TMI about square waves

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
My preference is to bandpass signals appropriately in front end circuits to avoid the risk of depending on other designers to do the heavy lifting. Making their work easier can improve the entire audio chain's sound quality.
Yes, it used to be a 'best practices" design rule (of thumb or otherwise) that "garbage in = more garbage out"
 
My Mom sent me this, it was a part of the literature and sales brochures for my dad's Hi-Fi rig, which was comprised of a Harmon/Kardon Citation 12 Deluxe Power Amp, H/K Citation 11 preamp, Dual turntable, University dual 12" coaxial speakers, etc.. There is a printing code on the back cover "90731099", I would guess it's from around 1973?

It's 8 pages in total, here is the front and back cover:

IMG_5321.jpgIMG_5322.jpg
 
Last edited:
This is a great thread. I notice a few things about 2x speed recording. My experience was trying to use a cheap 96K desktop interface in the hopes of recording things above 22K on my bench but it made no difference what speed it recorded at it seemed the inputs and outputs were rolled off heavily about 22k making higher freq recording impossible. Does anybody's ADC and DACs actually record to 48K? What op amps are in the ins and outs?

Another thing is recording at a 24 bit depth is more accurate than 16. So does anyone have anything to say about how that affects the HF?

I remember the slew rate issues but since the designs I worked on were open loop vac tube gain stages, passive RIAA and relatively low NFB power amp circuits, around 10 db I didn't run into slew issues. Mostly tubes and tube fet hybrids.
 
This is a great thread. I notice a few things about 2x speed recording. My experience was trying to use a cheap 96K desktop interface in the hopes of recording things above 22K on my bench but it made no difference what speed it recorded at it seemed the inputs and outputs were rolled off heavily about 22k making higher freq recording impossible. Does anybody's ADC and DACs actually record to 48K? What op amps are in the ins and outs?
most modern a/d front ends actually sample at very high rates and then decimate the result down to a roughly 20kHz bandwidth. You could probably trick a modern convertor chip to capture a wider (higher) bandwidth... That does not mean you will hear it.
Another thing is recording at a 24 bit depth is more accurate than 16. So does anyone have anything to say about how that affects the HF?
24 bit word length has finer resolution but accuracy is a different spec.

JR
I remember the slew rate issues but since the designs I worked on were open loop vac tube gain stages, passive RIAA and relatively low NFB power amp circuits, around 10 db I didn't run into slew issues. Mostly tubes and tube fet hybrids.
 
This is a great thread. I notice a few things about 2x speed recording. My experience was trying to use a cheap 96K desktop interface in the hopes of recording things above 22K on my bench but it made no difference what speed it recorded at it seemed the inputs and outputs were rolled off heavily about 22k making higher freq recording impossible.
This is quite common. Many cheap (and not-so-cheap) interfaces have a single decimating filter, on account they are designed for audible signals.

Does anybody's ADC and DACs actually record to 48K?
Several. I have or had interfaces that record up to about 82 kHz (at -3dB) with 192kHz sample rate. They allowed recording 42kHz at 96k SR. One I commonly use is the Tascam US 2X2 HR.
OTOH I had an interface that provided good response up to 85kHz @ 192k SR, but the DAC side used spectrum-shifting for noise reduction in the audio band, which resulted in very low S/N ratio at ultra sonic frequencies.
What op amps are in the ins and outs?
Irrelevant. It's the filters that define the HF response;
Another thing is recording at a 24 bit depth is more accurate than 16. So does anyone have anything to say about how that affects the HF?
It depends very much on the spectrum of the recorded signal. If the HF content is very weak, some of it may be "lost" at 16bit, whilst 24bit would make it more resolved. As JR mentioned, you're not very likely to hear it in a single-pass recording.
24-bit captation (and subsequent 32-bit) processing allows multiple passees of processing to be cleaner.
 
Last edited:
Thank you, gents.

