|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
129.33.19.254
In Reply to: RE: cMP - the open source high-end Memory Player posted by 40926 on December 30, 2007 at 05:42:01
What I mean is that there are obviously detrimental effects to pretty much anything - and that includes video adapter. Classic cMP recipe favors video, with control via KB/mouse/remote - with everything pertaining to networking disabled.It seems logical to check whether the opposite approach does less harm. Of course, we're not talking about enabling WiFi, either onboard or via adapter - that would unacceptably degrade the sound quality, in my experience.
Instead, music PC is either hardwired to the router, or connected to wireless AP - with video adapter disabled in BIOS. Control is performed by a tablet with remote app specific to the media player.
Of course, disabled video can be major invonvenience, including inability to perform some functions on startup, that can't be automated via scripting - for instance, setting PCI latency with configuration tool.
Thoughts?
Edits: 01/18/13Follow Ups:
1. Classic cMP - no network, monitor attached, control via KB with touchpadversus
2. No monitor, no KB, video adapter is either disabled (in BIOS or in Windows - whatever doesn't prevent XP from loading), or its impact is reduced by butchering Registry. PC hardwired to router, control via laptop for the lack of a tablet.
In preparation for cMP build, I'm doing some tests on AMD-based Windows 7 laptop, more or less optimized using cMP recipe, connected via Wyred4Sound uLink async USB-SPDIF converter. There, comparing sound quality with all networking hardware disabled, I clearly hear large negative impact of simply enabling 4 services required for wired networking - without even enabling hardware. The sound becomes duller, with reduced dimensions of the soundstage.
So, I guess my results are different from what's reported in this thread, and they don't look promising for networked solution. The only question is whether defeating video to some extent (if I'm successful at it on real cMP build) will outweight the negative impact of the network.
Edits: 01/22/13
Hi Carcass93,
I could never hear any sound quality degradation when using the wired network on my cMP setup. I also don’t see much more latency when using wired network.
But I did hear a sound quality improvement when reducing the USB-polling frequency as suggested by Ryelands. And I also could see much more latency when moving the PS2 or USB mouse around. So I thought it would be nice if USB would not be needed at all and could be disabled in the BIOS.
When I started too power the ESI Juli@ digital part with separate clean power I also thought about polluted ground connections (I don’t mean the safety earth) so in the end I followed Cics recommendation to use a high quality optical connection between cMP and DAC for galvanic isolation. For this reason I also did not like my Sony Bravia LCD TV (which I use as PC monitor) being hardwired too my cMP via the VGA-connection.
But when disabling USB and disconnecting the VGA to my Sony TV, I have too look for some methode of remote control.
Using the wired network and the VNC server software on my cMP setup did not lower sound quality in my setup. Also there’s hardly any rise in latency.
I liked remote controlling my cMP setup with VNC very much. There are lots of VNC clients available for tablets and smartphones.
When inmates started slimming the XP OS software and reported very nice sound quality improvements, I tried too follow that recipe, but I hated how that crippled my cMP machine. Since other inmates reported very good results with running Music Player Deamon on a slimmed Linux OS (Puppy Linux), I switched too MPDPUP. Which has also lots of vey nice MPD-clients available for tablet and smartphones. I like the MPD-clients even better, than the VNC-remote controlling.
So in short:
I could not hear any sound quality degradation from using wired network. But I could hear sound quality degradation from using USB.
Not using a hardwired VGA connection to a (TV/PC)-monitor, did not give me any sound quality improvement, but I want it for ease of mind when thinking about clean power and clean grounds.
Mark
cMP2 PC > ESI Juli@ > Van den Hul Optocoupler MkII > Lavry Black DA10 > XLR Vovox Direct S > Klein & Hummel O300
LOL! Mark, your post comes as I was about to mention something that runs - seemingly - contrary to what you just said about USB. In fact, contrary to what I had expected, given other people's comments in the past. So, this is just reporting in process, not "saying". I'll post more details later as I dig deeper. For now, I am just curious if someone has seen this before.To put it in some context, I am tweaking this new cMP/cPlay box. Previously, I used cPlay in my HTCP box. The phenomena I'll describe ALSO happened in the HTPC with cPlay, albeit too subtle for me to be sure. (I now suspect it was being masked by it not being a dedicated box like the cMP box.)
Here we go. In recent days, I've had my "test music selection" in the internal hard drive only. Yesterday, for the first time, I plugged my music USB drive to the new cMP box. I was bracing for a drop in SQ. It never happened. Just the opposite.
The USB drive beat the internal one. I spent last night and today confirming and trying various config variations. No dice. USB is beating internal.
