Inconsistent HDCD detection

[pangit]

Hi all,

I have quite a few HDCDs in my collection, and have been enthusiastically ripping them using the HDCD DSP effect to extract the maximum possible resolution from the disc. I have also discovered many albums in my collection with HDCD encoding that I wasn't aware of. I have turned on the "Track Technical" column in the CD ripper so I can see whether a disc has the encoding or not. I want to know before I rip, because I've been ripping all my HDCDs to a separate directory for easy accessibility during demos.

What I've discovered is that HDCD decoding detection is often very inconsistent. CD ripper has identified several of my discs with only a few tracks having HDCD encoding, but when I refresh the disc, many more (or often all) of the tracks are subsequently identified as HDCD. In case anyone wants to attempt to corroborate what I'm seeing, try:

Beach Boys, Pet Sounds
Ocean's 11 soundtrack

Both of these consistently show only a partial list of HDCD tracks, whereas if I refresh them they'll show all of the tracks as having the encoding. This got me thinking - I wonder if I've ripped discs that had HDCD encoding I didn't know about, that wasn't identified at all? I used the batch converter to sort through everything I've ripped by the "bit rate" column, and sure enough, I found several (one example was Beck's Mutations). I also checked the bitrate of a few discs I'd already ripped that showed inconsistent HDCD encoding, and the good news is that the DSP effect appears to always find the HDCD encoding (but see my last example below before being to reassured). For instance, when I first ripped Pet Sounds only the last several tracks were identified as HDCD, and I just accepted that, thinking that perhaps only the stereo tracks at the end were mastered with HDCD. Batch converter showed all of the ripped tracks, even the ones not identified as HDCD, as having the higher bitrate.

One disc I encountered (Reference Recordings 61CD, Turtle Creek Chorale's "Postcards") consistently showed everything but tracks 1, 7,8, and 11 as having HDCD encoding, no matter how many times I refreshed the disc. I can, however, play any of the four supposedly non-HDCD tracks in WMP11, and the HDCD logo always appears in the lower left corner. When I look at the bitrate of these same four tracks, it's the same as all the others I ripped from the disc, suggesting that HDCD encoding really was detected and processed by dBpoweramp.

One more anomaly: I have an album called "Thank you Shirl-ee May" by Shawn Amos. It doesn't have any HDCD identification on the packaging anywhere, but I've learned that many "stealth HDCD" discs don't. When I first load this in CD ripper, none of the tracks are identified as having HDCD. If I refresh the disc (no matter how many times), track 6 is shown as having HDCD encoding. However, if I look at the bitrate of the track rips, they all have the higher HDCD rate (2,116Kpbs). If I play this disc in WMP11, only tracks 6 and 9 are consistently identified as HDCD - no others. This disc has me really confused. I've read elsewhere in this forum from Spoon that HDCD is applied on a track-by-track basis, not by disc. Based on this, it would seem that only tracks six and nine on this disc should have the higher bitrate, but instead they all do. Anyone know what is going on here?

FYI - these are all pristine discs - no major scratches or other warts - ripped with a brand-new LG CDROM.

Bottom line: it seems to me that HDCD detection needs to be more robust and consistent. I'm left doubting the reliability of the detection (or even the decoding, based on my Shawn Amos disc).

I would also like to register my strongest possible vote for having something definitive about HDCD written into the disc rip log file. Currently, the log shows that the HDCD DSP effect is turned on, but not that the encoding was actually found to be present in any particular track. Something simple like "HDCD detected" for each track would be much appreciated. Others seem to have asked for something similar in a few posts I've read, so I think this feature would be a very worthwhile and popular enhancement.

[Spoon]

The detection for the technical column is on a limited frame count, now depending on the drive and offset it could affect the detection, where as the application of HDCD by teh DSP effect will allways correctly process the file as it uses the whole file.

[pangit]

Thanks for responding Spoon. I see your point. I think it would be cool to have an option the user could enable like "extended HDCD detection" or something like that, if they were willing to wait for a longer but more accurate detection. I know I would be.

This also reinforces my point (I think, anyway) about having some indication in the log file that HDCD was indeed detected and processed.

[Jeff Flowerday]

Thanks for responding Spoon. I see your point. I think it would be cool to have an option the user could enable like "extended HDCD detection" or something like that, if they were willing to wait for a longer but more accurate detection. I know I would be.

This also reinforces my point (I think, anyway) about having some indication in the log file that HDCD was indeed detected and processed.

Ultimately I store my files as wma lossless, but I rip to flac with the HDCD DSP on.

1) if all flac files created are 24bit, I convert the to wma lossless 24 bit
2) if the disc is partial HDCD I rerip directly to wma lossles 16 bit. I don't want 6db disparity between tracks in a album.
3) if the disc isn't HDCD i convert to wma lossless 16 bit.

Looking at technical data and or jewel is never completely accurate, looking at the resultant flac is.

Now if explorer dbpoweramp columns actually worked in vista, my life would be simple. Mouse over on each file is a real pain.