OK. Now I'm more up to date. I haven't worked on one since 90s. I'm a tad older so I can't hear much above 10K however on a good playback system I can hear the difference in the final results, especially in the top end, much cleaner and sibilants don't sound so harsh in 96/24.

IMHO the op amps are not irrelevant. They all have sonic signatures as do the filter they are in. I bypassed some in a hi-end DAC and lo and behold, much nicer and cleaner. But that was in the 90's. I think it was a Krell.

I was trying to use the audio interface to use a mac as a scope for noise but the one I had had RF all over it -inside and out. There was no way to stop it. I finally gave in and got a Siglant 4 ch 1104X-E 200 MHZ scope. I was using a Hitach 2 ch 35 mhz scope. I still use Leader analogue equipment - ACVM LMV 185,LDM 171 Harmonic Dist analyzer, LAG 120 10hz to 1 MHZ audio generator and a second LMV 189 ACVM. All quiet and good enough to find trouble. A and nd my highly trained ears. Cymbals and percussion instruments and strings tell me a lot.

I've been standing next to drummers for the last 56 years. I know what these all sound like.
 
IMHO the op amps are not irrelevant. They all have sonic signatures as do the filter they are in. I bypassed some in a hi-end DAC and lo and behold, much nicer and cleaner.
Last century HiFi/audiophile designers were known to be a tad awkward with opamps, often blaming opamps instead of their own shortcomings. If bypassing an opamp results in better sound, the opamp is often not at cause, it's the designer, who chose the wrong opamp or made a mess of its implementation.
I was trying to use the audio interface to use a mac as a scope for noise but the one I had had RF all over it -inside and out. There was no way to stop it. I finally gave in and got a Siglant 4 ch 1104X-E 200 MHZ scope.
Remember that with a scope, you're looking at noise within the scope intrinsic bandwidth. If the scope has a BW of let's say 20 MHz, white noise is 30dB higher than the "audio" noise (limited to 20kHz).
Now, visualizing out-of-band noise is instructive, since it allows detecting oscillations and RF interference.
 
Last edited:
I looked up the Tascam. $129 now from B&H in NYC. Have you ever scoped one to see how much RF is sitting on it? Case and all, that is.
Yes I have, particularly after the disappointing behaviour of the Native Instruments Komplete Audio 2 DAC.
i have upgraded my system to a combo of a Cosmos ADC and Topping E30, because i wanted the possibility of operating at 384k SR, which gives me correct measurements up to 160kHz, but for less frequency-hungry measurements I still use the AKAI.
The Cosmos/Topping combo requires using an ASIO wrapper, that combines the separate drivers into a virtual one, and the DAC and ADC being separate sometimes create noise issus due to USB power, so the stand-alone AKAI is less prone to issues.
 
Last century HiFi/audiophile designers were known to be a tad awkward with opamps, often blaming opamps instead of their own shortcomings. If bypassing an opamp results in better sound, the opamp is often not at cause, it's the designer, who chose the wrong opamp or made a mess of its implementation.
Probably both but it was the 90's and nothing but Red Book cd's which were pretty choppy to begin with. It just sounded much better right out of the dac. I figured the following preamp and power amps would bandwidth limit the signal. Customer was happy, I got paid..

That said I don't believe all op-amps sound the same or work the same. 120 db of open loop gain squashed down to 20 may look good on the bench but it can kill inner detail of a performance. Same thing is true of a textbook 12AX7 2 stage preamp with NFB - 70 squashed to 20. I changed AX to AU and tailored the component values. Now just 34 db squashed down to 20. Much more open!! Same thing true in text book nfb riaa preamps. Passive RIAA is better. I only did this for the last 40 years so I'm just getting the hang of it.

Remember that with a scope, you're looking at noise within the scope intrinsic bandwidth. If the scope has a BW of let's say 20 MHz, white noise is 30dB higher than the "audio" noise (limited to 20kHz).
Now, visualizing out-of-band noise is instructive, since it allows detecting oscillations and RF interference.
OF course. That's how you see all the spiky rf all over the place. My leader ACVM limits the BW to 1mhz when I use it. Of course I've used it since 1986 so I know what I'm seeing. My scope is good to 200 mhz now. I can see screen grids and mosfets go on the air on top of the audio.

