What is the steepest realistic audio transient in terms of V/us?

[Waly]

the Square wave is filtered with a 30KHz low pass.

Mr. Curl is quite confident that his target audience will remain in awe and therefore not read beyond the title and perhaps the abstract.

I think he is correct in his expectations.

 

[Chris Daly]

I've looked, don't see anything for CD enthusiasts to be embarrassed over


there are a literal count on your fingers number of exotic custom analog mag tape machines made for > $100M budget movie sound production that can exceed CD dynamic range

but then the fair comparison would be today's digital audio mastering process which is generally 24/96 with the best ADC again exceeding those custom analog machines with greater than 120 dB dynamic range

Given the majority of CD players have JRC4558 ( and a lengthy thread on DIY
describes them as the worst op amp ever made ) there is a limitation.

SPDIF is a limitation, installed in about 60% of all CD players

Poor transports and lack of attending to jitter reduction, is another limitation

Poor I/V stages is another

Mastering processors yes address many limitations of CD players
namely dither shaping ,compression and gating.

There are many ways to achieve the benefit of greater dynamic range
but will invariably involve companding, so a working knowledge of
the different companders and dither types is called upon.

Avoiding dither altogether also has appeal and the TDA1541
managed to get 16 bits pretty right other than current output
being unequal on each channel.

The best audio I think presently I know of is achieved using a largely jitter free CD player - with many74HC or AC14 chips tidying up data and clock lines, into a Type 1 DBX
Tape In then tape Out into a Type 4 as seen in the DBX Quantum mastering processor, set to 24 bits bypass other than Type 4 processing. Its dither to HPTPDF
either analog out of the DBX Quantum back into the Type 1 and Type 1 then
to the best known attenuator. Type 1 DBX 150x having very low impedance
capability.... until something else comes along.

Cheers / Chris

 

[Waly]

Presumably a fashion audio designer who doesn't care about actually reconstructing the signal?

And specializing in dirty sand resistors, right here on DIYaudio. Anno Domini was 2009.

 

[woody]

In the old Pass A40 article Nelson makes a little comment on this near the end of the article. Something to the effect 1.5v/us was the highest he had recorded and that was with a special Japanese recording using a straingage cartrige.

 

[SY]

Avoiding dither altogether also has appeal

Only if one is thoroughly ignorant of digital technology.

 

[billshurv]

Given the majority of CD players have JRC4558 ( and a lengthy thread on DIY
describes them as the worst op amp ever made ) there is a limitation.

SPDIF is a limitation, installed in about 60% of all CD players

Poor transports and lack of attending to jitter reduction, is another limitation

Poor I/V stages is another

I've never owned a digital player with JRC4558 in so dunno what this majority is.

how does SPDIF affect DR? DR is in the bits?

Look at any stereophile measurements and show me what these dynamic range limitations are. They measure low level linearity...

 

[jcx]

comparing mag tape mechanical motion noise scrape flutter, wow to >4 orders of magnitude lower Quartz Xtal Osc jitter? - you apparently don't know the numbers

read Jamie's white papers sometime http://www.plangentprocesses.com/index_save2.htm

better yet attend one of his talks/demos

 

[billshurv]

I take my hat off to Jamie. Some of the stuff he is doing is very impressive.

 

[john curl]

If people can understand what we did back almost 40 years ago on this subject, then let's talk! '-)

 

[scott wurcer]

better yet attend one of his talks/demos

OK with Bob Ludwig OK by me. I do have to go up and visit that guy.

 

[billshurv]

Well you stated that a good pre-amp should have a filter to limit slew rate being passed to the power amplifier. I hope all your pre-amps have that!

 

[GoatGuy]

Kind of directed at Mr. Curl…

Here's a big virtual "Go Bears!" hug from a fellow Cal man. The upper division courses in analog design were just wonderful stuff. Hard, too. Anyway… now that that's behind me…

Since you replied (thank you!), I've been perusing some of my old textbooks on the matter, as well as just stewing. I think it comes down to this: an amplifier with some maximum slew-rate will also have a bandwidth that numerically supports that impulse response.

Which kind of answers (proving myself wrong) one of my early conjectures: that amplifiers that are slew-rate limited will “do better” at lower amplitude output. Nope. Step response is a function both of the size of the step, and the bandwidth of the amplifier itself.

Which then answers to some of the stuff you put out there: from experiments that you did some 30+ years back (and one presumes, that you cobble together from time to time to demonstrate the effect…) you found simply that amplifiers with 100 kHz response sound better than ones with demonstrably capable but 20 kHz response. Which I now get.

