had to be re-enabled. In spite of all that, there was [is] a marked improvement in SQ.
* If I set the HDD to spin down after three minutes of inactivity, it does just that, suggesting that there is no disc-related R/W activity during replay. The disc has in any case a good dedicated PSU, is carefully mounted outside the CPU box, its SATA lead is fairly well shielded from the motherboard and its platters spin at a sedate 4,500 rpm.
* For all practical purposes, system power consumption can reasonably be assumed to be the same. I haven't measured it but it is unlikely to be any lower.
* Given the disparity in data rates between an SSD, an HDD and a flash drive, they are unlikely to be the cause - it takes roughly ten times longer to copy an album to the flash drive than to the HDD whereas an SSD is faster than an HDD. But the HDD seems to score worse than either for SQ.
* The only difference I can spot is that music data are being read from flash memory via USB rather than a disc via SATA.
That a similiar effect is reported if the HDD is replaced with an SSD suggests - though it certainly doesn't prove - that it's not to do with bus topology. It seems to be a case of flash memory v rust.
This should not be and, no doubt, some will tell me in no uncertain terms that it isn't.
Whatever, those little 8GB Kingston SSDs are also being sold off cheap in the UK so I've ordered one. With a bit of luck, I should be able fairly soon to repeat hfavandepas's experiment of putting the OS on one.
Watch this space.
++++
* I see that hfavandepas encountered the same issue as I did when pulling flac-format data over a LAN to a cMP^2 box. It's a bit of a relief that the problem has been replicated even if he seems to get it less often than I did. (I managed to "repair" some files by re-ripping the CD but it wasn't foolproof and in the end I had to work out a different playing routine.)
Edits: 01/31/11
Hi Ryelands,
* Just too be sure I understand everything correctly (I’m not a native English speaker):
In your cMP^2 setup you use a 2,5” HDD with dedicated PSU with the OS + music files stored on it.
You than plugin a USB flash drive, you copy some music files onto that USB flash drive and you play the music files from that USB flash drive.
The result is: the files played from the USB drive sound better than when the same files are played from the HDD.
Hmm well yes…. That’s remarkable.
I haven’t tried that, but I still have the Samsung HDD lying around too plug it back in and see if I can replicate the phenomenon.
* On sound quality and spinning HDD.
If after 3 minutes the HDD spins down, I don’t hear any sound quality differences in my cMP setup.
Do you hear sound quality differences?
In my setup, I just swap the HDD out and the SSD in place.
So: same power supply situation, same sata port, same sata cable.
Just the fact that there is a HDD connected (with dedicated power supply through a ‘dirty’ picoPSU 150 XT) results in sound quality degradation (compared too a SSD).
It doesn’t matter if the HDD is spinning or not.
* On strange behaviour from FLAC files in cPlay.
I noticed some more strange behaviour when cPlay handles flac-files.
Sometimes cPlay also doesn’t list all the individual tracks in a flac-file.
Though when using Foobar on the same flac file, Foobar is listing these individual tracks.
I haven’t looked closely into these differences between foobar and cPlay because I convert flac files back too wav-files.
But it seems Foobar does a better job handling flac files.
Soo may be cPlay needs ‘a closer look under the hood’ too see if it’s handles flac files correctly.
* I’m looking forward too hear from your experience with implementing that little SSD in your setup.
Mark
fully optimized cMP2 PC -> ESI Juli@ -> Van den Hul Optocoupler MkII-> Lavry Black DA10 -> XLR Mogami Gold -> Klein & Hummel O300
the files played from the USB drive sound better than when the same files are played from the HDD.
Yes and it's not a small difference. Blame Theob. I do - it was his idea.
Note that re-enabling the USB port and inserting the stick are the only changes I made. It's weird; if you can repeat the trial, I'd be delighted.
If after 3 minutes the HDD spins down . . . Do you hear sound quality differences?
No but I didn't try very hard, to be honest. (I don't normally set the disc to spin down as it would be stopping and starting almost every track.)