We're on the same page.
 
Yes I have, particularly after the disappointing behaviour of the Native Instruments Komplete Audio 2 DAC.
i have upgraded my system to a combo of a Cosmos ADC and Topping E30, because i wanted the possibility of operating at 384k SR, which gives me correct measurements up to 160kHz, but for less frequency-hungry measurements I still use the AKAI.
The Cosmos/Topping combo requires using an ASIO wrapper, that combines the separate drivers into a virtual one, and the DAC and ADC being separate sometimes create noise issus due to USB power, so the stand-alone AKAI is less prone to issues.
Could you measure noise with it into a laptop in the microvolt region or is there still too much stray noise? I use Waveforms by Digilent as the software to make my Mac a scope. good enough for recording events over a long time.
 
Probably both but it was the 90's and nothing but Red Book cd's which were pretty choppy to begin with. It just sounded much better right out of the dac. I figured the following preamp and power amps would bandwidth limit the signal. Customer was happy, I got paid..
by the 90s CDs and off the shelf op amps were pretty decent (better than vinyl).
That said I don't believe all op-amps sound the same or work the same. 120 db of open loop gain squashed down to 20 may look good on the bench but it can kill inner detail of a performance.
? The behavior of NF is pretty well understood. Modern test bench instrumentation can parse out non-linearity deviations with much finer resolution than I can hear.
Same thing is true of a textbook 12AX7 2 stage preamp with NFB - 70 squashed to 20. I changed AX to AU and tailored the component values. Now just 34 db squashed down to 20. Much more open!!
not a tube guy so no comment
Same thing true in text book nfb riaa preamps. Passive RIAA is better. I only did this for the last 40 years so I'm just getting the hang of it.
don't get me started on riaa playback... I killed way too many brain cells optimizing riaa playback stage designs over a few decades.

If you use a passive RIAA EQ you still need tens of dB post make up gain so this becomes 6 of one half dozen the other, just like passive (cough) mixers :cool:. FWIW my last riaa preamp used a passive 75uSec pole, the relatively easy 318/3180 uSec EQ poles and zeros were active.

JR
OF course. That's how you see all the spiky rf all over the place. My leader ACVM limits the BW to 1mhz when I use it. Of course I've used it since 1986 so I know what I'm seeing. My scope is good to 200 mhz now. I can see screen grids and mosfets go on the air on top of the audio.

We're on the same page.
 
RE: Early CD v. Vinyl - That argument is legion among audiophiles, though a very small portion of the market are influential. Each media has their strengths and weaknesses, conveniences vs inconveniences. Each have great recordings and dogs.

I think recording engineers should attend some the Hi End HIFI shows and just listen.

I can hear the differences. My theory is: the brain is always searching for information. Driving in fog is fatiguing because the brain is deprived of necessary info. Of course the stakes are higher. If you really want to hear all the detail in a recording, the reverb tails, the sound of the room, the contrapuntal lines that accompany the lead, if it's in there but hard to hear it's fatiguing. That's the realism factor. It can get lost fast, Bad switches, controls, caps, everything everyone gets into on this forum are detail killers.

And then there's MP3s and AAC. Good enough for the car but fatiguing for me. It's like looking at a compressed jpeg photo. You get the idea but you can't see the peach fuzz on the model's face. And compression creates artifacts that aren't there in the original. But the convenience of living in the digital world has created a tolerance for both video and audio compression, which actually make the market viable.

So much has changed in the digital age.

Maybe my bias is from being an active jazz musician all my life and growing up with records. I know how instruments sound on stage and in the studio.That said, I always use CDs for audio test listening because anything I designed had to sound good with them. I have a SACD RCA recording of Van Cliburn direct 3 track made on tape in 1958. One of the best recordings I've ever heard - 3 mics. And the piano player was pretty good, too. NYC gave him a ticker tape parade.

