|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
98.66.220.8
In Reply to: RE: cPlay - the open source high-end audio player using ASIO posted by cics on May 05, 2008 at 13:31:58
Grant and I were talking last week about all things cMP when the subject turned to how and why minimizing the available memory would be useful.
At the beginning cics wanted to use a small memory to minimize power usage. But when we do this I suspect the memory still uses the same amount of power.
Well, the only way to find out is to do it.
I took it out of my boot.ini and I must say it does make a difference. Taking it away has reduced noise not unlike removing the video and audio chips on my board.
Previously I had noticed a strange electronic noise when cPLAY was engaged. While the cMP screen was showing there was silence but as soon as I selected something to listen to, and especially with orchestral pieces where there are a few moments of silence before things get started this noise would begin. When the music began it was relegated to the background but of course it was still there.
With it removed so is the noise and I notice a general quieting in the noise floor. I was hearing some random soft pops - I listen to LPs so this was not bothering me but now that is gone.
I am upsampling at the maximum level. I am using a 512kB memory and I am using JULI@ so I wonder if what Serge is hearing with this is very different than what I am hearing. If he is not oversampling maybe this has a different effect with his system. We know he is not using JULI@
I am hearing a more robust sound. The sound is not as recessed which is an effect I liked but now I wonder if that was a sound effect instead of accuracy.
Serge, please explain what this is supposed to do.
I get the feeling the system must actually have to work harder with this in the system. Obviously the memory is and I do not think it a bad thing for the CPU to be given a load. I wonder if the CPU like a class AB amplifier might sound better working harder but there is more to it than that. It seems to me this is an obtrusive instruction that gets in the way of the machine doing what we want it to: simply process the information with as little interference as possible from unneeded files and registry entries that must be polled (I assume in my poetic/idiotic understanding of how computers work).
Follow Ups:
I tried taking out maxmem from my boot file and while I did not hear less noise I did hear many of the things you mentioned. Its not subtle. I hear soundstage more up front (but less depth). I hear a fuller midbass (but less deep bass). I hear more distinctness with maxmem back in. I think my reaction is the difference between our systems. My system (electrostatics and cone drive bass) is very inefficient (maybe 85 db) whereas your system (horns) is much more efficient. Bottomline I prefer maxmem back in.
But this is yet another way to balance one's system. Its just like, imo, reducing cpu frequency in bios...I tend to find the value that sounds dynamic/lively (not too low a freq) but not hard in the upper mids, highs (not to high in freq).
like I said I wonder if it is sound effects?
I did like the effect that maxmem gave, too, but the reduction of noise is something I think I will prefer over time.
There is the possibility that maxmem will work better after being re-inserted. I have noticed this many times with other tricks.
I still cannot figure out WHY this should do something worthwhile and I wonder if it has something to do with upsampling v. no upsampling?
You said: '...like I said I wonder if it is sound effects...'.I think its all about sound effects, that's what voicing is all about. Here's another item that has an air of absolute science about it but is really pragmatic in a sense ... 'lowering the voltage or cpu frequency' to lower rfi. That may be true (that it reduces rfi)but it is also just modifying the power supply. For example we're all told to lower the voltage and cpu speed as much as possible (to reduce rfi). But taken to an extreme if you lower both to zero and what do you get...no music right. Raise them both till you get something to work and get music...but it may be real weak music... soft, no strong bass, no high frequency extension. Raise it too much and you get strident music. So you try to set it just right...not too high not too low. There is much about audio design that is really just about voicing something just right. Yes new technology begets new electronic devices to produce music but they still have to be voiced. I believe since we have so many variables in computer audio to play with ... we should play with them to voice our system. Why not? I also believe those that have tried computer audio and dismiss it have just have not really tried to 'voice their system'. Look at it from an analogue perspective...you need to adjust cartridge vertical/horizontal tracking angles, stylus tracking force plus electrical loading (47k ohms or 10 ohms). If you don't do all that (plus a lot of other stuff) you don't optimize/tweak the system to get it...just right.
I bet if any who claim computer audio is not up to analogue or even good cd disk playback ever heard your cmp fully optimized as I know it is they would be instant believers in computer audio. I know I am.
Edits: 01/23/12 01/23/12
Adjusting a cartridge's load is not neccesarily about voicing though one could do that if they wanted. There are actual electrical parameters that need to be met due to the cartridge's characteristics.
All I can say is that I no longer am getting the blue screens and squeal and it is not as if the sound I am hearing is not attractive.
It is a more robust sound.
When someone loads a MC cartridge with too low resistance one will get a bassy bloated sound and lose output voltage; one can like it but it is not (really) what the designer intended.
One can take the voicing thing to the point where fidelity to the sound that was actually captured has little to do with what is actually ultimately heard. Sometimes this is good for a really bad recording but usually will introduce distortions (euphonic to some) that will mask what is there. This can be advantageous on some recordings but usually makes a mess of many others. I guess I want as much information as I can get and that is my bias. Neutrality is my goal. Sometimes you have to live with the warts.
I would like to know the reason for /maxmem since there is no reduction in power used by the memory when implementing this so the argument about reducing RFI does not apply here.
I think it is something that worked before and has been made superfluous with later developments.
Just my opinion.
For me the same experience: without the maxmem it sounds just a bit more musical, more body, more detail. Very subtle though.
I implemented it in a former setup (H55M), and continued this in the current atom d525 setup. So for me the lesson is every new setup needs its own check of the tweaks.
As for upsampling: i tested:
A. songs 44.1 upsampled to 88.2 by cmp
B. same songs upsampled 'by hand' and stored in a 88.2-file, no upsampling by cmp.
B sounds better, little more music, stage, detail
So in this case: the less work for cpu, the better
Other than the massive additional storage needed I do not doubt it is better to do this before but I have been concerned that there is something about SRC that is not available from, say, dBPOWERAMP which will upsample. I am sure there are others that I am not aware of.
My DAC upsamples everything to 192 and seems to me SRC/cMP does a better job so I leave cPLAY with 192/highest resolution.
What are you using for upsampling? I would like to give it a try.
Thanks,
For upsampling i use SoX (like cPlay and Foobar). Ik like it more than SRC. These are the settings i use in batch-resampling:
mkdir H:\MUZIEK_ORIGINEEL_3288\A_sox
FOR %%A IN (%*) DO sox %%A -b 32 -t wavpcm "H:\MUZIEK_ORIGINEEL_3288\A_sox/%%~nxA" vol 1.0 rate -v -s 88200 stat -v
pause
I don't recall the meaning of all settings (did it years ago).
I think 'stat' gives report which i use for checking on clipping. If clipping occurs, i set volume to 'vol 0.8' and retry.
32 stands for 32bit lots of 'zero's', huge files but better sound
My DAC upsamples to 176.4. Thats why i let cPlay/myself upsample to 88.2 In my setup this gives the best results.
Douwe
When I did upsample I also preferred sox. Now I listen to 44k sr. But the idea to restructure the file to 32 bits is very clever. I should try that.
I did not realize/remember that.
I think my problems with it are entirely due to upsampling.
I used to halve the duties in a similar way but eventually settled on full on cMP.
On cPLAY I seem to prefer SRC to SOX though I have no way to describe the differences I hear and wonder if I really DO hear a difference.
But upsampling the file does interest me.
Thanks for bringing it up!
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: