Head-Fi.org › Forums › Equipment Forums › Computer Audio › Linear Resampling for Winamp or Foobar
New Posts  All Forums:Forum Nav:

Linear Resampling for Winamp or Foobar - Page 4

post #46 of 74
8/27/07 at 1:41pm
Anyone else think this thread is just continuing due to some misconceptions b0dhi holds?
Reply

Head-Fi's Best Sellers

post #47 of 74
8/27/07 at 2:06pm
Thread Starter 
Hancoque, yeah, I agree with that.

Seidhepriest, looks like I won't have to code it. Check this out:


Top one is 5Khz, bottom is 8Khz. Not sure where those large notches are coming from, but it works fine on other frequencies I've tried.

The other weird thing is my DAC all of a sudden developed a ringing at 96Khz today (you can see it on the top graph). I can't get a DC signal out of it, whereas yesterday I could. Weird.

But yeah, apart from those two issues I'm happy with this upsampler. It retains decent square edges below 10Khz and keeps very smooth sine waves above that. The best of both worlds CPU usage is 50% though >_<

LawnGnome: Grow up.
Reply
post #48 of 74
8/27/07 at 2:09pm
Which upsampler?
Reply
post #49 of 74
8/27/07 at 2:19pm
Thread Starter 
http://otachan.com/out_asio(dll).html

SRC in Foobar should sound about the same though. My DAC must've had a brainfart or something, coz square waves are sharper today than they were yesterday, and that 96Khz component wasn't there yesterday or the day before either.

Thanks guys, problem solved
Reply
post #50 of 74
8/27/07 at 3:14pm
Quote:
Originally Posted by b0dhi View Post

Top one is 5Khz, bottom is 8Khz. Not sure where those large notches are coming from, but it works fine on other frequencies I've tried.

The other weird thing is my DAC all of a sudden developed a ringing at 96Khz today (you can see it on the top graph). I can't get a DC signal out of it, whereas yesterday I could. Weird.

But yeah, apart from those two issues I'm happy with this upsampler. It retains decent square edges below 10Khz and keeps very smooth sine waves above that. The best of both worlds CPU usage is 50% though >_<

LawnGnome: Grow up.
Why are you so focused on finding an upsampler that produces nice square waves? when it distorts everything else?

You are spending time to reduce the quality of the output, so you must have some misconception that this will help you in some way. Either that or you just want more distorted sound.
Reply
post #51 of 74
8/27/07 at 3:25pm
Thread Starter 
Quote:
Originally Posted by LawnGnome View Post
Why are you so focused on finding an upsampler that produces nice square waves? when it distorts everything else?

You are spending time to reduce the quality of the output, so you must have some misconception that this will help you in some way. Either that or you just want more distorted sound.
The only misconceptions being conceived here are yours. The upsampler you see in those graphs is a very high quality one, and if you actually read the post you quoted with your brain switched on, you'd know that it doesn't reduce the quality in any way. "It retains decent square edges below 10Khz and keeps very smooth sine waves above that. The best of both worlds."
Reply
post #52 of 74
8/27/07 at 3:29pm
But why don't the resamplers in Audition or other professional audio applications use linear interpolation if it is so much better? And why is a linear resampler generally labeled poor quality whereas other resamplers like SSRC or SRC "best sinc" are considered to be high quality? Explain that please.
Reply
post #53 of 74
8/27/07 at 3:36pm
Thread Starter 
Quote:
Originally Posted by Hancoque View Post
But why doesn't the resampler in Audition use linear interpolation if it is so much better? And why is a linear resampler generally labeled poor quality whereas other resamplers like SSRC or SRC "best sinc" are considered to be high quality? Explain that please.
That's a question you'll want to pose to whomever it was that stated linear resampling was objectively "so much better".
Reply
post #54 of 74
8/27/07 at 3:45pm
@OP: Do you know Sampling Theory?
http://en.wikipedia.org/wiki/Nyquist...mpling_theorem

That theorem says that any signal which does not contain frequencies >= 22050 Hz can be reconstructed _perfectly_ from a time series sampled at 44100 Hz. This is what makes digital media possible in the first place. For this reason the stuff on your CDs had been band-limited at about 20KHz before being sampled, so trying to tickle out 40KHz artifacts by linearly interpolating will give you no information which was present at recording time. You cannot get anything above 22050 KHz from your CDs, and the best resampler for this job is a sinc resampler.
Reply
post #55 of 74
8/27/07 at 3:55pm
Thread Starter 
"That theorem says that any signal which does not contain frequencies >= 22050 Hz can be reconstructed _perfectly_ from a time series sampled at 44100 Hz."