Anyway, to each his own. The joy of music is the bottom line. For me. I hope everyone enjoys their's, too.
 
JohnRoberts said:
"The behavior of NF is pretty well understood. Modern test bench instrumentation can parse out non-linearity deviations with much finer resolution than I can hear."

Only you would know that. My test bench isn't all that modern. But I can hear when it ain't right. Listener fatigue.

You may have the best recipe in the world for dog food but if the dog won't eat, there you jolly well are.

Most importantly:
It don't mean a thing if it ain't got that swing. Do wah do wah do wah do wah do wah do wah do wah! - Duke Ellington

I'll leave it at that. Thanks

BTW I just played a concert last night with Houston Person, tenor sax. Keeping my ears tuned. That's me on bass.
 

Attachments

  • 1F7A7573-10x8.jpeg
    1F7A7573-10x8.jpeg
    91.8 KB · Views: 1
Last edited:
JohnRoberts said:
"The behavior of NF is pretty well understood. Modern test bench instrumentation can parse out non-linearity deviations with much finer resolution than I can hear."

Only you would know that. My test bench isn't all that modern. But I can hear when it ain't right. Listener fatigue.
I don't think I am the lone ranger here about that. I figured out several decades ago that I could measure things that I couldn't hear, but could always
iu

measure stuff that I could. My test bench was cobbled together on a budget. Nowadays kids with a decent soundcard and some free aps can do better.
You may have the best recipe in the world for dog food but if the dog won't eat, there you jolly well are.
"Don't let the same dog bite you twice", Chuck Berry
Most importantly:
It don't mean a thing if it ain't got that swing. Do wah do wah do wah do wah do wah do wah do wah! - Duke Ellington

I'll leave it at that. Thanks

BTW I just played a concert last night with Houston Person, tenor sax. Keeping my ears tuned. That's me on bass.
Sounds like fun...

JR
 
Hi folks
Really Don t know it s the right post but a fast reading makes me think this... Always used old adat recorders such as fostex ad8 and tascam mx24, this last one is so great! With not so esoteric soundcard. What I found is that even working at 44.1 16bit the sound is always bigger realistic and 3D even compared with modern high quality converters....
I believe it s the oversampling that these old vintage gital gear converters have
Even with modern music production, people is always excited when I mix down passing trought these devices, the answer is always the same... Hey Richi what you did?
My answer is Nothing! Only pushed play!
So is the Oversampling offered buy that converter that makes so big difference?
Thanks for your attention
Best
 
Hi folks
Really Don t know it s the right post but a fast reading makes me think this... Always used old adat recorders such as fostex ad8 and tascam mx24, this last one is so great! With not so esoteric soundcard. What I found is that even working at 44.1 16bit the sound is always bigger realistic and 3D even compared with modern high quality converters....
I believe it s the oversampling that these old vintage gital gear converters have
Even with modern music production, people is always excited when I mix down passing trought these devices, the answer is always the same... Hey Richi what you did?
My answer is Nothing! Only pushed play!
So is the Oversampling offered buy that converter that makes so big difference?
Thanks for your attention
Best

Do you mean Tascam MX2424. I can't find TASCAM MX24.
Doubt it's ADAT since TASCAM had their DTRS machines. Not that it's particularly relevant to the converters.
 
Do you mean Tascam MX2424. I can't find TASCAM MX24.
Doubt it's ADAT since TASCAM had their DTRS machines. Not that it's particularly relevant to the converters.
yes it s the mx2424 hd recorder with adat bridge expansion....
sometimes it s hard to boot with that crappy IDE hd and also the display has problems.... should be a psu fault has i read in past....but once started it s the main clock for my setup and my favourite converter
should be cool to upgrate it to Sata....searched a lot for mods but nothing found
Fans are very noisy! even this should be a further upgrade:ROFLMAO:
 

Latest posts

Back
Top