BW getting up there with the best LP
And the sound quality is still judged to be 90-95%
THx-RNMarsh
Last edited:
Did not one of the external DACs of lore do an interpolating re-sample at double the original 44 so they could then use a bit less steep filtering?
Some preamp designs used to have a "cartridge loading" (capacitor) selector. Now there's a user selectable filter. I recall discovering years ago one of the differences between how the vinyl version sounded versus the CD, was that the engineer clearly mixed the CD to sound like the vinyl using the wrong cartridge loading. Oops.
Some preamp designs used to have a "cartridge loading" (capacitor) selector. Now there's a user selectable filter. I recall discovering years ago one of the differences between how the vinyl version sounded versus the CD, was that the engineer clearly mixed the CD to sound like the vinyl using the wrong cartridge loading. Oops.
Thermal tracking between VBE multiplier (or whatever is bias generator)and
OP devices. It seems hard to fathom as MOSFETS have such low gm
compared to BJT's.
I build power amps with no to 0.1ohm OP stage emitter resistors so far
without any issues. I probably wouldn't sell such a thing and if I did it would
have either a solid copper block or a heat pipe underneath / connecting all
the relevant devices.
One thing I was going to suggest a long time ago was to be careful with the
bootstrapped IP stage. They can oscillate. It's a good idea to decouple the
bootstrapped collectors.
TCD
I was thinking about something like this:
Boashash, Boualem; Time frequency signal analysis: Past, present and future trends; Control and Dynamic Systems, 78 (C), 1996, 1-69
"Something to sell?" Good question, most people are "selling something" ; for example selfconverted "ex-golden-ears" like to sell the idea that they only could have been fooled because of a worldwide conspiracy.
Other people feel their ego is at stake, which might be not direct "selling something" but .....
Spending too much time with people only pretending to be interested in "audioscience" comes at a risk and the result IMO is unfortunately posting things like "What ever you say" in this context.
Good question, but doesn't it depend on the format which will be used for "selling/distributing"?
And we should not forget about all the records that have been released in the days of the not so modern SD ADCs.
As far as I remember, I've posted some examples of antialiasing filters used in the 1980s and we've seen the data from oversampling digital filters used in CD-players (halfband at 22.05 kHz) as well.
But I guess "Clueless in Gaza" covers it all.......
I find it an interesting technical discussion. As for the rest of your post, I don't agree with any of it.
I don't think so. I rechecked the specs on the TI (nee BB) PCM4222 and the modulator runs at more or less the same frequency whatever the output bitrate or format is. The decimation and digital LPF determine the output bit rate. The stock digital filter is 100dB down at 0.55fs. Same analog front end from 44.1 to 192k output.
I agree with the technical discussion, too. Some of it is interesting. Especially JN's work/point.
To each his own for the rest of it. "Long live the CD".
Onward and forward to better things.
Enjoy,
-Richard
Last edited:
I think this is all immaterial. I have heard a few 44.1/48 kHz recordings, including some made from tape, that blow away some 96 kHz and 192 kHz material. I've compared an album recorded at 192 kHz with downsampled versions of the same files as well. I tried to do a DSD comparison but almost every hybrid SACD I found had different content on the Redbook layer (i.e. it was obviously mixed differently or compressed even more).
Anyway, my personal opinion is that there is not a significant difference between formats. I'm not even sure the modern converters really make the difference. I think it's recording engineers finally learning to use the tools and software, along with the software drastically improving.
I'm sure my observations will be discounted, though. I did this uncontrolled ad-hoc test about 9 years ago with Sennheiser HD800 and both Benchmark DAC1 and EMU 1820m and 27 year old undamaged hearing.
Anyway, my personal opinion is that there is not a significant difference between formats. I'm not even sure the modern converters really make the difference. I think it's recording engineers finally learning to use the tools and software, along with the software drastically improving.
I'm sure my observations will be discounted, though. I did this uncontrolled ad-hoc test about 9 years ago with Sennheiser HD800 and both Benchmark DAC1 and EMU 1820m and 27 year old undamaged hearing.
Of course. It's obviously not settled, simply look at the technical discussion on both sides and the increasingly "desperate" arguments
That's correct.
The advised anti alias filter for the PCM4222 has its -3dB point at 590Khz.
Passband ripple is excellent.
