On Thu, 16 Sep 2004 11:51:51 -0400, Jerry Avins <jya@ieee.org> wrote:>Randy Yates wrote: > >> Not to look a gift whore in the mouse, but how do you know all this? > >Hey! it's Allan Herriman you're asking! That's as close to a dumb >question as I've seen here in a long time. He's our resident >encyclopedia for audio and broadcast specifications. ;^)I got it wrong though. The emphasis curve has a single continuous time zero with a time constant of 50us and a single continuous time pole with a time constant of 15us. The pole was added to limit the amount of high frequency boost (to about 10dB). Regards, Allan
A Basic Question or two on CD Emphasis/Deemphasis
Started by ●September 16, 2004
Reply by ●September 16, 20042004-09-16
Reply by ●September 16, 20042004-09-16
Randy Yates wrote:> Jerry Avins <jya@ieee.org> writes: > >>[...] >> >>>Not to look a gift whore in the mouse, but how do you know all this? >> >> >>Hey! it's Allan Herriman you're asking! That's as close to a dumb >>question as I've seen here in a long time. He's our resident >>encyclopedia for audio and broadcast specifications. ;^) > > > I wasn't aware of this title for him. How STUPID of me! (Or > at least ignorant...)Seriously, whenever he writes with apparent authority, I become a believer. Don't underestimate him. He came to meet a customer not far from me a few years ago, and we arranged to meet. At some point during the arrangements, he sent me a satellite photo of my house and asked if it was the right one. (It was.) I believe anyone who knows how to do that! Jerry -- Engineering is the art of making what you want from things you can get. �����������������������������������������������������������������������
Reply by ●September 16, 20042004-09-16
On 16 Sep 2004 12:17:18 -0400, Randy Yates <randy.yates@sonyericsson.com> wrote:>Allan Herriman <allan.herriman.hates.spam@ctam.com.au.invalid> writes: > >> On 16 Sep 2004 11:21:02 -0400, Randy Yates >> <randy.yates@sonyericsson.com> wrote: >> >> >Allan Herriman <allan.herriman.hates.spam@ctam.com.au.invalid> writes: >> > >> >> On 16 Sep 2004 09:53:37 -0400, Randy Yates >> >> <randy.yates@sonyericsson.com> wrote: >> >> >> >> >What is it? >> >> >> >> High frequency boost, defined by a single continuous time zero. >> >> IIRC the time constant is 15us. >> > >> >OK, thanks Allan. Is the high frequency boost done on the signal >> >at CD record time, then the inverse (high frequency attenuate) done >> >at CD playback time? >> >> Yes. > >OK, cool. > >> >> >Why is it used? >> >> >> >> Is it used? I don't think I have any CDs in my collection that use >> >> preemph. >> > >> ><Argh!> It's IN the DAMN PLAYER SO IT MUST BE USED! At least there was >> >the intention for it to be used. Most designers I know of don't design >> >stuff in for no reason. "Hey, let's add a waffle iron to this GlidePath >> >navigation radio...". >> >> Most designers follow the standard, if there is one. The Red Book >> says to support optional deemphasis in the player, so they do. >> Whether it gets used is another matter. >> >> There's a flag in the format somewhere that indicates that whether >> preemp has been used. I don't know whether it's a per disk or per >> track thing. As I said, most disks are recorded without preemphasis. >> >> I think your question should have been "why did they put preemphasis >> in the Red Book?" > >And how is that question different than: > > "Why is it used?" > >????????????????????????????????????????Well, it's *not* used much in practice, so the question "why is it used?" isn't that meaningful. It's much more interesting to ask "what caused preemphasis to be put in the CD standard?" and "why isn't it used in practice?">> The answer to that can be found in any undergrad communications >> course. > >> The audio signal being recorded tends to have less power at high >> frequencies. > >Why?You can convince yourself of this by pluging a microphone into an audio spectrum analyser. The microphone will have some upper frequency cutoff, but the audio itself tends to have less power at high frequencies. (An easier way might be to look at the spectrum analyser display in something like WinAmp.) Now hit a cymbal. There will be lots of high frequency power, but it will be short-lived. If you really want to know *why* I suggest asking some acoustics folk.