CD Pre-emphasis: How to detect? What software to rip? Tagging?

[miguelito]

Hi all,

 

@Archimagopost about pre/de-emphasis (https://archimago.blogspot.com/2020/09/how-to-cd-pre-emphasis-and-dealing-with.html) got me thinking about this...

 

I found a Roon-based filter that will do the de-emphasis: https://community.roonlabs.com/t/de-emphasis-curve-was-parametric-eq-functions/123336/11?u=miguelito

 

However, a few questions remain:

 

1- XLD seems to support detection of pre-emphasis in both the subcode and TOC (https://wiki.hydrogenaud.io/index.php?title=Comparison_of_CD_rippers). But how is this then used? Does it write a tag to the ripped file to indicate this? 

 

2- Apparently not all CD drives support extraction that would allow you to detect pre-emphasis in the subcode. How do I find out if my CD drives support this?

 

Thx!

[yamamoto2002]

I have only one pre-emphasis CD, a harpsichord Bach BWV 1080. It is recorded on 1980s.

 

When ripped with Exact Audio Copy, there is pre-emphasis flag is set on its CUE sheet. ripped PCM data is not de-emphasised: raw PCM is extracted to WAV or FLAC. When listened to the PCM without de-emphasis, high frequency is super harsh, with constant high frequency hiss and I cannot bear to listen to it.

 

It must be noticeable the sound is very wrong when you have a pre-emphasis CD and listen to it without de-emphasis.

 

Do you have any pre-emphasis CD ?

 

 

 

[miguelito]

Pretty sure I do. Some of the 80s releases are thin sounding and I always assumed they were “mastered for walkman” so that there was not that much bass, but I am starting to think that it’s probably pre-emphasis also.

[yamamoto2002]

If those thin sounding CDs sound normal with conventional CD player, it is definitely pre-emphasis CD.

 

Then it is possible for your USB CD drive to check if it support pre-emphasis detection or not.

with Exact Audio Copy, create cue sheet, open the cue file with a text editor and find "FLAGS PRE"

 

I attached the cue sheet of my pre-emphasis CD.

 

Quote

REM GENRE Classical
REM DATE 1985
REM DISCID 7F0B370A
REM COMMENT "ExactAudioCopy v1.0b3"
PERFORMER "Jean-Sebastien Bach"
TITLE "L'art de la fugue - CD 2"
FILE "CD.wav" WAVE
  TRACK 01 AUDIO
    TITLE "2 fugues miroir - Contrapunctus 12 a 4 voix - rectus"
    PERFORMER "Jean-Sebastien Bach"
    FLAGS PRE
    INDEX 00 00:00:00
    INDEX 01 00:00:32
  TRACK 02 AUDIO
    TITLE "2 fugues miroir - Contrapunctus 12 a 4 voix - inversus"
    PERFORMER "Jean-Sebastien Bach"
    FLAGS PRE
    INDEX 00 03:31:72
    INDEX 01 03:32:22
  TRACK 03 AUDIO
    TITLE "2 fugues miroir - Contrapunctus 13 a 3 voix - rectus"
    PERFORMER "Jean-Sebastien Bach"
    FLAGS PRE
    INDEX 00 06:59:05
    INDEX 01 07:10:37
  TRACK 04 AUDIO
    TITLE "2 fugues miroir - Contrapunctus 13 a 3 voix - inversus"
    PERFORMER "Jean-Sebastien Bach"
    FLAGS PRE
    INDEX 01 09:38:52
  TRACK 05 AUDIO
    TITLE "Fugue inachevee - Contrapunctus 14 a 4 voix,  fuga a 3 soggetti"
    PERFORMER "Jean-Sebastien Bach"
    FLAGS PRE
    INDEX 00 12:15:65
    INDEX 01 12:29:67
  TRACK 06 AUDIO
    TITLE "4 canons - Canon in Hypodiapason - Canon alla ottava, a 2 voix, theme inversu..."
    PERFORMER "Jean-Sebastien Bach"
    FLAGS PRE
    INDEX 00 21:53:67
    INDEX 01 21:57:42
  TRACK 07 AUDIO
    TITLE "4 canons - Canon alla decima, contrapunto alla terza, a 2 voix, theme inversus"
    PERFORMER "Jean-Sebastien Bach"
    FLAGS PRE
    INDEX 00 24:30:65
    INDEX 01 24:34:12
  TRACK 08 AUDIO
    TITLE "4 canons - Canon alla duodecima, in contrapunto alla quinta, a 2 voix, theme ..."
    PERFORMER "Jean-Sebastien Bach"
    FLAGS PRE
    INDEX 00 29:30:57
    INDEX 01 29:32:05
  TRACK 09 AUDIO
    TITLE "4 canons - Canon (in Hypodiatessaron), per augmentationem in contrario motu a..."
    PERFORMER "Jean-Sebastien Bach"
    FLAGS PRE
    INDEX 00 31:42:67
    INDEX 01 31:45:60
  TRACK 10 AUDIO
    TITLE "Fugue inachevee completee - Contrapunctus 14 a 4 voix, 4 themes (3e nom BACH ..."
    PERFORMER "Jean-Sebastien Bach"
    FLAGS PRE
    INDEX 00 36:58:12
    INDEX 01 37:07:12

 

preemphasis_example_cuesheet.txt

 

[semente]

Maybe this old topic can be of use:

 

 

[Miska]

At least HQPlayer detects it and applies the needed correction too. If Pro is used to rip the CD, then the output is with the correction automatically.

 

[audiofool]

Cool!  How does HQPlayer determine it needs pre-emphasis?  Is there a way to set a tag in a FLAC file to force it?   Does HQPlayer do anything with HDCD?  I am currently using ffmpeg to deal with HDCD.

 

[Miska]

3 hours ago, audiofool said:

Cool!  How does HQPlayer determine it needs pre-emphasis?

 

It is from the CDDA data.

 

3 hours ago, audiofool said:

Is there a way to set a tag in a FLAC file to force it?

 

No, once the data ends up in FLAC it should be already corrected without need for such measures.

 

3 hours ago, audiofool said:

Does HQPlayer do anything with HDCD?

 

No.... Only standard CDDA (RedBook), not any non-standard additions like HDCD or MQA.

 

But you won't find CDs that would have HDCD and pre-emphasis.

 

[audiofool]

39 minutes ago, Miska said:

 

It is from the CDDA data.

 

No, once the data ends up in FLAC it should be already corrected without need for such measures.

 

So even CD's I've ripped years ago should be handled correctly by HQPlayer for pre-emphasis since it is embedded in the CDDA data and not depending on the TOC?  Is this a newer feature of HQPlayer?

 

41 minutes ago, Miska said:

But you won't find CDs that would have HDCD and pre-emphasis.

 

Good to know, I didn't know they were mutually exclusive.

 

43 minutes ago, Miska said:

 

No.... Only standard CDDA (RedBook), not any non-standard additions like HDCD or MQA.

 

I have a number of CD's that are HDCD.  I am using music forums to find them since I don't believe there is any way to know for sure from the CDDA data stream?

 

[Miska]

2 hours ago, audiofool said:

So even CD's I've ripped years ago should be handled correctly by HQPlayer for pre-emphasis since it is embedded in the CDDA data and not depending on the TOC?  Is this a newer feature of HQPlayer?

 

No, you need to use the actual physical media. Already made rips are lost case.

 

2 hours ago, audiofool said:

I have a number of CD's that are HDCD.  I am using music forums to find them since I don't believe there is any way to know for sure from the CDDA data stream?

 

I have three CD's with HDCD. A HDCD decoder will detect from the data stream. I used dbPowerAmp to rip and decode those three to 24-bit FLAC. Note that the check box in the decoder is misleading. It needs to be unset, or otherwise the decoded result will be just clipping harder.

 

I think Microsoft owns HDCD these days. But there are so few discs with it that it is not worth trying to talk to Microsoft and potentially increase license price with decoding support.

 

[audiofool]

1 hour ago, Miska said:

 

No, you need to use the actual physical media. Already made rips are lost case.

 

 

For the lost cases, what about a tag in the flac file that HQPlayer could recognize? ie. preemp=yes

 

1 hour ago, Miska said:

I think Microsoft owns HDCD these days. But there are so few discs with it that it is not worth trying to talk to Microsoft and potentially increase license price with decoding support.

 

I have at least 16 discs and am using ffmpeg to decode, interesting that an HDCD decoder can detect from the data stream.  I'm curious how ffmpeg is avoiding the license fee.

