What about digital RIAA?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Disabled Account
Joined 2004
Last edited:
Disabled Account
Joined 2004
If you haven't already gathered, I'm not really concerned with winning friends;).

"Influencing people"? Don't know - I just tell the truth as I see it, and I'll listen to reasonable philosophy from others.

Zealots and fanatics I have no time for.

I see math and science are for zealots and fanatics, I guess because they don't have enough "wiggle" room for nonsense.
 
  • Thank You
Reactions: 1 user
Even people with hearing limited to 10KHz or less will notice clearer, more natural and generally more realistic instruments and voices.
published peer reviewed cites please?

for the thread subject of digital RIAA we expect that people will be using 24/96 ADC setting on their soundcards
and most of audio ADC will today be high oversampling ratio delta-sigma converters

so the audio frequency phase shift argument is moot even before looking for controlled listening test hearing evidence


and worrying about digital FIR filter's Gibbs pre and post ringing of analog source from a phono cart playing a vinyl recording is hilarious to any knowing the dynamic limitations of the media
the only place you could possibly expect see such effects would be from mistracking/clipping - never from the recorded audio signal that the medium and transduction can handle

the RIAA DSP EQ can and should be done at the 96k capture frequency, keep playback at the high sample rate and again there's nothing to argue over - you can't hear the ADC's 40+kHz Gibb's which is never remotely excited by signal content from a phonocart playback of recorded audio

only when bandlimiting in the decimation for putting the EQ'd output on a 16/44 CD of your own is there any reason to even look at the (non)issue

still no generally accepted evidence for even 22k ringing audibility with modern "hot" close miced digital mastering of CD 16/44

audio high frequency content/signal slew rate is more limited by passing through a vinyl recording/playback chain
 
Last edited:
Or backing up extraordinary assertions with actual data or references.

Sy,

I don't want to 'dis' you, but this really is a discussion bordering on the "philosophical".

You strike me as someone who possibly demands ABX or 'double-blind' discrimination of audio recordings.

This will NEVER provide answers, because listening to music builds over minutes and hours of listening.

An ABX/double-blind is "taking a test".

I vividly remember when I was laying on my back, listening to a rip of the AAD CD of Suzanne Vega's eponymous album, in 2003.

I've had this album around since its original release, when I played it on my Systemdek IIX/Rega/Nagoaka back in the late 1980's.

Like the rest of my CD's, I had "archived" it as 320Kbs LAME, because my 250GB HDD seemed to mandate using this codec to store my music collection.

I was listening via the SPDIF-out from an M-audio soundcard to an Audio-Alchemy 'DAC-in-the-Box', a Creek OBH11 headphone amp and the HD580's I still use to this day.

I was several tracks in when I was jolted awake by ..... something not right.
 
Disabled Account
Joined 2004
A quick exercise, 30 1/3 octave multitone with RIAA pre-emphasis. Run through SoX with a simple ~5000 point FIR RIAA de-emphasis. Let's put to bed the pro vs home software paradigm and any questions that digital RIAA taxes any limitations of the software. Vertical is dB, horizontal is linear frequency. BTW the accuracy is only around .0002dB across the band due in part to a little truncation of the impulse response.

EDIT-24 bit 96kHz TPDF dithered no noise shaping.
 

Attachments

  • riaa.jpg
    riaa.jpg
    69.8 KB · Views: 196
Last edited:
Disabled Account
Joined 2004
Is that all it takes? Declare victory?

Any insights on steep (or even brick-wall) filtering and how it affects the audio band?

(shakes head).

I gave you a reference to Doug Pries' work (over 30 Yr. ago) as usual evidence to contrary beliefs is ignored. The red flag OTOH is for the denial that the sampling theory is valid. You can deny first principles that is your prerogative but do it with like minded folks.
 
'band limiting', and in particular the VERY steep low-pass filtering needed with 16-44.1/48 KHz, whether implemented in the analogue or digital 'domains', is impossible to achieve transparently.

I would agree but probably not for the reason you'll be hoping for. Implementing band limiting when there's significant OOB content (otherwise why bother?) gives an improvement in transparency. The reason for this isn't any kind of rocket science - with reduced HF energy the IMD performance of anything downstream is less crucial.

I will concede there's a penalty in phase response, overshoot and ringing on HF transients but subjectively I've not found those issues a problem. The upside outweighs the down subjectively when the filter's put after a NOS DAC (which has high levels of OOB output).

I believe Ricardo (kgrlee) has even done conclusive DBTs which produced the same result - the aurum pinnae go for the bandlimited version every time.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.