Stereo Tool
https://forums.stereotool.com/

Soundprocessing Freakday - NL
https://forums.stereotool.com/viewtopic.php?t=3055		 Page 7 of 16
Author:  hvz [ Fri Apr 01, 2011 10:28 pm ]
Post subject:  Re: Soundprocessing Freakday - NL
Quote:
Is there some way to record the MPX via VAC ? What for program is best to use?
I tried Adobe Audition but it keeps telling me that sample rate of hardware is 44.100 instead of 192 kHz.
@hans Are you also going to make an MPX recording for the processingtest for on the processing freakday site?
1 is a bit problematic. Stereo Tool outputs 176.4 kHz output - you could record that. Alternatively, you can upsample the input to 192 kHz and use that as input for Winamp.

The latter is the easiest (no need for VAC, sample rate conversion is done BEFORE MPX signal is present), but... Declipping won't work properly anymore (because after resampling you don't know anymore which samples are clipped and which are not).

I will add something for that (it's needed also for analog signals; basically in that case if a sample is clipped, the sample before and after should be treated as clipped as well).


I will also upload an MPX for the 'test' file (will start listening to it in a moment, and also waiting for your feedback...).
Author:  eldoradofm [ Fri Apr 01, 2011 10:54 pm ]
Post subject:  Re: Soundprocessing Freakday - NL
I've had to read your text a couple of times to understand what you mean. If i understand it correctly the output is at 176.4 kHz because the input file is at 44.1 kHz. So if i resample the input file to 192 kHz i also get 192 kHz output from the winamp dsp?
If so, would 48 kHz as input then also work for 192 kHz output?

Ah i figured out that i can use rewire from esi juli@ to record with adobe audition. So no VAC necessary then for that part.

I checked the latest preset with the test file. It's doing a good job. Sometimes i hear some distortion in the lows but i begin to doubt if it's not introduced by my transmitter.
Author:  Bojcha [ Fri Apr 01, 2011 10:56 pm ]
Post subject:  Re: Soundprocessing Freakday - NL
Quote:
I checked the latest preset with the test file. It's doing a good job. Sometimes i hear some distortion in the lows but i begin to doubt if it's not introduced by my transmitter.
Maybe, i have some DC on transmitter too.
Author:  hvz [ Fri Apr 01, 2011 11:12 pm ]
Post subject:  Re: Soundprocessing Freakday - NL
Best way to check that is using MPX Tool...... :shock:

@ eldoradofm: Yes, but: By resampling to 48 kHz you also loose the info about which samples are clipped and which are not.
Author:  eldoradofm [ Fri Apr 01, 2011 11:37 pm ]
Post subject:  Re: Soundprocessing Freakday - NL
Quote:
Best way to check that is using MPX Tool...... :shock:
The problem to measure the tilt is the following for me:
I can create a 30 Hz square wave with breakaway. First thing is to check if the output of the soundcard also produces a square wave.
This is possible with a scope with DC input. So maybe it's time to buy one. If the wave is tilted i can correct it in breakaway. I'm waiting for this feature in stereo tool
Then i can also calibrate the input tilt of the soundcard when i connect output to input and check it with MPX tool.
I can then also check if the wave doesn't get tilted if i connect the output to the transmitter.
Maybe it's easier to buy Marian Trace Alpha sound card which supposed to have DC straight output. I guess it's also DC straight input.

But now the problem for me. I have a tuner with MPX output but the MPX output is not great and i still have to make a circuit for better linear response. I also don't know what's the tilt generated on the MPX output.
So now how to check if the transmitter is DC straight or if i have to have more tilt adjusted in breakaway/future stereo tool?
I can also check lowest deviation with pira analyzer but don't know how accurate that one is.

Does anyone know if i can measure how much tilt is needed for the transmitter with a spectrum analyzer?
Author:  hvz [ Fri Apr 01, 2011 11:45 pm ]
Post subject:  Re: Soundprocessing Freakday - NL
Until last week I didn't even know what kind of tilt is caused by analog audio equipment. Apparently it's a constant phase shift for all frequencies (a partial Hilbert transform). Shouldn't be too difficult to add.

For now, you can use the "FM Calibration" filter and set ALL the phase rotation values to the same setting. This does cause extra latency and - if the tilt is big - some artifacts though.
Author:  Bojcha [ Sat Apr 02, 2011 12:47 am ]
Post subject:  Re: Soundprocessing Freakday - NL
Quote:
For now, you can use the "FM Calibration" filter and set ALL the phase rotation values to the same setting.
Exactly, That's practicaly same! And that's why i "many" times asked for 'Link' checkbox.
Quote:
This does cause extra latency and - if the tilt is big - some artifacts though.
True again, best way is to have no-tilt soundcard.
Anyway, i have Juli@. One day, I take my E-MU 0202 to transsmitter to see how much is that tilt actually.
So, thanks to Leif, i grabb his software, build all cables, calibrate soundcard. More here: http://www.claessonedwards.com/index.ph ... &Itemid=86

I measured tilt and i get every channel in "phase difference" is ideal to be @ 0.002499998. Small isn't it?
------

This adjustmeton INPUT will be very ussefull.
Then i can play 30 or 60Hz square tone on my SOURCE, and calibrate Soundcard input.
Author:  eldoradofm [ Sat Apr 02, 2011 1:04 am ]
Post subject:  Re: Soundprocessing Freakday - NL
Thanks for the link. It's a simple circuit to make, with that i can at least calibrate my sound card. Although a hardware oscilloscope will be more accurate i see in the description.
Author:  Bojcha [ Sat Apr 02, 2011 1:14 am ]
Post subject:  Re: Soundprocessing Freakday - NL
Two things are important:

- Calibration of soundcard (for measuring)
- And when measuring, MPX from soundcard must be attached to transmitter, so something like "T" BNC connector is needed.
Author:  hvz [ Sat Apr 02, 2011 1:45 am ]
Post subject:  Re: Soundprocessing Freakday - NL
Hm, interesting. Never though of that! That's indeed the simplest possible solution to do this - assuming (???) that the printer port output is completely constant. And I don't know if it is.

If the tilt values are that low, linking the sliders would work (on my own, really bad quality, transmitter, I'm getting changes to over 120 degrees between frequencies!).

But what I really need is to write a filter that does a tilt without adding latency (or at least adding very little), and without causing artifacts. That's why I mentioned the Hilbert transform - that's a 90 degrees shift.
Page 7 of 16 All times are UTC+02:00
Powered by phpBB® Forum Software © phpBB Limited
https://www.phpbb.com/