[Miska]

Just now, audiofool said:

For the lost cases, what about a tag in the flac file that HQPlayer could recognize? ie. preemp=yes

 

Technically possible, but just not implemented right now.

 

1 minute ago, audiofool said:

I have at least 16 discs and am using ffmpeg to decode, interesting that an HDCD decoder can detect from the data stream.

 

CD players with HDCD support do it the same way. The watch the least significant bit (LSB) for the encoded pattern. Normally this bit would be used for dither, but on HDCD it is used for control data.

 

2 minutes ago, audiofool said:

I'm curious how ffmpeg is avoiding the license fee.

 

Cutting some corners in this explanation... ffmpeg has a lot of such stuff. It is free and open source, nobody is doing it commercially, so it doesn't make any money someone could sue for "damages".

 

[audiofool]

2 hours ago, Miska said:

 

Technically possible, but just not implemented right now.

 

 

CD players with HDCD support do it the same way. The watch the least significant bit (LSB) for the encoded pattern. Normally this bit would be used for dither, but on HDCD it is used for control data.

 

 

Cutting some corners in this explanation... ffmpeg has a lot of such stuff. It is free and open source, nobody is doing it commercially, so it doesn't make any money someone could sue for "damages".

 

Thanks Miska,

For those interested here is a link to using ffmpeg to detect and convert HDCD, though the site claims the patent is expired it doesn't mention Microsoft's involvement so I suspect licensing is still involved.

https://opensource.com/article/20/7/hdcd

 

Also a link to another opensource library for HDCD

https://github.com/bp0/libhdcd

 

 

[miguelito]

Funny enough, I've encoded about 1500 CDs and I cannot find one with pre-emphasis!

[FIndingit]

Classical cds from the late 80s, early 90s used pre-emphasis, especially Harmonia Mundi France, BIS. 

[miguelito]

A brief recap on how I am going to deal with pre-emphasis:

 

The fool-proof way to determine if a CD has pre-emphasis is at rip time (via the cue sheet - I don't think there are other ways). If this indicates pre-emphasis then run the rip through the appropriate de-emphasis fiter (sox everywhere or xACT on the mac). However, I have thousands of ripped CDs (and have them all in storage) and I am not going to bother to re-rip just to be sure.

 

The next most accurate method for CDs you've already ripped is to capture the first track from a streaming service (Qobuz or TIDAL), calc the frequency spectrum (eg with MusicScope on the mac) and compare that to your rip's spectrum. Normalize by considering the difference around 200Hz, and if the 16KHz region of the spectrum is roughly 9dB down in the capture vs your rip, then your rip has pre-emphasis. This assumes of course that streaming services dealt with pre-emphasis to begin with, which I think is NOT the case across the board (see side note below).

 

Finally the simplest method - and the least accurate one - is to play your rip and compare the sound to the streamed version. If your  rip sounds bright and harsh in comparison, then you probably want to try a de-emphasized version. This again assumes that the streaming service file is properly de-emphasized in the digital master in the first place.

 

If you use Roon, you can also create a simple high shelf filter based on the post linked below and toggle that filter on and off and determine which sounds more correct.

 

As a side note, there's one CD that I always thought was bright and harsh - Vivaldi's Stabat Matter feat Andreas Scholl (Harmonia Mundi). The streaming version sounds the same as my rip, but if in Roon I engage the de-emphasis filter, it sounds a lot more natural to my ears.

 

De-Emphasis in Roon:

https://community.roonlabs.com/t/de-emphasis-curve-was-parametric-eq-functions/123336/22?u=miguelito

[kilroy]

I have ripped CD flac files with pre-emphasis kept in the ripped file. Is there an easy way to correct this? (I do have audio editing software if that can easily be used.)

 

Also have a couple pre-emphasis CDs. I've tried to use foobar with a plugin to remove it in a rip but never figured out how to use it right. Playing with plugins and command line stuff is not really my thing, maybe there's an easy way I haven't figured out yet.

 

I prefer EAC for ripping but haven't seen a way to remove the pre-emphasis with it. Is it even possible?

[miguelito]

You can use sox on Windows.

 

Take a look at:

 

https://archimago.blogspot.com/2020/09/how-to-cd-pre-emphasis-and-dealing-with.html