Like I said, this is not new. It was happening before. The difference in the HTPC is so faint (but repeatable) that years ago I chalked it up to "astral strangeness"...not worth mentioning here. Yet, on the cMP box the contrast is remarkable.
It is interesting that the ONLY difference between the USB drive and the internal one is the degree of glare. The internal drive has more. The USB drive barely any. The violins give it away very quickly but it really is pervasive as one listens. Nothing else changes. This "glare" difference is easily heard in speakers and headphones.
In the new cMP box, the USB card is a USB-3. In the HTPC box, the USB card is USB-2. The external drive is USB-2 (WD MyBook). The internal drives are SATA. All drive or partitions where the music is located are NTFS. However, I just realized the main partition (of 2) on the cMP box is FAT32 (I now have to try placing music in there to see if it makes a diff). The cMP box has no Windows cache in disc (it is all memory)After many tests, one nagging suspicion is that the phenomena "could" be caused by the AMD chipset on both mobos (same chipset but different motherboard brands & AMD CPU types, though). The HTPC runs Win 7 Pro. The cMP box is XP Pro.
On a related subject, I just found something surprising. If one leaves a USB cable unhooked from the computer but hooked to the USB drive, the cable projects a major electromagnetic field (detectable 1 foot+ away). When close to other audio equipment, it can affect SQ. However, once the cable is connected at the other end also (the PC), the field collapses.
This came about yesterday. The equipment started sounding funny at one point. I ran my trusty electric field detector around. Amazingly, it was from the music USB drive that I had unhooked from the HTPC an hour before (planning to connect it to the cMP box later). The drive box had a strong field but the USB cable itself carried it all over! More by instinct than by design, I plugged it into the cMP box. The field collapsed instantly.
Edits: 01/20/13
I don't think it is contrary
------------------------------
Hi JBen,
I don’t think what you are hearing is contrary with mine or other peoples
comments.
I suppose your internal SATA drive is a HDD-sata drive, not an SSD-sata
drive?
In my opinion you compare the electrical noise injected by the HDD-sata
drive with the electrical noise injected by the (self powered?) USB-HDD
drive.
Removing HDD’s from your cMP-setup and using a little internal SSD-sata
instead, will give you a sound quality improvement ‘bordering on the
absurd’. As one inmate here on the PC-asylum nicely described the sound
quality improvement he got, when he removed the HDD(s) from his cMP setup
and mounted a little SSD inside.
A side note: do you hear a sound quality improvement when you lower the
USB-polling frequency? You do not mention this.
Also I would preferably use a NAS too store my music-library on (NOT an
external USB-HDD). If your MoBo has a modern 1Gb/s Ethernet port and also
your NAS has a modern 1Gb/s Ethernet port, you can connect your NAS
directly too the ethernet port on your MoBo. No home-network or special UTP
crossover cable needed.
Chipsets and processors all do have their own sonic finger-print. But I do
not know if one chipset is more sensitive ‘sound quality wise’ too injected
electrical noise than an other. So I do not dare too speculate if it has
something too do with the chipset and processor you use.
“the cable projects a major electromagnetic field, when close to other
audio equipment, it can affect SQ.”
That’s why I completely switched too pro-audio-gear. When using pro-audio
gear you can use shielded, balanced XLR connections system-wide. (although
my XLR VoVox direct-S interconnect wires between pre-amp and active
speakers are not shielded. But VoVox explicitly stats this on the package.
Since I use shielded mains cable everywhere and no other major
electromagnetic field sources are around (at least not any that I can think
off that would cause distortion), I dare too use these unshielded VoVox
direct S XLR-interconnects between pre-amp and active speaker.
Back to topic: I think you should first remove your HDD’s from your cMP
setups and change too a little SSD for OS only. Than check again what
impact the external USB-HDD has on sound quality.
Also start with bringing injected noise from the ATX-PSU further down
through simple filtering on the P4 and P24 (extra caps as describe by
Ryelands) or if funds allow use SOtM-filter sets. I have not tried SOtM
filters myself, but their concept is clear too me and people on the net in
audio-forum report these filters do work.
Mark
cMP2 PC > ESI Juli@ > Van den Hul Optocoupler MkII > Lavry Black DA10 > XLR Vovox Direct S > Klein & Hummel O300
Thanks, Mark. Yes, an SSD is in plans.
For now, I am going step by step into this; to gain a better feel for "what does what". Since, at its worse, the new cMP box is still better that the previous setup, I can live with a few weeks of incremental steps.