Hancoque already made that point earlier in the thread. Music, though, contains harmonics above 22050Hz, and electronic music would contain infinite harmonics if it weren't bandwidth limited by the reproduction medium.

If it wasn't clear in post #47, I decided on an upsampler that isnt linear. Let the thread die already.
Reply
post #56 of 74
8/27/07 at 4:21pm
Quote:
Originally Posted by b0dhi View Post
"That theorem says that any signal which does not contain frequencies >= 22050 Hz can be reconstructed _perfectly_ from a time series sampled at 44100 Hz."

Hancoque already made that point earlier in the thread. Music, though, contains harmonics above 22050Hz, and electronic music would contain infinite harmonics if it weren't bandwidth limited by the reproduction medium.

If it wasn't clear in post #47, I decided on an upsampler that isnt linear. Let the thread die already.
So you wanted to distort the rest of the spectrum so you could try to eek out harmonics which are probably much much higher than your range of hearing?

Makes great sense.
Reply
post #57 of 74
8/27/07 at 4:32pm
Thread Starter 
Quote:
Originally Posted by LawnGnome View Post
So you wanted to distort the rest of the spectrum so you could try to eek out harmonics which are probably much much higher than your range of hearing?

Makes great sense.
Yeah, ok.

For everyone else, I thought this was interesting: -

Quote:
It’s been determined that time delay differences of 15 microseconds between left and right ears are easily discernible by nearly anyone. That’s less than the time difference between two samples at 48kHz (about 20 microseconds). Using a single pulse, one microsecond in length as a source, some listeners can perceive time delay differences of as little as five microseconds between left and right. It is therefore, indicated that, in order to provide a system with exact accuracy concerning imaging and positioning, the individual samples should be less than five microseconds apart. At 96kHz (a popularly preferred sample rate) there is a 10.417-microsecond space between samples. At 192kHz sample rate there is a 5.208-microsecond space between samples. This reasoning suggests that a sample rate of 192kHz is probably a good choice.
http://digitalproducer.digitalmedian...le.jsp?id=7408
Reply
post #58 of 74
8/27/07 at 5:28pm
Quote:
Originally Posted by b0dhi View Post
Yeah, ok.

For everyone else, I thought this was interesting: -


http://digitalproducer.digitalmedian...le.jsp?id=7408
Yeah, and your point?

What does that have to do with you looking for a linear RS which introduces distortions instead of a superior algorithm?

Also, what does that excerpt prove?

Yes it is well known that a delay from one sound source over the other, will make the earlier source sound dominant.

However, where does that come into play here? The sampling rate won't effect channel balance or the likes. The logic in the author's argument is flawed.

Why would sample rate change time difference between the left and right channels? You can even test this yourself, use a resampler and greatly reduce the sample rate, if the authors logic is correct, one speaker would become more and more dominant sounding over the other.

However, this is not the change, the channel balance/time difference does NOT change.
Reply
post #59 of 74
8/27/07 at 5:39pm
Thread Starter 
There's a reason I said:
Quote:
Originally Posted by b0dhi View Post
For everyone else...
I don't have time to sit here and educate someone with no tact, no knowledge and an overabundance of arrogance. Stop threadcrapping.
Reply
post #60 of 74
8/27/07 at 7:04pm
Quote:
Originally Posted by b0dhi View Post
There's a reason I said:


I don't have time to sit here and educate someone with no tact, no knowledge and an overabundance of arrogance. Stop threadcrapping.
Your entire thread is a threadcrap to this forum.

You post a thread asking for advice, and then ignore everything.

Maybe you should try posting some ideas that you can actually explain your reasoning behind. Because throughout this entire thread, you have been unable to do so. And your ideas have been illogical and ridiculous.

You having this need to use a linear resampler because you think it will bring out extra harmonics about 22khz, even though you most likely can't even hear above 18khz, and more importantly, it would distort the music a fair degree. Also, not to mention CD's often have highpass filters at 20khz. It shows your logic all too well.
Reply
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Computer Audio

Head-Fi's Best Sellers

Head-Fi.org › Forums › Equipment Forums › Computer Audio › Linear Resampling for Winamp or Foobar