>> The digitising process tends to make more noise at high >> frequencies. > >Why would you say that? The standard assumption that quantization >noise is white says there is not more noise at high frequencies.I was referring to the ADC/DAC combination (although I see now I used the wrong name). The early audio DACs generated noise (see below). You are right, of course - white noise is a pretty good model for (ideal) ADC quantisation.>> So boost the highs during recording and cut them on >> playback to improve the overall SNR. > >Oh, yeah - I should have known THAT. How *STUPID* of me!!!!! > >EXCEPT I didn't "know" that a) the audio signal has less power >at high frequencies and b) quantization noise has more power >at high frequencies. > >:) > >And what happened to your reconstruction filter conjecture???Snipped. I believe it's still true, but it wasn't relevant to the point I was trying to make.>> Note that (US) FM radio (1940s?) uses preemphasis with a 75us time >> constant. Many other parts of the world (which got FM radio much >> later than the US) use 50us. CD (1970s) is switchable to 15us or >> none. >> I believe this trend (of smaller time constants = less boost) over >> time is related to the improvement in recording techniques, >> microphones, etc. that causes audio to have more high frequency >> content. > >Hmmm. > >> (BTW, 16 bit DACs weren't available until the mid '80s, and the early >> players used 14 bit DACs with a simple PWM arrangement (no noise >> shaping, etc.) so any extra LP filtering after the DAC was welcome.) > >Why would DAC resolution (at least for a "standard" DAC like these, >what, R2R ladder types) have any bearing on the filtering that >was required?The audio was 16 bit and the DAC was 14 bit. The extra bits were made up with PWM. This generates high frequency noise. (NB. This seemed like a good idea at the time, but that was before people believed in noise shaping or could make true 16 bit DACs.) Regards, Allan
Reply by ●September 16, 20042004-09-16
Allan Herriman wrote: (snip)> Note that (US) FM radio (1940s?) uses preemphasis with a 75us time > constant. Many other parts of the world (which got FM radio much > later than the US) use 50us. CD (1970s) is switchable to 15us or > none.Not to mention that Dolby encoding with 25us preephasis is also used. Dolby, on average, increases the high frequencies, which is somewhat corrected by going through 75us deemphasis for those without Dolby decoders.> I believe this trend (of smaller time constants = less boost) over > time is related to the improvement in recording techniques, > microphones, etc. that causes audio to have more high frequency > content.Maybe also trends in music. Rock, at least, tends to have a lower dynamic range than classical, so usually the average signal level is higher. People tend to register the average, where recording systems are limited by the peak signal level. Some music styles likely have more in the high frequency range. -- glen
Reply by ●September 16, 20042004-09-16
"Randy Yates" <randy.yates@sonyericsson.com> wrote in message news:xxp656ew5m9.fsf@usrts005.corpusers.net...> Allan Herriman <allan.herriman.hates.spam@ctam.com.au.invalid> writes: > > > The audio signal being recorded tends to have less power at high > > frequencies. > > Why?I think this is a pretty well-established principle for "normal audio". For example, from http://www.rane.com/pdf/note128.pdf: "Studies show the typical spectral energy for different types of music have high frequency energy considerably lower in level than low frequency energy [2]." [2] R. A. Greiner and Jeff Eggars, "The Spectral Amplitude Distribution of Selected Compact Discs," Journal of the Audio Engineering Society, vol. 37, pp.346-275 (April, 1989) ----- Obviously, if you're recording synethetic signals, this may not be the case, but for the broad categories of "music" and "speech" it holds as a generalization.
Reply by ●September 16, 20042004-09-16
"Randy Yates" <randy.yates@sonyericsson.com> wrote in message news:xxpllfaxqu6.fsf@usrts005.corpusers.net...> What is it? > Why is it used? > How is it usually implemented (digitally or analogically)?Here are some quotes from the web courtesy of Google: "Emphasis came about because of early converter design. The entire sampling process was new, and A to D converters exhibited low level noise because of bad linearity in the conversion process. This process added some high frequency broadband noise to the digital signal. Manufacturers overcame this byproduct by boosting (emphasis) the high frequencies during the conversion from analog to digital, and then rolling off (de-emphasis) the high frequencies by the same amount after the conversion back fro digital to analog. This process was optional and there was a switch to select emphasis on each track during record. A flag was set in the digital bit-stream, which automatically activated de-emphasis during playback. All CD players, DVD players, and DAT machines detect this flag and turn on a high frequency roll-off in the analog domain during playback. If the digital signal contains emphasis and the flag is missing or turned off, then the roll-off does not occur and the audio will be brighter than normal." http://www.rogernichols.com/EQ/EQ-2003-07-08.htm "Preemphasis has been used since the earliest days of commercial recording. In general, the high-frequency content of the music (or whatever) being recorded is low and the noise is high. Therefore, treble was boosted and lows were cut by a preemphasis curve which was removed in playback. The standard RIAA curve for turnover and rolloff (the amount and frequency for treble and bass, respectively) was not accepted universally until the 50's, and some fine preamps offered selectable values with presets for the common curves into the early transistor era." http://www.cdrfaq.org/faq03.html#S3-12
Reply by ●September 16, 20042004-09-16
On 2004-09-16 19:55:11 +0200, Jerry Avins <jya@ieee.org> said:> ...At some point during > the arrangements, he sent me a satellite photo of my house and asked if > it was the right one. (It was.) I believe anyone who knows how to do > that! > > JerryHi Jerry, I don't know if he did it that way but there are some programs that can do that (http://www.chimoosoft.com/terrabrowser.html). Now, if I were to send you such a picture, I really hope you don't believe me everything...! :-) -- Stephan M. Bernsee http://www.dspdimension.com
Reply by ●September 16, 20042004-09-16
"Randy Yates" <randy.yates@sonyericsson.com> wrote in message news:xxpd60mxlvn.fsf@usrts005.corpusers.net...> Randy Yates <randy.yates@sonyericsson.com> writes: > > [...] > > Allan Herriman <allan.herriman.hates.spam@ctam.com.au.invalid> writes: > > > > > I guess it seemed like a good idea at the time. (Remember that the > > > red book CD format was thought up in the 1970s, with 1x oversampling > > > and analog reconstruction filters in mind. > > > > > The extra lowpass pole in > > > the player helps.) > > > > That could be the reason. Is it? > > 1. If this process is done in the analog domain at both ends, then > what you gain in the reconstruction filter is lost in the anti-aliasing > filter.True, but the ADC end was done on expensive mastering equipment, whereas the DAC was in cheap consumer CD players. So this trade-off could be a winner.> 2. Why couldn't this have been done independent of the medium? That is, > if the *player* DIGITALLY emphasized the signal, then the analog > reconstruction filter, which would include the deemphasis filter, > would have a better time of things, and the CD proper need not be > aware of any of this.I don't know for sure, but perhaps at the time digital processing like this was very expensive (especially compared with an op-amp, R's and C's). Also, as mentioned elsewhere, it is difficult to deal with the high frequencies digitally due to frequency warping effects.> 3. Why isn't this done even today? Even a delta sigma D/A could use > some help in the analog filtering stage.I'm guessing that the amount of "help" you get is small compared to the hassle of encoding it and the possibilty that some obscure player out there won't handle it correctly.> I have been hearing about CD emphasis/deemphasis since they came out > around 1983 and have never heard a satisfying, authoritative answer > as to why it is used. Frustrating!Probably nobody really talks about it because it actually isn't used! To me, it seems like an artifact of applying well-understood old principles to a new and very immature digital system. Also, things get added to standards for various political reasons not directly related to technology. I wasn't in those committee meetings, so this is just conjecture.
Reply by ●September 16, 20042004-09-16
"glen herrmannsfeldt" <gah@ugcs.caltech.edu> wrote in message news:ciclbp$uv$1@gnus01.u.washington.edu...> > > > Note that (US) FM radio (1940s?) uses preemphasis with a 75us time > > constant. Many other parts of the world (which got FM radio much > > later than the US) use 50us. CD (1970s) is switchable to 15us or > > none. > > Not to mention that Dolby encoding with 25us preephasis is also > used. Dolby, on average, increases the high frequencies, which > is somewhat corrected by going through 75us deemphasis for those > without Dolby decoders.What's the deal with specifying filters in terms of time? Wouldn't a frequency be more intuitive? A naive question, what is the cut-off frequency (3dB-point I presume) of a 25us filter? Is it 1/25us = 40kHz or 1/(2*pi*25us) = ~6.37kHz or something else? I'm guessing it's the 6.37k.
Reply by ●September 16, 20042004-09-16
Jon Harris wrote: (snip regarding preemphais in CD recording)> Probably nobody really talks about it because it actually isn't used! To me, it > seems like an artifact of applying well-understood old principles to a new and > very immature digital system. Also, things get added to standards for various > political reasons not directly related to technology. I wasn't in those > committee meetings, so this is just conjecture.It might be that it is not used for political reasons, also. It is widely believe that less analog processing equals better sound. (Except by analog purists with vacuum tube amps.) By the way, CDs can have a maximum of 99 tracks because the track number is BCD coded in 8 bits. BCD coding makes it easier for the player to display the track number. Compare the amount of logic to do 8 bit binary to BCD decoder to a digital filter and you have some idea where things were at the time. -- glen