. . . may be cPlay needs ‘a closer look under the hood’ too see if it’s handling flac files correctly.
The only time I get trouble with cPlay and flac data is with some tracks when pulling them over a network under a hair-shirt cMP setup with issas disabled. In every other situation it works absolutely fine. (I've not tried the same setup with Foobar for SQ reasons.)
That has to be, by any reckoning, an obscure bug and I don't even know if it's a cPlay one. It could equally well be a flac bug or an OS bug or . . .
As I'm asking cMP^2 to do something its author never intended, I devised a work round that works fine for me. You did much the same, I think.
Here is my experience/take on the ssd or flash drive sq phenomena. I noticed 18 months ago or so that my then 'noise problem' with cplay was positively affected by playing music off a flash drive. So I tried (unsuccessfully) to devise an algorythm to automatically load a flash drive from which cplay could play music files. Any way at that time I read just about all of the cmp thread to understand what was going on. I remember cics posts showing current spikes when using different power supplies for hdd's. Extending this to hdd's and flash drives it seemed logical that yes although the hdd's in my and Ryelands system are still spinning when running off a flash drive and therefore using the same rotational energy the little arm that moves across the hdd is not moving. And this starting/stopping/reading mechanism is no longer in play when reading off a flash drive. Hence less energy/power draw and better sonics. Running an OS off a ssd takes this to another level and perhaps explains Mark's wonderful sonic results. I think cics also mentioned that use of an ssd was a sonic opprtunity. When you say blame theob I turn it around and say blame Mark for resurrecting this idea for me to attempt the Ryelands work around. Dave when you get your ssd and load your os and it moves the needle further on sq I think I may go this way as well and try the Ryelands work around 1st and perhaps then the NAS approach.
Hi Theo,I can replicate your USB flash drive experience.
Playing a wave file in cPlay (44.1 -> 44.1, buffers: tiny, AWE allocated) of an USB Flash drive improves sound quality.
However, improvements are very subtle.
I had too do some careful A/B comparisons,
but sound quality improvements are also audible in my setup.
More space, more dept, more micro details, etc.It gives the same type of sonic improvements as an SSD gives.
It gives a glance at what a small 8 Gb SSD did for sound quality in my setup.
Although using an SSD gave 20 times this type of sonic improvements.Mark.
Some setup info:
- core speed: 600, Bus speed: 100, FSB: 400.
:-)) lowest possible bus speed setting in bios !
- Cruzer 2Gb USB flash drive
- 500 Gb Samsung spinpoint sata HDD
- USB flash drive in same separately powered USB bracket as wireless mouse/keyboard
- pico 160-xt on p24
- linear psu on p4
- USB bracket and Samsung HDD separately powered by pico PW-200-M
fully optimized cMP2 PC -> ESI Juli@ -> Van den Hul Optocoupler MkII-> Lavry Black DA10 -> XLR Mogami Gold -> Klein & Hummel O300
Edits: 02/08/11
Guess I have to get an ssd
Hi Theo,Your suggestion works flawlessly.
It only took me 5 minutes too (re-)connect my Buffalow NAS directly too my cMP setup.I didn’t have to change any network settings.
I just disconnected my NAS from the switch in my home network,
and re-connected it (directly) too my cMP setup.* use of fixed IP-addresses (than a DHCP service is not needed to assign them)
I already had assigned (manually) a fixed IP address too the NAS (192.168.1.3).
I also already had assigned a fixed IP address too the network card on the cMP-mobo (192.168.1.100)
So I didn’t had too change these already manually assigned IP addresses .
I could use the same fixed IP addressed as I already had assigned for use in my home network LAN.* Workgroup membership.
Both the NAS and the cMP where already members of workgroup: “vandepas”.
So I didn’t had to assign the same names to both devices again.
They already where member of the same workgroup.
* default gateway settings seem too be irrelevant in a direct connection.
I also didn’t had too change the default gateway settings (192.168.1.1.) which are necessary too function in a home LAN network setup.* no special network (cross over) cable needed.
