AES 2007 – DSP-Based RIAA Filters (Attn: Vinyl Fans) - Part II

Andy W — Fri, 02 Nov 2007 22:24:00 GMT

For those of you waiting with bated breath (LOL), here is part II, finally…

The author presents an argument that if instead of normalization after truncation (see part I of this blog for an example of truncation), the digital samples are processed by a digitally implemented inverse-RIAA filter, and then carefully matching the gain, the residual error is greatly reduced. In fact, he claims that the magnitude of error is only on the order a few LSBs out of 24 bits, and he claims that this proves to be over four bits less than expected 6 or 7 bit loss.

I have no way of knowing what filtering or algorithm is applied in DSP so I can’t comment on his findings, however today’s floating point DSPs have incredible amounts of processing power, and the accuracy of the computations involved in not in question.

The other argument presented was based on the review of perhaps thousands of available vinyl recording, his own and others. He selected around 40 recordings that were singled out based on the amount of high frequency content. His theory is that if a recording has a high level of high frequency or treble content, the resolution of the bass frequencies would be the most impaired.

This stems from the fact that when doing a digital recording (or an analog to digital conversion), you will eventually run out of headroom, and any transient that exceeds 0dBFS (full-scale) will be clipped.

When doing an inverse-RIAA filter digitally it is imperative that the maximum level on the recording is known, otherwise you will either 1) clip the signal on the peaks if the level is too high or, 2) underutilize the available resolution if the level is too low.

By sampling these specially chosen recordings, and using peak responding, peak-hold audio analyzer software, he discovered that the peaks were lower that what would be expected based on the RIAA curve alone (which would imply that you would have several of resolution for free).

That is not strictly the case however, since the analog gain level must be set based on the peak value of the coherently summed signal of bass, mids, and treble, not just the peak value at discreet frequencies across the spectrum. But remembering that the mids and especially the bass are de-emphasized, and by analysis of the program peak level on the selected recodings, validated the use of the peak responding, peak-hold analysis, and that with LP recordings this is unlikely to exceed 1 bit (6dB).

His argument continues that RIAA equalization in the analog domain followed by digital conversion would be a reduction in fidelity since 1) the program peak level would now be defined by the recording’s bass level, 2) there would now be less resolution available for the treble, so the treble will be truncated instead of the bass, 3) humans have better hearing in the treble range, 4) treble doesn’t benefit from the reconstruction effects of the filter.

So there you go… If there are errors in getting his point across they are probably mine.

AES 2007 – DSP-Based RIAA Filters (Attn: Vinyl Fans) - Part I

Andy W — Fri, 12 Oct 2007 18:49:00 GMT

Last weekend I got to attend the 123^rd AES Convention in New York City. The one paper presentation that perhaps would have the most interest among some of our resident KlipscHeads was one presented by R. S. Robinson of Channel D Corp. entitled “Filter Reconstruction and Program Material Characteristics Mitigating Word Length Loss in Digital Signal Processing – Based Compensation Curves used for Playback of Analog Signals.” In layman’s terms that means “Don't worry about lost resolution when doing a RIAA filter using DSP.” For those interested in reading it for yourself it is Convention Paper 7185, but unless you have access to the AES online library, that will be difficult. After his presentation I asked him if this was available in some consumer product; it is… Virtual Vinyl has more information about it, and you can purchase the software to do this on your computer (if you have a Mac).

The author also said to me (which I was rather surprised to hear) that many people use the DSP filter during playback of the actual vinyl LP… I was rather surprised at this since I would have assumed that most vinyl fans would rather play the LP once for archival purposes (saving wear on the irreplaceable recording), and then listen to the archived digital file through the DSP… or at the very least transfer it to CD…

But I suppose that an extensive vinyl collection would take a lot of drive space… and the reason they are still listening to vinyl in the first place is that they believe vinyl is better than CDs.

I will not reprint the article here, but I will attempt to give a synopsis. I have not used this product and I am not endorsing it in any way. Neither do I pretend to be a DSP Engineer.

First a little something for those who don’t have a clue what I’m talking about…

Before making an LP master, the treble is boosted (~20dB at 20kHz) and the bass is cut (~20dB @ 20Hz). This is done to overcome limitations of the LP playback system (mainly the phono cartridge) and to pack more music onto a record (making al LP possible). A phono preamp is then used to 1) amplify the weak signal from the phono cartridge (a.k.a. the stylus, or needle) and 2) to undo the RIAA equalization (boost the bass and cut the treble).

Now back to the paper… The author’s intent is to apply DSP equalization for LP playback and archiving with the reasoning that the accuracy can exceed that of an analog design and avoids channel to channel variation. He also believes restoration is facilitated since the click and pop removal can be done before equalization.

A major consideration of any DSP processing is dynamic range and headroom. In the case of the LP you must consider a 20dB headroom allowance for the treble; likewise this implies the full digital dynamic range can’t be utilized for bass frequencies. Taken together (40dB overall) this accounts for six to seven bits of potential loss of resolution. The author claims that this analysis “omits the consideration of the reconstructive effects of the low frequency emphasis curve (digital filtering), plus the frequency balance of typical musical program material.” Hence the title of his paper “Filter Reconstruction and Program Material Characteristics Mitigating Word Length Loss…”

So what is word length loss? If I have a digital word (for simplicity I’ll use 16 bits) say, 1101101100111011. If I have the full dynamic range of a 16-bit ADC available I will have no losses, but if signal is attenuated (like in the RIAA curve) I can’t use all those bits. If I attenuate the signal 40dB (multiply by 0.01) I will end up with (forgive me if I make any mistakes in my binary math) 0000001000110001. If I simply apply scalar gain to this result I end up with 1101101100100100. Notice the error in the last 5 bits.

Perhaps it would be easier to see if I used a factor of 2…

Start with

1101101100111011 and divide by 128 you get…

0000000110110110 (simply stated this is word length loss),

and now multiply by 128, you get

1101101100000000. Compare that to the original…

1101101100111011. Notice the error in the last six bits.

This is the same thing that would happen if you sampled the output of an LP (where 20Hz is 40dB down from 20kHz) and then applied scalar gain to the low frequencies, you’re losing resolution in the low frequencies.

More to come next week…

Andy's Blog : RIAA

AES 2007 – DSP-Based RIAA Filters (Attn: Vinyl Fans) - Part II

AES 2007 – DSP-Based RIAA Filters (Attn: Vinyl Fans) - Part I