Hans
That illustrates what I've meant. I'd think that the recording people usually don't use these filters but after editing/mixing/summing use some of their own DAW filters, but some of that weren't such a good choice expecially one or two decades before.
If using 192 kHz there should be usually no problem with aliasing or imaging filters but in case of 44.1 kHz the datasheet shows one of the typical half-band filters (means 6 dB down at Nyquist) and if we take into consideration for example the NPC digital filters used for oversampling in CD-players (also typical half-band filters 6dB down at Nyquist) it becomes obvious why sometimes it might be a less than optimal solution.
Edit: On the recording side, the usual suspect like Bob Katz seem to think that it is mostly a matter of filter quality, while the bandwidth extension is not of importance. Dan Lavry still maintains that in his opinion 44.1 kHz is not sufficient and that it should be 60 kHz (or I'd guess the next Fs above, broadly used like 96 kHz for practical reasons).
Others still think differently.....
Last edited:
Not sure if that is related to my post, as usually the editing/mixing/summing takes place in the digital domain at (for example) 192 kHz and at the end the downsampling/dithering process will happen, hopefully done nowadays with a filter not being a halfband at 22.05 kHz.
With all this extended bandwidth talk, what about the microphones?
Recording a cymbal with a mic that doesn't extend up far doesn't produce a recording useful for format comparisons, "nothing up there to see".
JC showed a mic running up to 40khz, do studios go there? Would it be useful to use high BW mics only on specific instruments?
What is the BW of a lead guitar pickup, and is the instrument capable of transients over Fs? I suspect a hard pick generates some whopping leading edge transients at string let-go.
Jn
Recording a cymbal with a mic that doesn't extend up far doesn't produce a recording useful for format comparisons, "nothing up there to see".
JC showed a mic running up to 40khz, do studios go there? Would it be useful to use high BW mics only on specific instruments?
What is the BW of a lead guitar pickup, and is the instrument capable of transients over Fs? I suspect a hard pick generates some whopping leading edge transients at string let-go.
Jn
I find the biggest determinant of recording quality is the recording itself - not the medium.
I listened to Yosi Harikawa's 'Bubbles' on MP3 via Spotify over the weekend at my son's house (he's a KEF LS50 fan as well ) and was blown away by the 3D sound stage both way beyond the left and right of the speakers and very deep. Another one is Yo-Yo Ma's 'Peace' CD recording - very holographic recording.
So, while 24 bit 192 may offer some small things you can hear if you really listen hard, its really only adding a few tweaks on top of what the good recording is delivering in terms of a 'sonic' experience. The same thing applies to good vinyl as well - my favourite example being 'Ella Fitzgerald Sings the Irving Berlin Songbook'
I listened to Yosi Harikawa's 'Bubbles' on MP3 via Spotify over the weekend at my son's house (he's a KEF LS50 fan as well
) and was blown away by the 3D sound stage both way beyond the left and right of the speakers and very deep. Another one is Yo-Yo Ma's 'Peace' CD recording - very holographic recording.So, while 24 bit 192 may offer some small things you can hear if you really listen hard, its really only adding a few tweaks on top of what the good recording is delivering in terms of a 'sonic' experience. The same thing applies to good vinyl as well - my favourite example being 'Ella Fitzgerald Sings the Irving Berlin Songbook'
First recording made with 230 Volt on the ESL63.
Recorded on axis from 1 meter distance with pink noise, smoothed in 1/3 band octaves.
Microphone with MDF correction table flat to 40Khz.
I'll have to wait for 24hrs before the 5kV bias has been dropped and the next recording with 10% less mains supply can be made.
Hans
Recorded on axis from 1 meter distance with pink noise, smoothed in 1/3 band octaves.
Microphone with MDF correction table flat to 40Khz.
I'll have to wait for 24hrs before the 5kV bias has been dropped and the next recording with 10% less mains supply can be made.
Hans
Excellent, thank you.
As the discussion included polar plot variations with DC bias, would it be possible to do a few off axis?
Thanks, jn
Edit: Hans, didn't you say your neon blinked every minute or so? That would mean your losing roughly a volt per second.
If you have a variac, you could dial the line down and just wait for the neon to wake up.
Last edited:
Yes but that has already been filtered. You can access the 6bit modulator output and roll EVERYTHING yourself if you want.