I also didn’t have to use a special pathed network cable to make this direct network connection work.
This means both network sockets: in the Buffalow NAS as well as on the Gibabyte mobo, are auto cross sensing.
So because of this auto cross sensing, it’s not necessary too use a special pathed (cross-over) network cable.
If network sockets are not auto cross sensing, than a special patched (cross-over) network cable has too be used too make direct connections.
( I used the 4 meter network cable that already came with the Buffalow NAS in the box).Except for funding and budget, their seems little in your way too start using a NAS (directly coupled too your cMP) and a small SSD.
:-)
The 8Gb Kingstone will set you back: $ 45,-.
A Buffalo LinkStation V-Series NAS: 1 Terra bite, will set you back: $ 189,99.Enjoy !
Mark
** edit **
I read on this wiki-page:
http://en.wikipedia.org/wiki/Ethernet_crossover_cable
“in practice Auto-MDIX (auto crossover sensing) is implemented on all 1000BASE-TX interfaces”.So indeed (as I already experienced this afternoon) no special crossover network cable is needed when connecting 2 modern network devices directly together.
fully optimized cMP2 PC -> ESI Juli@ -> Van den Hul Optocoupler MkII-> Lavry Black DA10 -> XLR Mogami Gold -> Klein & Hummel O300
Edits: 02/08/11 02/08/11 02/08/11
Hi Theo,
I think your idea too connect a NAS directly too a cMP setup is good idea. Better than connecting a USB- or Firewire HDD
This way, the SSD for OS can be a very small one.
A small one, used less power, etc, ect.
Also using a network connection, will not degrade sound quality.
At least, in my setup it does not degrade sound quality.
But this is also reported by others.
An additional nice thing about a NAS is: it already uses a separate power supply by design.
And the LAN network socket on the mobo is transformer de-coupled.
I couldn't stand, not knowing exactly the answer to your question if a NAS could be connected directly too the cMP mobo network socket.
I only vaguely know, it should be possible somehow.
So I will do some google-ing and see if I can fill the gaps in my network knowledge and see if I can get such a configuration work with my own NAS and own cMP.
I'll get back too you in this one.
As I think your idea too use a NAS instead of a USB/Firewire HDD, is a better approach.
It could be very useful for other inmates too.
Mark
fully optimized cMP2 PC -> ESI Juli@ -> Van den Hul Optocoupler MkII-> Lavry Black DA10 -> XLR Mogami Gold -> Klein & Hummel O300
I remember cics posts showing current spikes when using different power supplies for hdd's.Fair point except that with memory loading (AWE) enabled, the R/W heads are moving for only a small period during each track. The SQ lift is apparent even while that activity is taking place (assuming that the disk activity lights give a fair indication of when it's happening).
Running an OS off a ssd takes this to another level
There is little if any disk activity during play other than the reading of data at the start of each track (tracks #1 and #2 at the start of #1, #3 at the start of #2 and so on except for long tracks that need to be split).
I am baffled as to possible cause though the effect is palpable. It may be that it's that different motherboard devices are involved in reading from USB or SATA. That would at least explain why the effect is noticed even though the HDD is still present. That an SSD sounds better than an HDD may not be the same issue.
I need to look something or other up somewhere or someplace but I don't know where to start right now.
Whatever, what I have here is a cumbersome but effective "tweak" and v-e-r-y cheap.
Can we twist arms to get someone else to try it? It'd take about 10 minutes.
++++
I may . . . try the Ryelands work around 1st and perhaps then the NAS approach
As far as the cMP^2 box is concerned, there is little if any difference between the two. The OS sees a shared folder on the LAN but can't tell if it's on NAS or (as mine) a drive attached to another XP box.
In any case, if you use wav-format data, you won't hit the flac bug, making my trick of copying the data before playing superfluous. H's way is then probably easier.
* I just tried for old times' sake pulling flac data over the network. I can't get it to work at all now. Aaaaaaargh! Well, I can live without it . . .
Edits: 02/01/11
Post a Followup:
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: