Part One is only 20,000 posts, so shouldn't be much problem to read. The thread diverges from preamp discussion to "general" audio (and automotive and whatnot) discussion after, I dunno, maybe a few hundred posts.
John Curl's Blowtorch preamplifier
Part Two is 99,999 posts. If you contribute to the forum it activates a Thread Tool "enhaanced print view" that lets you see a whole thread on one big long webpage (so you can print or save the whole thread for offline reading), except Part Two is too big for the forum software to load all at once (the mods discuss this in a Forum Problems thread). Maybe that's a Good Thing ...
Poll for the participants of "John Curl's Blowtorch preamplifier part II"
No, wait, that's not it, here it is, 1001 pages at my setting of 50 posts per page!
John Curl's Blowtorch preamplifier part II
It's probably due time to lock this one and make a Part IV thread.
Last edited:
I find this is because of how much coffee I've been drinking (decades ago it was due to ingesting other substances but I quit them long ago), whether I've cleaned my ears out and maybe the spiciness of the food I've recently eaten.
But if you don't at least try to measure it, you'll never know.
Now, are you prepared to say that there was no difference between the files with the same checksum and that you where wrong on this?
Speaking of sockpuppets did 2DTV ever make it to USA? They were very cruel to George dubya as he had to have everything explained to him by a sock puppet.
YouTube for an example
Ive just got myself a miniDSP 2x4 and shall soon be getting a calibrated Umike. My intention is to measure and calibrate everything in sight and sound. Then I can talk about measuring things. I am very good talking about things that can’t be measured, only described - outrageous, eh?
tapestryofsound
It does sound different everyday until warmed up (30-45 mins usually) then it sounds the same.
This I realized.....what I did not realize was how much it was cooling off in the amount of time it takes me to make such a change as speaker wire (10 mins +/-)
The exponential cooling is something I’d not considered until marrk brought it up, his explanation was acceptable enough to satisfy my curiosity, I will not swear on anything important to me that’s definately what it is but it sure makes sense to me.......sometimes it’s something obvious that you just don’t put together that’s why I ask. btw....thanks Mark
Thanks for your concern cbdb
YouTube
Was the amp turned off when you change the cable? Amps usually do not sound the same at turn on and after being settled. Some amps have bias current that is increasing after turn on. A few amps have constant bias current at turn on and after a few minutes (at no music or even at high volume!), but even so, the sound will often change also when the transistors get some heat.
I monitor currents (and DC) in my amps lively when listening to music, for above reason.
Not always better, can become worse too. In CFP output amplifiers, bias current is very critical and a few mV (or mA) can change sound quality (not just sound difference). So when bias current is increasing after turn on and the lower bias is the good one, then the sound will be worse after settling in. The time to settling in condition can vary from seconds to half an hour, depends on the design of the Vbe or temperature compensation scheme (and stability too, of course).
Nope.
I have rerun the experiment several times of transferring two copies of the same wav file to a thumb drive and have found them to sound different in blind testing with witnesses.
I don't know the mechanism and neither do you, I standby my proposal that different system noise during write process causes different write errors and subsequent different error correction during read process.
When you have a product that you can insert into car cigarette lighter socket and change the car stereo sound let me know, until then you are out of your depth.
Dan.
If you have speaker switching this might be preferable.,
Dan.
For real ?
There are no write errors in a non pathological drive. You are imagining things.
Any significant BER will manifest itself as poor performance. In the case of a NAND flash SSD, in the event there are recoverable read errors it is corrected by the controller seamlessly.
You do understand that the data is not streamed off of disk directly to a peripheral, right? It goes to RAM first on a PC. There are few cases of peer to peer DMA in a PC. In any case, this data is buffered somewhere before it sees a DAC.
Last edited:
Thanks, but, no thanks...
From what little I've picked up in these recent posts, it's not at all interesting to me.
Bunch of babbling.
So with that, I'll just "unsubscribe it from my list.
😱
I suppose your personal definition is just as valid as mine in Mark's book and just as easily understood 😉
if the source during playback is a USB stick and there is other activity on the USB buss there can be issues. Usually they are not subtle but anything from increased noise radiation to data collisions since the USB is not a priority on the buss. A standard USB audio device would be but some USB audio implementations are not isochronous and suffer for it. This is all really simple stuff to identify and isolate. I would not play from a USB stick if there is another option (and I would prefer a PCI sound card for its deterministic interface as well. . .). Could different parts of the USB memory device sound different? Maybe they too can have different noise profiles depending on how the USB device manages its internal memory. But move the data to the host PC and try again for the "same" differences. Or set up to measure the radiated EMI/RFI noise during the different playback cycles.
I don't think we should confuse USB Audio Device (endpoint) issues with the act of reading from a file and playing it back through a piece of software. They are two different things.
Given that the USB Audio Class requires isochronous transfers, they occupy a reserved portion of the USB frame. It's not possible for USB bulk transfers from a Mass Storage Device like a flash drive to take that away. Any implementation of a USB sound card that is not a USB Audio Class Device is very rare these days. It is true that poorly written drivers exist and so do poor host controllers (Synopsys / DesignWare, older ASMedia) with an errata list that is pretty scary.
People that experience USB audio dropouts or glitches have a misbehaving driver. Windows and Linux aren't RTOSes and don't have fully preemptable kernels, and they are also monolithic kernels. Which means any idiot writing drivers can write a driver that starves the rest of the system of kernel time, causing pops/clicks/dropouts and worst case BSODs.
I *really* doubt there is a measurable difference in noise anywhere in reading from one block of flash to another on the same device.
Modern systems, with properly behaving drivers and EFI don't have problems with USB contention. Intel desktop chipsets since the X58 days have at least something like 6 separate USB host controllers. You can count the number of USB controllers and root hubs in Device Manager in Windows.
Anyway, despite the technical details leaving a 0.00001% chance of a real effect, we know that Max would say this happens anyway even with a PCIe (onboard) audio device using files played back off an internal hard drive. Maybe I'm an ***, but we are not talking about someone who has been painstakingly attempting to measure common-mode conducted noise on the USB cables connected to his PC.
I guess that's a long winded way of saying you're right in a sense, but I would not put money on that being what's going on or what Max claims to hear.
Given that the USB Audio Class requires isochronous transfers, they occupy a reserved portion of the USB frame. It's not possible for USB bulk transfers from a Mass Storage Device like a flash drive to take that away. Any implementation of a USB sound card that is not a USB Audio Class Device is very rare these days. It is true that poorly written drivers exist and so do poor host controllers (Synopsys / DesignWare, older ASMedia) with an errata list that is pretty scary.
People that experience USB audio dropouts or glitches have a misbehaving driver. Windows and Linux aren't RTOSes and don't have fully preemptable kernels, and they are also monolithic kernels. Which means any idiot writing drivers can write a driver that starves the rest of the system of kernel time, causing pops/clicks/dropouts and worst case BSODs.
I *really* doubt there is a measurable difference in noise anywhere in reading from one block of flash to another on the same device.
Modern systems, with properly behaving drivers and EFI don't have problems with USB contention. Intel desktop chipsets since the X58 days have at least something like 6 separate USB host controllers. You can count the number of USB controllers and root hubs in Device Manager in Windows.
Anyway, despite the technical details leaving a 0.00001% chance of a real effect, we know that Max would say this happens anyway even with a PCIe (onboard) audio device using files played back off an internal hard drive. Maybe I'm an ***, but we are not talking about someone who has been painstakingly attempting to measure common-mode conducted noise on the USB cables connected to his PC.
I guess that's a long winded way of saying you're right in a sense, but I would not put money on that being what's going on or what Max claims to hear.
Last edited:
Chris, multiple pairs of files sounded different on dedicated playback devices such as shelf mini-system and car audio. One explanation might be flash memory less errors on write and less errors on read causing less correction and timing noise of decoded/corrected and operating system level delivered data.
OTOH somehow it seems that there is something deeper going on and more is recorded than just ones and zeros and that energy storage, energy transfer and energy transduction are more complex than we consider.
Dan.
Last edited:
We are in the same camp on this issue. While finding something is possible it's really really unlikely with reasonably current hardware.
However I can make some amazing Ethernet cables for $1000 a foot with a $2000 termination fee. . . Cash up front and they need a 2 year aging for stability before they can be shipped.
However I can make some amazing Ethernet cables for $1000 a foot with a $2000 termination fee. . . Cash up front and they need a 2 year aging for stability before they can be shipped.
Recording something other than 1's and 0's? You lost me there.
I do understand your gripes with USB audio, though. I haven't written a driver for it, so I don't know exactly why, but a lot of vendors seem to have trouble writing decent ones. The fact that it has another layer, running above the host controller driver may have something to do with it. I have a Focusrite Scarlett 2i2 Gen 2 and the Windows drivers are BAD. They will literally crash one of my PCs randomly during playback. That PC is is a 6 core / 12 thread Xeon @ 4.2 GHz that can use every single input and output channel of an EMU 1820m over a PCI to PCIe bridge without a hiccup.
Last edited:
- Status
- Not open for further replies.