HDCD rescue - it's possible, but man...


I spent the better part of today sick, but working on finding a way to decode music I might have that is HDCD encoded.

As a refresher, HDCD was an enhanced CD format. In some ways the predecessor to MQA. HDCD was an engineer's toolbox, allowing the mastering engineer to select a number of features. This would then be decoded by a matching chip on a CD player or DAC. The most famous feature of HDCD was bit-compression. Getting a 24 bit signal encoded in a 16 bit music file.

As an aside, the Pacifics Microsonics AD converters were highly prized by engineers for their sound quality. Anyway, the format got bought by Microsoft and died.

Of the 670 CD's I have ripped only about 11 were HDCD encoded. But man, what a pain. I ripped everything to FLAC, but the HDCD decoder only does WAV. I had to download source, compile it, then write a script to go through every CD and decide if it's HDCD or not. Once found, I have to convert from FLAC (44/16) to WAV, decode the WAV file (now 24 bits) and convert back to FLAC to compress again.

The discovery process was pretty fast.  About 10 minutes to go through them all by cheating. :) More time was spent figuring out how to pass apostrophe's in file names than finding the files.  Nathalie Merchant was one author who consistently used HDCD by the way.
erik_squires
Hi @dtc

I found a very complete discussion of it here:

http://wiki.hydrogenaud.io/index.php?title=High_Definition_Compatible_Digital#Regarding_the_Playback...

According to that, the transient filters were part of the specification, but were never licenced and therefore never implemented.

So this leaves dynamic range expansion, a-la dbx, kind of. So it makes me wonder what good the RR HDCD encoding is doing? Expanding 24 bits to 28??

Best,

Erik
Following on the link, above, I found a very interesting section. It is true that PM equipment was pretty popular in recording and mastering studios. I quoteth the results:

Much of the time, the engineers turned all HDCD features off, but the HDCD control packets were still inserted into the discs anyway. For these discs, there is absolutely no benefit in decoding HDCD.

This explains why so many disks may be identified as HDCD even if not labeled as such. Chances are good the mastering engineers had no idea it was on, and therefore did not utilize any of it’s features. Meaning, HDCD is irrelevant for those disks. Though I imagine there’s probably a byte or two at the start of each track which gets the hidden HDCD markers removed. Otherwise they might as well be RedBook.

It's also interesting how they describe it as a scam, originally marketed with grandiose claims and very little information. I can think of at least one digital format that seems like that. Grandiose, unsubstantiated (to my ears) claims, but a ton of data about it. Not all scams are lacking in data. Some like the perpetual dark energy scams have tons of explanations. They're just not true.

Erik
Erik - You have to be a little careful of the hydrogenaudio comments. Many on there are so caught up on their methodologies and rules that I wonder sometimes whether they actual listen to music. I only have a few HDCD, but they are very good quality and I do believe the HDCD decoding improves the sound.

As to  disks that are not labeled as HDCD - no idea. Many of them may be a mistake, but it seems like an honest mistake since they were never labeled to be HDCD.  I just know the labelled ones I have I like.

RR is holding onto an old technology. WIth high rez available (including 24/44), I see no reason to continue with HDCD given the lack of proper decoders and the cheapness of disk space.
To me a HDCD disc, decoded using PMD100 or PMD200, far outperforms what MQA can do with the same disc without HDCD.

Cheers George