This USB external vs SATA internal is one curious thing...just one of more to come, I suspect. I should add that the USB drive that I tested with does have its own walwart switching power supply. These are nasty little fellows, as you know. Thus, it is intriguing that the darn drive still manages to out-smooth the SATA internal. Or, maybe that's just it! Maybe the fact that the SATA drive is inside the box, closer to the PS, is the cause. I'll be looking into it later (easy to place that drive "outside the box") LOL!
BTW, pending more strict testing, having the ethernet on vs off is not making an audible difference so far. However, I have not tried listening to files drawn from the network yet. I left testing this for last.
On the electromagnetic fields, I do have to be careful here. I have a very crowded setup and no XLR capability for now. Thus, all power cables are shielded and I also place much emphasis on cable management & routing. It is challenging but it works. However, the new cMP box is not yet inside this environment. I placed it far away enough such that once I maximize its capabilities, I'll know if something gets screwed up once it is in "its final place".
It seems logical to check whether the opposite approach does less harm.
It's a fair question. It's difficult to say whether it does more or less "harm" as it's hard meaningfully to compare the two setups. Eschewing "classical" cMP2 pretty well from the off, I've been running a "headless" cMP2-type system without KVM for several years. (There's a description of an early setup of mine on the cMP2 website.)
I can see no reason why using KVM and a PCI-based soundcard is inherently better than a headless setup that, say, pulls data over a LAN and outputs it via USB. Sure, there's a LAN overhead but, especially with a "super-slim" XP setup, it's pretty small.
Obviously, there's an overhead with video as well. Some users report sound improvements from unplugging the monitor, separately powering the mouse and keyboard and so on. That suggests there's something to be gained by not having them in the first place. So who's to say which is better?
I chose to go "headless" because it suits the way I use my system; I set it up as best I could (there's not much left to tinker with) and got on with, well, using it. The difference between a more-or-less "unslimmed" XP config and a full hair-shirt version is, I suspect, far more substantial than worrying about whether there's keyboard connected or not.
I forget who reported some trials on this (tests would be too strong a word) but he came to much the same conclusion. Ditto enthusiastic reports from folk who've gone from cMP2 to an mpdpup setup.
I'm not sure that XP (unless perhaps XP embedded) can even load if the video is disabled in BIOS but I have deleted a heap of video-related stuff in the Registry. If I connect a monitor to my audio PC, I get a blank screen - I can only access the PC via the LAN. There's certainly no Registry entries left for keyboard or mouse, unused USB ports, etc. Many entries associated with the one USB port that XP knows about have also been deleted. A good asynch USB-to-I2S is a big step up from a what's it called (adaptive?) but even that responds well to OS slimming. No doubt some will say (once a week for several years) that in an ideal world a DAC shouldn't be affected by the source but for now they are.
All this faffing about does make for great for clarity of sound but it won't win any prizes for its flexibile interface. Then again, my record deck can't run spreadsheets either.
HTH
A few months ago a friend and I gave up on attempts to mod minixp into accepting USB the way I needed. I had been counting on this before jumping to cMP. (cPlay however, has been in use for years in a "compromised" HTPC box. It did very well if I ran some scripts and turned things off).
I then looked at "mpdpup" and salivated a little...but it will have to wait for the next upgrade cycle (no drivers for my current USB sound device).
So, recently, I put together the basic cMP box. I am still in the process of tweaking things.
Relative to the video, which I have at its most base setting, EVERY step on getting closer to remove it has brought added benefits.
However, I am mystified by an unexpected result. For now, to turn off the monitor, I use UltraVNC. I can then access the screen from a tablet or another PC.
The expectation was that using VNC (and its CPU load) would STILL damage the sound. Yet, even with the dual core CPU basically using one core and at 900Mhz, using UVNC still fails to change the sound character. It stays great! At this CPU speed, I can force minor dropout if I move the screen elements too suddenly...which the "remote" function makes easy to avoid.
OTOH, just turning the monitor on is enough for me to notice a slight degradation...which BTW does dot even register as a change in the Task Manager, if it is on. Yet, the same Task Manager yells out if UVNC is on. Nevertheless, the sound is virtually the same in all respects; tonally, dynamically and image-wise, etc.
The mystery got to me so bad that I've had to try different audio systems and various headphones instead of speakers to make sure, to the extent that I can. A neighbor even blind-tested me and I him...none of us could cath a difference repeteadly.
Has anyone seen anything like this before? BTW, UVNC is arriving via ethernet. The wireless access is elsewhere.
BTW, UVNC is arriving via ethernet.
Thanks for an interesting note. If you're not already using UVNC's "mirror driver", it's well worth loading it onto the cMP2 box. Much like you, I don't find it makes for a difference in sound but do find it makes the system more responsive.
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: