@Amir: request for testing DAC's pre/de-emphasis support

[little-endian]

Hi Amir,

although it has been discussed here 1, 2 and there, it is undeniable that supporting to deemphasise preemphasised sources for DACs seems to remain a niche topic and annoyingly enough, many, such as the otherwise great Toppings, don't seem to support it.

Neither you so far unfortunately test any DAC for its support in that regard. Given that most devices end up in either the audio performance being above or way above human hearing capabilities, it is however a feature of still way higher magnificence compared to ranting about the last SINAD digit.

Which shows the irony as the public interest of the last point which makes zero difference anyway, is a lot higher than the lack of a feature which when in doubt lets your recordings sound overly bright.

So in short: preemphasis is part of the CD-standard (and even for LaserDiscs!) and thus also S/PDIF and to include a test routine for that shouldn't be a too much of work once established. I guess manufacturers would start to actually give a shit then once Amir says so.

So far, unfortunately only RME seems to do it entirely the right way. Not only do their DACs support it (at least for 44.1/48 kHz), but they also allow to manually enable the deemphasis circuit for sources not aware such as PCs (unfortunately, the PC-support of preemphasised material is also half-assed at best in form of cuesheets).

Please start to test for such details!

Thanks a lot in advance for your consideration.

 

[AnalogSteph]

At this stage, you're arguably best off taking care of de-emphasis yourself at or just after ripping or alternatively at audio player level on playback. Emphasis CDs are rare enough that a bit of extra manual effort shouldn't be too much of an issue (they make up about 1% of my collection [1]). I'm tagging the rip as required by the foo_dsp_deemph DSP and have it transcoded with that enabled into 24-bit FLACs using a converter preset, then remove the tags. I used to have the foo_deemph postprocessor installed but that could get a bit annoying.

Still, a DAC that accepts S/P-DIF and has a DAC chip that supports de-emphasis should also implement it (ideally this would be part of compliance testing at the manufacturer). There may not be too many left who want to connect a CD player directly and play 30+-year-old CDs with pre-emphasis, but those folks have no way of accommodating the issue any other way and may be caught unaware.

In a computer-based setup, there is absolutely no point in throwing the entire DAC into de-emphasis mode ever, hence why I'm pretty sure it's not even a thing with USB audio. This becomes particularly obvious with shared mode output, where you might have multiple audio sources going at the same time.

 

[little-endian]

Thanks a lot for sharing your thoughts.

At this stage, you're arguably best off taking care of de-emphasis yourself at or just after ripping or alternatively at audio player level on playback.

Well, that goes in the direction of acknowledging that CDs (and LaserDiscs in particular) are entirely outdated from a storage media's perspective and thus should be backed up anyway to a better accessible storage system, no doubt. Still, such physical media just is damn cool and nostalgic.

One advice I can give for the ones facing a similar issue: some players apparently are also able to deemphasise the CD data by themselves in addition to outputting it unaltered and flag it accordingly as noted here previously.

An example of such a player is the Panasonic DVD-S47. By default, it uses the deemphasis for the analog circuit and outputs the data via S/PDIF without any changes together with the appropriate flag for a DAC to perform the deemphasis. Exactly as it should so far. However, when using the otherwise pretty useless "remaster" feature (as no one hears any difference anyway), it actually starts to deemphasise the CD audio and outputs that via S/PDIF without the flag so even non-standard conform DACs play it back correctly.

I discovered that in conjunction with the Topping DX3 Pro+ and EX5 which both don't support preemphasis.

Emphasis CDs are rare enough that a bit of extra manual effort shouldn't be too much of an issue (they make up about 1% of my collection [1]).

Indeed, however what is rankling is that one can't be sure right away without analyzing the ... Q-subchannel entirely if I remember correctly. With a standalone player - it just works and would continue doing so if DACs would always respect the S/PDIF - flag.

Still, a DAC that accepts S/P-DIF and has a DAC chip that supports de-emphasis should also implement it (ideally this would be part of compliance testing at the manufacturer).

That's the exactly my point.

There may not be too many left who want to connect a CD player directly and play 30+-year-old CDs with pre-emphasis, but those folks have no way of accommodating the issue any other way and may be caught unaware.

True, I'm for sure part of the exception who is still filled with pleasure, putting in a physical copy of Michael Jackson's Thriller (JP release) into the player, being astonished how damn good the audio quality was - mind you 42(!) years ago and how it still sounds the same. Well, nominally as although the error corrected data literally hasn't changed one bit, the ears certainly have and will do unfortunately.

 

[amirm]

How many such CDs were produced? I remember only one from Pink Floyd that I bought circa 1983 or 1984.

