Welcome, Guest
Username: Password: Remember me

TOPIC: Converting sample rates

Converting sample rates 2 years 9 months ago #28602

  • denver8me4dinner
  • denver8me4dinner's Avatar
  • Offline
  • Gold Boarder
  • 3038564518
  • Posts: 226
  • Thank you received: 40
  • Karma: 5
Hi pos.
Sorry to bug you again , I had a question about converting sample rates.

So if the HD uses 96k and the correction I want to use I can't get done at 96k (2048)- is there ANY way I can make the correction in 44-48k and somehow convert it to 96k?

If I use a 44k file the 2x4HD reads it but changes the correction (as we talked previously about). Is there a way to export a file into some other FIR editor and convert it to 96k and just add zeros but make the appropriate changes to the filter so that the values are not changed?

For example: (none of these numbers I'm using were correct I can't remember but for sake of discussion I'll make some up ;) ) I made a 400hz Lin phase crossover at44k imported it into the mini HD and looked at the graph saw my correction was there and it worked but instead of being -6db down at 400 it was -6db down at 280 +/-. So I made a 600hz +/- filter at 44k and it turned out being what I wanted -6db at 400.

So I was just using the graph on the FIR screen in the HD and looking where the cursor said and had to make a good guess because I don't know the math to convert it.

Is it possible to do it that way or would the filter ring or have any bad effects? Any help much appreciated

Andrew
The administrator has disabled public write access.

Converting sample rates 2 years 9 months ago #28607

  • pos
  • pos's Avatar
  • Offline
  • Platinum Boarder
  • Posts: 345
  • Thank you received: 196
  • Karma: 39
Hi Andrew
denver8me4dinner wrote:
So if the HD uses 96k and the correction I want to use I can't get done at 96k (2048)- is there ANY way I can make the correction in 44-48k and somehow convert it to 96k?
No you can't: 96kHz is what the 2x4HD is running the convolution at, and there is no way around it.
If you generate a bin file with a different sampling rate then frequencies will only get shifted by the ratio between the two sampling rates.
Then if you adjust frequency points with the inverse ratio to get the correct ones then you get exactly the same result as if you had used the correct 96kHz sampling rate and original frequency points.
download the last version of rephase here: rephase.org
Last Edit: 2 years 9 months ago by pos.
The administrator has disabled public write access.
The following user(s) said Thank You: devteam

Converting sample rates 2 years 9 months ago #28634

  • denver8me4dinner
  • denver8me4dinner's Avatar
  • Offline
  • Gold Boarder
  • 3038564518
  • Posts: 226
  • Thank you received: 40
  • Karma: 5
pos wrote:
Hi Andrew
denver8me4dinner wrote:
So if the HD uses 96k and the correction I want to use I can't get done at 96k (2048)- is there ANY way I can make the correction in 44-48k and somehow convert it to 96k?
No you can't: 96kHz is what the 2x4HD is running the convolution at, and there is no way around it.
If you generate a bin file with a different sampling rate then frequencies will only get shifted by the ratio between the two sampling rates.
Then if you adjust frequency points with the inverse ratio to get the correct ones then you get exactly the same result as if you had used the correct 96kHz sampling rate and original frequency points.

Okay thank you pos ,
I get bit rate conversions mixed up with sample rate conversions. I see what you mean now.
So let me ask you this than. And some may not like this way to do it and it might not be the best way but is it possible to make a 2 part FIR filter.? Like , let's say I want to make a correction on sub I always run out taps,. So could I make the sub run the left channel out of the HD and loop back with a short RcA cable into the right channel input and do half of a correction on one of the 2048 tap banks and the other half of convolution on the other 2048 bank (abandoning two outputs)?

Yeah it would mean 2 trips through adc/dac , but just for a sub I don't think it will be as bad as I think. If it's possible to do it that way than maybe I can squeeze a LF crossover linearization and a box GD correction , because right now I can only do one or the other and not both corrections on one convolution, and no matter what I do I can't make a high order linear crossover below 60hz.

Any ideas on how I could pull that off :)
Andrew
Last Edit: 2 years 9 months ago by denver8me4dinner.
The administrator has disabled public write access.

Converting sample rates 2 years 9 months ago #28636

  • pos
  • pos's Avatar
  • Offline
  • Platinum Boarder
  • Posts: 345
  • Thank you received: 196
  • Karma: 39
Hi Andrew,
Yes, you certainly can. This will not let you correct as much as a single 4096 taps convolution, but still, if it gets the job done why not?
The double conversion should not be a problem with subs.
download the last version of rephase here: rephase.org
The administrator has disabled public write access.
The following user(s) said Thank You: devteam

Converting sample rates 2 years 9 months ago #28645

  • denver8me4dinner
  • denver8me4dinner's Avatar
  • Offline
  • Gold Boarder
  • 3038564518
  • Posts: 226
  • Thank you received: 40
  • Karma: 5
pos wrote:
Hi Andrew,
Yes, you certainly can. This will not let you correct as much as a single 4096 taps convolution, but still, if it gets the job done why not?
The double conversion should not be a problem with subs.

Pos,
That's great! Okay I can't wait to get started. So would I have to do double the delay offset d1+d2 or just whichever filter has a longer length?

Seems logical to just add them but just want to make sure I don't know if the 2nd correction will alter the time on the 1st in any way.
The administrator has disabled public write access.

Converting sample rates 2 years 9 months ago #28666

  • denver8me4dinner
  • denver8me4dinner's Avatar
  • Offline
  • Gold Boarder
  • 3038564518
  • Posts: 226
  • Thank you received: 40
  • Karma: 5
^^^^
Cont'd


I know it's the processing of the filter that the delay is manifested. So it still kinda messes with me as to what it is.

Again I can't know for sure. I'm just not smart enough yet to try to think it out.

If that right it would be the longer of the two filter offsets .... but what if... I just don't know :(
The administrator has disabled public write access.

Converting sample rates 2 years 9 months ago #28749

  • pos
  • pos's Avatar
  • Offline
  • Platinum Boarder
  • Posts: 345
  • Thank you received: 196
  • Karma: 39
Hi Andrew,

Sorry, I missed your question.
Delays will have to be added together, including ADC/DAC stages.
download the last version of rephase here: rephase.org
The administrator has disabled public write access.
The following user(s) said Thank You: devteam

Converting sample rates 2 years 8 months ago #29047

  • denver8me4dinner
  • denver8me4dinner's Avatar
  • Offline
  • Gold Boarder
  • 3038564518
  • Posts: 226
  • Thank you received: 40
  • Karma: 5
Sorry that was truly a dumb question...

I way over think that. :(

I tried it tho. It definitely works and I am now able to correct more than 23ms GD on sub and port. Quite the adventure getting it to sound right tho. It's almost as if the GD is needed for the port to work right or to have plentiful output.
It seems like there's some minor problems with sub harmonics now... it's a bit echoie sorta.

I'll keep playing with it .
The administrator has disabled public write access.
Moderators: devteam