Skip to main content ↑ ↓

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: Ubuntu + pulseaudio + max settings + m-audio fast track pro = beyond i (Read 14175 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Ubuntu + pulseaudio + max settings + m-audio fast track pro = beyond i

Reply #25 –
unity100.  your reference to the wikipedia page for sonic weapon... it's called a weapon for a reason.  I personally wouldn't want any super high frequencies directed towards me with significant amplitudes. 

also, someone mentioned that upsampling from 44.1 to 96 does nothing.  while there are certainly proponents of this argument (including Schiit Audio who claims it "destroys" your music) I'll just point out 2 things:

1.  imagine looking at a picture of a set of stairs from the side.  that's fine they're stairs right?  what if someone pointed out that it's not actually a picture of stairs, but a diagonal line.  it's just an 8 bit photo.  then they hand you a 24 bit upsampled version of the same photo which shows a diagonal line.  if you get out a magnifying glass to look closer you see that it is still stairs, but just smaller steps.  converting to analog is like trying to figure out what the original object was.  now, if the original object was a diagonal line then the upsampling was helpful to you.  if instead if had imperfections or waviness that made it unique and special, but the original 8 bit photo lost that detail, then upsampling can't help as much.  sure upsampling will make it look a little more like the original, but it can never be as good as if the original photo was taken with a 24 bit camera. 

2.  there are world leaders in "digital" that say upsampling is great, such as DCS of England.  they make DAC's and such.  one of their reps once used an anecdote that went something like, "audiophiles can argue all they want, but when your DSP is in the nosecone of a plane and some guy on a sand dune has a surface to air missile pointed at you, you don't mess around."  by this he meant that they supply parts for defense contractors for life and death situations... and they (DCS) use upsampling.

Ubuntu + pulseaudio + max settings + m-audio fast track pro = beyond i

Reply #26 –
Quote from: galtthedestroyer on
[...] (including Schiit Audio who claims it "destroys" your music) [...]

Did they come to that conclusion using accepted standards such as double blind testing? Please give references. If not, their claims are anecdotal and can be disregarded.

Quote
1.  imagine looking at a picture of a set of stairs from the side.  [...]

There are no stairs, just sample points! By definition we know that these sample points were sampled from a continuous, band-limited wave. At the output stage the reconstruction filter will produce a continuous, band-limited wave that looks exactly like the one that was recorded. There are no stairs!

Upsampling will not introduce data outside of the original band limits. The reconstructed wave will still look exactly like the original band-limited wave (without stairs).

Quote
2.  there are world leaders in "digital" that say upsampling is great, such as DCS of England. [...]

Upsampling as in oversampling has it's merrits, sure. Your product can be built cheaper than a non-oversampling equivalent. You could even argue that specs achievable with good oversampling equipment can not be obtained with classic, non-oversampling equipment (or only with ridiculous effort and cost). The oversampling is done in hardware or perhaps the driver -- both of which you can't influence -- usually at frequencies far above 96k.

So what is your point? Didn't you just refer to Schiit who allegedly claim the exact opposite? Well, "audiophiles can argue all they want"...

Ubuntu + pulseaudio + max settings + m-audio fast track pro = beyond i

Reply #27 –
Quote from: galtthedestroyer on
also, someone mentioned that upsampling from 44.1 to 96 does nothing.  while there are certainly proponents of this argument (including Schiit Audio who claims it "destroys" your music)...


Upsampling always add undesired harmonics when you pass the signal to a DAC.

Ubuntu + pulseaudio + max settings + m-audio fast track pro = beyond i

Reply #28 –
Upsampling, done properly, will not add harmonics (or other stuff not present in the source) of any significant amplitude to the signal.

Where and how does the DAC discriminate between signals upsampled to frequency X or signals natively sampled at frequency X?

Ubuntu + pulseaudio + max settings + m-audio fast track pro = beyond i

Reply #29 –
Quote from: galtthedestroyer on
watch out Nezmer.  I got busted for making a less drastic statement than you.  I'm still at a 33% warning level because of it.  I said something like "OSS is more audiophile".  What I meant was that some audiophiles were snobbish about it.  What it was taken to mean is that I claimed it was better than ALSA.  and I didn't reference an ABX. 

so, if you have it, for your own good, change your statements to "so and so have reported OSS to be better and here's my reference to so and so." YMMV

or here's my ABX data.


Off-topic:
I didn't know I can't edit my posts in this forum (Can I ?)
I'd like to fix some spelling errors and a couple of missing words.

On-topic:
Is it possible to ABX this directly?
OSS4 disables ALSA when installed. So, As far as I know we can't ABX this in a controlled environment.

References:
I don't know where to begin or what to choose:
What about the ALSA Wiki?  Or a Hydrogen Audio thread? Or distribution wikis ?  Or Blog Posts where kernel developers don't even try to rebut in the comments the fact that OSS4 is superior compared to ALSA? They just complain about OSS4 developers past practices.

Ubuntu + pulseaudio + max settings + m-audio fast track pro = beyond i

Reply #30 –
Quote from: Gecko on
Upsampling, done properly, will not add harmonics (or other stuff not present in the source) of any significant amplitude to the signal.


You named a key word here: "Significant". Each pass of conversion or sampling adds quantization errors to the resulting analog signal, and you must evaluate if those errors are small enough to be insignificant. Sometimes upsampling errors are insignificant (Example: Upsampling an audio signal from 96 kHz to 192 kHz), sometimes are not (Example: Upsampling from 44.1 kHz to 48 kHz).

Quote
Where and how does the DAC discriminate between signals upsampled to frequency X or signals natively sampled at frequency X?


The DAC does not know (And it does not care) if the signal is upsampled or not, or if the data has upsampling errors due to data stuffing. The clock in the DAC latches itself to the frequency of the pulse train, then the DAC start processing the data to do the conversion.


Ubuntu + pulseaudio + max settings + m-audio fast track pro = beyond i

Reply #31 –
Judging by what the OP posted, he is resampling everything via pulseaudio to 96kHz and claiming vast sound improvements. I'm saying: resampling should be transparent. Pulseaudio may well use a worse algorithm than SOX, but does this lead to a perceived improvement? Probably not.

The OP has not yet responded to whether he has introduced sound altering mechanisms in the processing chain which could explain what he hears.

Quote
Sometimes upsampling errors are insignificant (Example: Upsampling an audio signal from 96 kHz to 192 kHz), sometimes are not (Example: Upsampling from 44.1 kHz to 48 kHz).

I just tried the following experiment:
Generate a full-amplitude sweep from 20-20000 Hz with foobar at 44.1kHz and save to a 16bit dithered wav.
Load said wav into foobar and convert to a new wav with the SOX resampling DSP enabled (target frequency: 48kHz, quality: very high, passband 95%, do not allow aliasing, phase: 25%)
When you load the resampled wave into Audacity and use the spectrum view (window size: 4096, window: blackman-harris, f_min: 0Hz, f_max: 48000Hz, gain: 0dB, range: 115dB, frequency gain: 0dB/dec) you can clearly see the sweep and the dither noise. But no harmonics.
When you change the range to 120dB you start to see faint lines of the harmonics.

Retrying the experiment with 24bit, there is nothing to see even if you crank up the range to 160dB.
Retrying the experiment with 16bit but without dither, there is nothing to see but the noise floor.

In my book, that is insignificant!

Ubuntu + pulseaudio + max settings + m-audio fast track pro = beyond i

Reply #32 –
Quote from: Gecko on
It seems to me, the discussion of possible benefits of reproducing the frequency range from 22.05kHz to 48kHz is irrelevant to the topic at hand. The resampling of CD quality material to 96Khz (as is done here) will not introduce any significant content above the original Nyquist frequency (22.05kHz) unless the process is broken. Let's talk about things that actually exist.

unity100, what are the "max settings" you refer to? Are they what you posted above or have you introduced additional processing?


its what i have included. if there are other settings already buried into this .. aaah. pulseaudio thing, i dont know about that. i dont know whether pulseuadio uses ALSA behind either. these are out of my area.

i found those settings after a google search. even 48000 setting for resampling rate had put a toll in my cpu/sound card. after i upgraded, cpu no longer gets clogged by 48000 or even 96000, but this time one of the new sound cards i bought fails to give out proper sound. (skips). the other one (audio production grade one) copes up very well though. both cards were advertised as 96 khz sampling capable. so far one doesnt seem to deliver it in linux. not too big a matter tho since the others' subjectively perceived sound quality seems better so far.

Quote from: Gecko on
The OP has not yet responded to whether he has introduced sound altering mechanisms in the processing chain which could explain what he hears.


this part is the interesting bit actually.

while using windows with my old card (and i also tested one of the new cards i bought - the non-audio production one) i always used crystalizer (comes with these cards' software), also SRS audio sandbox dsp. with these i would set crystalizer to max, and i also would set srs audio sanbox's 'definition' to max and high focus settings.

however on linux, the sound i got was clearer and better than these, without using any kind of dsps - when i used the settings i posted in the initial post of this thread.

then i used audacious player's crystalizer plugin and set it to 4.5 - 5 or so, and the subjective result i had perceived was much much better than what i had before. this led me to buy these new cards, and run them with linux.

then i tried to match linux sound with windows by using a lot of plugins and vsts and whatnot. (learned a lot in the process). what i used were exciter, expander plugins, equalizers, resamplers, many many things. so far, i wasnt able to match linux sound in windows.

i even attempted to set up pulseaudio in windows. but couldnt succeed. (over my head atm, and i cant do that much research for it since i figured out i would have to get into compiling). i also attempted to see if i could use alsa or something else - whatever that was processing sound in linux.

i wasnt able to succeed. so far, i am switching to linux when i want to listen to music for prolonged durations.

Ubuntu + pulseaudio + max settings + m-audio fast track pro = beyond i

Reply #33 –
In summary:

Through the use of a chain of DSP effects you do not fully understand you were able to create a sound while in Linux you were unable to recreate through the use of a different chain of DSP effects in Windows.
You prefer the sound of one chain over the other.








Could someone please explain what the surprise is?
Creature of habit.

Ubuntu + pulseaudio + max settings + m-audio fast track pro = beyond i

Reply #34 –
please read over the post you responded to, again. you havent understood what i wrote.

sound quality was much better in the later case WITHOUT any kind of dsps. EVEN when applied dsps onto, the sound quality in former case was not able to even come close to the latter.

when applied dsps to, the sound quality in latter case becomes worlds apart from the former case.

so, there has to be something that is differentiating non dsp'ed sound in latter case, from non dsp sound in case a, and from dsp'ed sound in case a. im wondering what it is.
 

Ubuntu + pulseaudio + max settings + m-audio fast track pro = beyond i

Reply #35 –
Invite a friend and see if he/she feels the same way.

Such a large difference should be easily detectable.