Thing that the higher bandwidth amplifier does is keep more of the energy within the impulses that are being reproduced. It doesn't need or even benefit from being absolute and perfect, but rather “good enough such that no improvements are perceptible amongst Golden Ear types in double blind tests”. Empirical testing. Which you found to be about 100 kHz at the top end.

So, all then one has to do is use DF96's math (or mine, or the textbook's), and relate 'em all together with a couple of predictive formulæ:

V_risetime = V_peak × 2 π • f_{hi limit};
V_risetime = 1.0 × 2 π × 100,000
V_risetime = 0.63 V/μs per volt, peak, at output

Great number to remember. Then working with Ohms, Watts and the rest. Lets use 100 watts 'design spec', and RMS watts, and 8 Ω nominal impedance speakers:

V_peak = √(2 × P_output • Z_load)
V_peak = √(2 × 100 W × 8 Ω)
V_peak = 40 V peak

So the 100 kHz bandwidth factor now can be worked in:

Slew design = 40 V × 0.63 V/μs per volt
Slew design = 25 V/μs

And there you are. It all links together.
Thanks for the inspiration to flop open the old books.

GoatGuy

 

[DPH]

Goatguy--just remember that small signals will not slew as hard as large signals by definition. Slew rate is a large signal property. 😀

At the risk of repeating (oh, heck, flat out stealing) Waly: a 5x oversize of power bandwidth over (ostensible) audio is pretty harmless and shouldn't cause problems with answering your cellphone for you. 😀

 

[Chris Daly]

I've never owned a digital player with JRC4558 in so dunno what this majority is.

how does SPDIF affect DR? DR is in the bits?

Look at any stereophile measurements and show me what these dynamic range limitations are. They measure low level linearity...

SPDIF in most devices is limited to 20 bit
https://en.wikipedia.org/wiki/S/PDIF

and has problems related to the use of biphase mark code causing bandwidth limitations http://audioworkshop.org/downloads/AES_EBU_SPDIF_DIGITAL_INTERFACEaes93.pdf

 

[jcx]

I don't understand this limping rearguard action against digital audio in 2015

how many are really actually spinning CDs in a dedicated player, using SPDIF to a outboard DAC?

and where are the resounding, peer reviewed, crushing demonstrations of 16/44 inadequacy in commercial music playback in only 3 decades since its launch?

 

[benb]

Trying to get a post in before the lock.

[famous designer from the off]

Yeah, but dithering destroys the sound.

I'd like to see the Fletcher-Munson curves for THOSE ears.

 

[Pano]

I don't know how we can talk about good Slew rate without mentioning Hialeah back in '77.
That was a Slew rate of over 5 furlongs/minute. Darn fast!


(race fans will understand)

 

[john curl]

It is almost impossible to put enough rise time control into either a preamp or a power amp with rolling off the hi's noticeably. Best to increase slew rate, it's easier.

 

[DF96]

Now what makes high transient signals? Not a summation of a few sine waves, but percussion or just hitting a couple of wood blocks together might do it.

You have microphones which can go to 100's of kHz?

ONLY an added filter to the square wave can give you a defined or controlled rise time.

But there is ALWAYS a filter! The microphone is the first filter, then there is the mike preamp, then there is either the cutter amp (plus vinyl elasticity) or the tape head gap or the anti-aliasing filter.

For example, a 100Hz square wave could have a 1ns rise time (10us is more probable) yet on a spectrum analysis be 40dB down at 10K from 100Hz and still falling at higher frequencies. This can be misleading, because it might be almost impossible to pass the 100Hz square wave at modestly high output levels (100Hz audible) without slew rate limiting of the amplifier that is not designed for 10us rise time and faster audio signals.

This would only confuse those who have never calculated the Fourier analysis of a square wave. The rest of us would know that 10kHz has to be 40dB down, and 100kHz has to be at least 60dB down even with infinite slew rate and without any filtering.

Yeah, but dithering destroys the sound.

That's funny. I thought it was supposed to improve the sound, and indeed does so.

 

[billshurv]

But that is NOT what you said all those years ago?

As the topic is about signal slew not amplifier slew we can safely say the OP was answered on page one. You assert that 100V/us is needed for a power amp and 10 for a preamp. Reading the article I worry about correlation vs causality, but I know that has been argued to death on other threads and 100V/us on a power amp is not excessively hard for a DIYer to do.

Hardly worth arguing about now I'm sober...