 

[little-endian]

I think most if not all of Sony's early 35DP/35DC/35DH ("P", "C", "H" and so on depending on the genre; the "35" seems to have been derived from the 3500¥ price point back in the days). But also some other releases (Lionel Richie's "Can't slow down" early US release comes to mind which by exception has the preemphasis fact also listed in the TOC).

Surprisingly, apparently also some LaserDiscs such as Madonna's "Virgin Tour" feature it.

For some releases, the preemphasis is mentioned in the Loudness War Database.

So yes, it is scarse nowadays but part of the CD standard so any DAC with a S/PDIF input should support it in my opinion.

 

[AnalogSteph]

How many such CDs were produced? I remember only one from Pink Floyd that I bought circa 1983 or 1984.

There have been multiple attempts at complete lists over the years, the latest at SHF. I downloaded the table (what kind of bleeping bleep puts a massive table inside a word processor document instead of using a spreadsheet?! It's slow AF) and inserted a column for count, which got to 624 releases. I know of at least 3 more (France Gall - Babacar, ZELDA - 空色帽子の日 and Tomoko Aran - More Relax, see link in my previous post). Presumably digging through '80s Japan and France more thoroughly would unearth a bunch more.

IME emphasis CDs aren't something you come across all the time but neither are they rare. RYM says I only have 166 '80s releases, so for a good half a dozen of these to be with PE is quite significant.

 

[amirm]

How do we get a test signal for it and verify operation?

 

[Old_School_Brad]

Is this some sort of mastering technique? What are its advantages?

 

[tw99]

Honestly, the number of discs produced with pre-emphasis was tiny. Just Google it, there seem to be low hundreds at most, and all discs from the dawn of the CD era, mostly Japanese made. It is just not an issue worth considering today.

 

[amirm]

Is this some sort of mastering technique? What are its advantages?

It increases the dynamic range of CD.

 

[restorer-john]

How do we get a test signal for it and verify operation?

I have several such test tracks on my various test CDs. Let me know if you want any of the files.

CBS-CD1:

Denon test CD does pre-emphasis on and off separate sweeps for each channel.

Pierre Verany Test Set:

The Philips special test set (3 CDs) does not have pre-emphasis on/off tests.

 

[Chester]

only RME seems to do it entirely the right way

Chord DACs also.

 

[audio_tony]

There have been multiple attempts at complete lists over the years, the latest at SHF. I downloaded the table (what kind of bleeping bleep puts a massive table inside a word processor document instead of using a spreadsheet?! It's slow AF) and inserted a column for count, which got to 624 releases. I know of at least 3 more (France Gall - Babacar, ZELDA - 空色帽子の日 and Tomoko Aran - More Relax, see link in my previous post). Presumably digging through '80s Japan and France more thoroughly would unearth a bunch more.

IME emphasis CDs aren't something you come across all the time but neither are they rare. RYM says I only have 166 '80s releases, so for a good half a dozen of these to be with PE is quite significant.

I converted it to an Excel spreadsheet.

There are a couple of empty rows, but at least it's much faster and more accessible now.

It's a zip file as the forum doesn't accept .xlsx uploads.

 

[restorer-john]

I converted it to an Excel spreadsheet.

There are a couple of empty rows, but at least it's much faster and more accessible now.

It's a zip file as the forum doesn't accept .xlsx uploads.

I have a significant (many) number of listed, original discs in that list.

Some of my players have relay operated analogue de-emphasis and it can easily be heard tripping in once the disc starts.

 

[little-endian]

It is just not an issue worth considering today.

I obviously disagree. As tiny the amount might be, nostalgic purists including me still like to play back the original releases and when collecting CDs from that early area, the likelihood to end up with one with preemphasis is certainly higher than for the average consumer. How silly it is to have some super-duper DAC which doesn't play back that correctly then?

The biggest argument however is that we are making ourselves crazy about really ridiculous low levels of distortion, fractions of dB derivation in the frequency range which no one is going to hear anyway, discussing that back and forth even on this forum which should be the non-voodoo alternative draft to the usual head-fi & Co. forums and yet, it remains unknown which DAC follows the CD-standard if S/PDIF is used and that acutally makes a major difference.

The irony is staggering and should be fixed. Plus again, it wouldn't be a lot of work to test for proper deemphasis in the future whenever already measuring the devices anyway.

 

[restorer-john]

I obviously disagree. As tiny the amount might be, nostalgic purists including me still like to play back the original releases and when collecting CDs from that early area, the likelihood to end up with one with preemphasis is certainly higher than for the average consumer. How silly it is to have some super-duper DAC which doesn't play back that correctly then?

