Read the paper and then comment - you've just reacted.
The TH article is hardly obscure (in vinyl replay engineering terms) - I've seen it referenced on DIY audio enough. And in the day 'Audio' was hardly an obscure publication for the cognoscenti.
I don't get the 'no one has ever bothered to replicate' bit. It was a serious attempt to plot the absolute limits of phono replay levels with some now fairly widely accepted recommendations.
No one is knocking 'Digital RIAA', but the dynamic range across records is what it is and the Holman paper is a very credible reference IMV.
Oh, got the wrong end of the stick completely on that one ....... I am sorry to have flounced out of the room - I should have tried a bit harder - I apologise.
Still, I was a bit puzzled by what you said, and upon reflection, you are right. I too, am uneasy with genres, as in there is 80's pop ...... and there is 80's pop. Like wasps and bees, they have stripes and stings, but are two totally different things.
Onwards and upwards. ToS
Last edited:
The Holmann article is great, but I question the sources of his overload margins. Other than the shure V15III advertorial there have been no attempts to replicate, test methods are not shown and the records are not listed. Therefore I have to consider that to be a grain of salt data point. You can accept it but its not validated.
Same as people said babies should be fed every 4 hours because in the 1880s someone thought it was a good number.
Same as people said babies should be fed every 4 hours because in the 1880s someone thought it was a good number.
No probs... Glad we are back on the same track!
And as has already been said - your bit about musical illiteracy was spot on...
But I do not get your logic. Why would you use a compressor after RIAA preamp? Noise gate even less.
Maybe you mean an attenuator?
How can the Holman article be 'great' if in the next breath you consider the data points to be a 'grain of salt'. So you just diss everything that doesn't fit with your opinion? Why don't you just adopt the same attitude to everything put up on this site, after all it has not been validated by anyone else has it?
Last edited:
It is just because the mentioned data points were published by Shure , according to various articles, based on a survey of dozens (if not hundreds) of records known for the "hotness" (means high levels) at that time.
But up to now, we haven´t seen the real data, neither the real number of records analyzed nor which records it were. Shure used the data to explain why their best system needed an advanced redesign, and disclosed that it still (although better than its predecessor) was not able to track the "hottest" of the analyzed records, which might be an argument against the "it was all marketing" - assumption.
I must admit not having contacted Shure about this questions, which could be a good idea.
But up to now, we haven´t seen the real data, neither the real number of records analyzed nor which records it were. Shure used the data to explain why their best system needed an advanced redesign, and disclosed that it still (although better than its predecessor) was not able to track the "hottest" of the analyzed records, which might be an argument against the "it was all marketing" - assumption.
I must admit not having contacted Shure about this questions, which could be a good idea.
They are data points that were published and I have to assume (as with anything like this) that it was based on hard measurements. If you can't take the technical argument as it is then you have to put something else up that is equally as compelling. And then someone can say the same thing about that data set. The Shure data is referenced by Holman, so it was available when he wrote the article - he would have been stupid to list it IMV if it was not and would have opened himself up to exactly the kind of comments flying around now, 40 years later.
A typical phono pre has to deal with MM carts that output from 2.5mV to 6mV (some exemplars as high as 8mV, but we'll drop that for now). A record is nominally 5cm/sec at 1kHz, with peaks of at least 5x that (+14 dB). So, I don't see that Holman's recommendation of c. 30 dB overload capability to handle all eventualities is out of order and that's why shooting for > 20 dB O/L is not outrageous. If you are doing a passive phono stage with the ~2.1 kHz breakpoint in the second stage I honestly don't know how you will do that and not run into overload problems on some recordings with the standard 30-40 dB gain, let alone feeding it directly into a 2V FS A-D.
(NB Shure pulled out of the cartridge business a year or twoc back)
Last edited:
See this is what is wrong with this thread, thanks to certain people un named any disagreement is seen as some sort of attack and raises hackles. Get a beer/wine as its friday and calm the **** down!!!
Everything in that article that Holman is writing from personal experience on the bench with his own measurements with test setup given is good and still very valid (if not more so based on some of the stuff being sold as MM phono stages these days). However in the case of overload he has chosen not to measure anything and do (as many others have done). The Shure V15III was a fine cartridge and I'm s(h)ure if John and Matti had had one of those instead of a poorly tracking ortofon MC the conclusions on TIM might have been different but shure published the design information on this as an advertorial with NO backup data for anyone to go back and check. What are these super hot and untrackable records? They rarely get mentioned beyond the telarc 1812.
The fact that I question where the overload margin number comes from doesn't suddenly make me a slathering, rabid objector does it?
First, none of this really applies to the ICs of the scale found in an audio device. I’ve seen dozens of devices fail EMC testing and not a single one was fixed by adding poorly grounded copper tape to the top of low power digital ICs. When you find an audio device with a 1B transistor CPU that dissipates 300W let me know. I mean, really, did you even read the PDF you linked? If so, I don't think you understood any of it.
Slide 4: EMI mechanisms for VLSI chips
The Heatsink acting as a monopole antenna against the first solid plane of the PCB (tall Heatsink). The Heatsink and the first solid plane of the PCB creating a patch antenna (wide Heatsink). The low inductance power distribution of the VLSI chip allows noise injection into the PCB, which re-radiate (especially through closely placed DC-DC convertors).
Hmm... guess what. NONE of those apply to most audio devices. Again, let me know when you sell an audio device with a 18 core Xeon CPU with a giant heatsink on it.
Oh yeah - check out how they propose to ground the heatsink, sure isn't a long wire.
More from your PDF:
Grounding will work at lower frequencies, but not above 1GHz Heatsink shielding works at higher frequency (grounding is implicit)if the contact is continuous, 360 degrees, creating a Faraday cage. Not effective above aprox. 1.5GHz (any grounding will be too inductive to really matter)
Putting copper foil on top of an IC with a shitty ground connection might even cause issues at very high frequencies, the very same ones this paper is trying to solve.
How is an IC going to radiate lower RF frequencies by itself efficiently? Do you know what it takes to make an antenna?
Further, I’m not taking about equipment I own personally, but equipment I have access to at work, which is much more modern than I’ve seen from his posts.
My point stands. Anyone who has done product level EMC troubleshooting would know that you’re not going to fix **** with some squares of copper tape and a pigtail. There also was no evidence presented that there is even a problem to solve.
You might as well put the copper foil on your head.
Last edited:
I don't understand. What matters is the dynamic of the signal. Do-you pretend that the one of a vinyl signal could be superior to the one of, even, a 16 bit sound card (>90dB) ?
What you need is just to adapt all the levels in a correct way in your chain.
Last edited:
Build it out of discretes, run it at high supply voltage (e.g. ±30V after all filter doohickies), and get all the overload margin you want.
Those preamp designer favorite R-Core power transformers from China only come in large wattage ratings, so order them with higher voltage secondaries. The reduced max current will still be plenty big enough to feed your shunt mode power supply and the rest of your preamp.
Those preamp designer favorite R-Core power transformers from China only come in large wattage ratings, so order them with higher voltage secondaries. The reduced max current will still be plenty big enough to feed your shunt mode power supply and the rest of your preamp.
I am quite, quite calm about the whole thing. On the other hand, clearly the same cannot be said about you.
The 105 cm/sec figure is from a Woody Herman recording Verve V8558 side 1 band 2. You have still not read the paper have you?
Show a stack-up of signal levels that indicates 20 to 30 dB overload margin is not a valid engineering design target given the points Holman discussed in his article.
I've shown my numbers (and the 5/25 cm/sec numbers are quite well documented if you root around on the web) as are the output ranges of commonly available MM carts.
Kindly do me the courtesy of doing the same before you dismiss something out of hand without even reading it and then launching into a personal attack.
https://www.analogplanet.com/images/vinyl-dynamic-range.pdf
I think you understand these people better than me:
Does Vinyl Have Wider Dynamic Range Than CDs? Here's Some Math | Analog Planet
We should not forget the huge quantity of noise in the LP compared with the cd!
Because I can't edit the previous post any further, let me add that I have a lot of respect for Richard and know that he has an impressive personal collection of test equipment.
I happen to disagree on this one point, which was presented without any measurements.
My personal collection ranges from 1.8mV to 20mV. I need something atypical, which is fine.
Not from where I am sitting
I have Holmans 1976 paper in front of me, no mention of a record. Which paper are you talking about? Interestingly the woody Herman is a mono recording. Big band is not my thing but tempted to get a copy for the research pile.
I never said the number wasn't valid. I said the source data was not backed up with any measurement method and people have just copied it. Which is odd
I read it, I did not attack you, I attacked the data as being a shure advertorial. Back to what I said about what is being wrong with this thread!!!!!!
People who rip vinyl tend to set -23dBFS to the 1kHz reference level even with flat ripping and that seems to work well with few complaints of overload. I have seen no evidence of music above that level on vinyl but happy to discover it and see if anything I have can track it. Anything more than that is for handling non-musical signals and how much you need depends upon topology and how that topology handles an overload.
- Status
- Not open for further replies.