The biggest argument however is that we are making ourselves crazy about really ridiculous low levels of distortion, fractions of dB derivation in the frequency range which no one is going to hear anyway, discussing that back and forth even on this forum which should be the non-voodoo alternative draft to the usual head-fi & Co. forums and yet, it remains unknown which DAC follows the CD-standard if S/PDIF is used and that acutally makes a major difference.

The irony is staggering and should be fixed. Plus again, it wouldn't be a lot of work to test for proper deemphasis in the future whenever already measuring the devices anyway.

All compact disc players built during the halcyon years of the format have proper de-emphasis circuitry built in. It was either done in analogue, or later, in the digital filter itself.

And yes, a large amount of early compact discs have pre-emphasis with the appropriate flags to trigger the correct de-emphasis. And incorrect filtering, or the lack of de-emphasis makes a disc virtually unlistenable in some modern "players".

I have no idea how the various software players handle correct de-emphasis and whether the de-emphasis flag even survives certain ripping/container/delivery formats for digital audio. Roon for example had a whole lot of issues with de-emphasis. But that was 2019, maybe it's fixed now?

and this in the same thread. It would be really funny if people have been listening to musical content with no de-emphasis and not knowing it. If so, they should hand in their audiophile badge immediately. LOL.

Foobar had a plug-in? The rest- who knows.

 

[audio_tony]

The sox audio utility can provide de-emphasis.

"deemph Apply Compact Disc (IEC 60908) de-emphasis (a treble attenuation shelving filter)"

Something like this: sox test.flac -b 32 test2.wav deemph

 

[AnalogSteph]

How do we get a test signal for it and verify operation?

You've got an AP. I would be very surprised if it didn't have some way of just setting the PE flag in S/P-DIF output (the effect on frequency response would be immediately obvious) or even had a complete routine for verifying frequency response flatness in this mode by applying PE to the recorded side digitally.

 

[little-endian]

Chord DACs also.

So do they also allow do set it manually if the source lacks the proper signaling capability?

When it comes to the application of proper deemphasis of sources via S/PDIF which are flagged accordingly, the Benchmark DACs also do it right (in contrast to topping).

I have no idea how the various software players handle correct de-emphasis

Also pretty badly and neither I am sure whether 3,5"-drives pass along the information. Essentially, there are two modes such a drive can operate:

a) like a CD player using single speed reading and delivering the audio to its analog and optionally digital S/PDIF output. Here, it's excepted to act just like a standalone CD-player, deemphasising on the analog output and flagging it accordingly on the digital one, leaving the original data unaltered for the DAC to do the rest.

b) in DAE mode, often at way higher reading speeds, delivering the data to the PATA or SATA output, more or less including the subchannel data and that's often the crux.

[...]and whether the de-emphasis flag even survives certain ripping/container/delivery formats for digital audio.

Rarely as also badly implemented on PC systems till today which will probably never change anymore due to its importance decline.

As far as I'm aware, the only way is via an additional cuesheet which refers to the raw PCM files or the images in the CloneCD - style with their additional .sub file containing the PQ- or PW-subchannel data.

Since many original CDs with preemphasis don't list that fact in the TOC properly as mentioned before, odd things occur when one rips such CD with e.g. CloneCD: the cuesheet for e.g. Michael Jackson - Thriller will only state the "FLAGS PRE" statement for some of the tracks although the album uses it throughout entirely. If one would burn such an image only based on the cuesheet, the result would be only correct for those tracks. Using the entire CloneCD image with the .sub however leads to a correct result. It's really weird and since programs such as EAC only check the TOC, most actually preemphasised CDs will claim "no" there.

So if possible, Amir should test DACs for the following, taking RME as reference:

a) are S/PDIF sources properly deemphasised when flagged so?
b) do they allow to manually enable deemphasis for incompatible sources?
c) if preemphasis is supported, up to what sample rate? (guess this would be optional for really enthusiastic testing as although such sources could exist, I'm not aware of any commercial production using it at higher sample rates. However, Benchmark stated even preemphasis support for 96 kHz sources for their DAC1 back in the days (Page 4), so there might be a use case after all)

 

[little-endian]

The sox audio utility can provide de-emphasis.

Yep, and pretty decently as far as I've read.

However, while there is no doubt that on a PC of course there are plenty of ways to "fix it in post", it has nothing to do with any convenience as it's cumbersome* to do and one has to be aware of the preemphasis at first, something not necessary when using a proper player and DAC.

* not only that, it may also increase the file sizes if one decides to "academically" save the deemphasis result in containers > 16 bit in order to keep the advantage of having applied preemphasis in the first place. However, maybe that's voodoo by today's DAC quality achieved already at 16 bit and thus only about having something correct sounding at 44.1 kHz / 16 bit.