Digital vs Vinyl

[Cbdb2]

you might wanna look into upward compression. this will make quiet parts louder without compromising transients like a regular compressor

How does a regular compressor "compromise" transients? How does upward compression not? They both boost (theres usually make up gain in regular compressors) the sound below threshold and when a transient hits they both decrease the gain. And this is totaly dependent on the attack times. Most compressor settings let the transients thru. Limiters don't. I dont doubt they will sound different, but if one type was better than the other you would think it would be more popular in recording studios. And most, if not all the iconic studio compressers are downward types.

 

[Cbdb2]

I learned something. New York compression is upward and done by mixing a downward compressed signal with the original.

More info here:

"If the above mentioned technique seemed laborious, and you don’t feel like trying to find a dedicated Upwards Compressor, you can still achieve upwards compression using a normal compressor. By reducing the amplitude of the Peaks using a high ratio setting, and then boosting the makeup gain of the compressor, it effectively does the same as Parallel Compression. This technique arguably gives you less control over the upwards compression process. Personally, I use Parallel Compression when I don’t use a dedicated upwards compressor, because I feel it gives me more control over mixing the signals together."

From https://soundphysics.wordpress.com/...ssors-upwardsdownwards-compression-explained/

 

[dasdoing]

How does a regular compressor "compromise" transients? How does upward compression not? They both boost (theres usually make up gain in regular compressors) the sound below threshold and when a transient hits they both decrease the gain. And this is totaly dependent on the attack times. Most compressor settings let the transients thru. Limiters don't. I dont doubt they will sound different, but if one type was better than the other you would think it would be more popular in recording studios. And most, if not all the iconic studio compressers are downward types.

maybe I should have said peaks instead of transients. If you just want to make quiet sounds louder the upwards compression is much less destructive. think about it, when there are no quite sounds, the upwards teqnique will do nothing, while the normal compressor will act all the time. in production a compressor is used as a "sound designer", in this case you don't realy want to change the sound, right? that's the reason for my suggestion. personaly I have never tried this out btw. my room is quite quiet.

 

[Cbdb2]

A normal compressor does not act all the time, The signal has to be above treshold. So it does nothing to the quiet parts except add make up gain. So yes it affects the peaks ( but most transients are allowed thru ( not limiter attack times)).

So whats worse, affecting the peaks or the quiet parts. And with the loudness wars there are no quiet parts in most music so an upwards compressor will not even kick in.

 

[Phorize]

I am curious if any of you who prefer digital -- think there may be some limited situations wherein there would be objective reasons to choose a vinyl record over the CD counterpart. For example, what if the CD had too much dynamic range? In other words, on the CD version, the listener could not find a volume level that would be satisfactory for both soft and loud passages.

Some other thoughts that I have involve the channel separation. Some listeners like mono recordings and mono is considered by some to be great for a soloist performing without other musicians. So when playing back in stereo (rather than mono), perhaps less channel separation could be better for certain types of music that involve solo performances?

Consider a stereo system wherein the listener had positioned the two speakers too far apart. Would less channel separation then be desirable to fill the center void? I realize in some cases the simple answer might be to just move the speakers. However, this is not practical in all situations.

What about a case wherein the digital mastering (or recording and mastering) was done poorly? It seems there are a lot of cases such as that. How do you think a vinyl version that was recorded and mastered very well (as well as the medium could allow), would compare with a poor digital version?

Do any of you think the CD (16 Bit / 44.1 kHz) as a medium (in general) has too much dynamic range? I have seen some post in ASR that praise dynamic range and condemn compression. Yet, other posts have claimed that compression is absolutely necessary to some recordings or we would not be able to hear the difference between soft and loud passages. Perhaps it is a case where the very competent studios know whether to use it or not?

Some people claim that when they have recorded their CDs on analog tape, that the tape playback sounds better. This could also relate back to the hypothesis that CDs have too much dynamic range.

I can’t imagine that a scratch mix battle would be very compelling if performed with a digital sources, at least not if the knob twiddling that 20’s DJ performance been reduced to is anything to go by.

 

[John Dyson]

A normal compressor does not act all the time, The signal has to be above treshold. So it does nothing to the quiet parts except add make up gain. So yes it affects the peaks ( but most transients are allowed thru ( not limiter attack times)).

So whats worse, affecting the peaks or the quiet parts. And with the loudness wars there are no quiet parts in most music so an upwards compressor will not even kick in.

About 'attack times'. Almost all normal, commodity consumer recordings have a compression applied to them with a dynamic attack time of approx 2msec to 40msec (contoured to the waveform) at HF, and 4msec to 80msec (contoured to the waveform) at 3kHz and below. The release time of the ubiquitious compression is 80msec (contoured to the waveform) for the HF, and equivalently 160msec for < 3kHz. The compression ratio is approx 2:1, segmented into 10dB ranges, except at very low levels where the ratio increases a little, and above 9kHz. There is little or none of the compression that I am speaking above about -20dB at midrange, threshold just a little lower at LF (<80Hz), and the threshold at >3kHz is approx -10 or so dB. (I could be off a little, I haven't recently double checked.)
For an existence proof of the compression, and IT IS realtively standard -- compare the general sound of the ubiquitious compressed consumer stuff with a microphone with a mixer, how quiet the ambience is with the unprocessed mic/micer and a natural ambience decay. For the compressed consumer stuff, there is a sustained decay, and especially on old material there is a heck of a lot more tape hiss than tape recorders of the day. There is often measurable hiss on so called digital recordings (often not digital all they way through), when the background hiss should be very low. This is all about the compression that I described.

It is amazing how audio can withstand a super well designed compressor system. Slow attack/release doesn't cut it -- even though such attack/release can be useful. The compression system that I speak about is not just one compressor, but is layered so that the attack/release doesn't have to happen with just one attack/release time. Since the attack/release is carefully contoured, the actual distortion is very unlike conventional designs... Who designed the scheme? One of the most well known audio people of all time.

Very fast attack is in the realm of both compressors and limiters, but limiters do go down to the microsecond or 'synchronized' timeframes. There is one h*ll of a lot of processing on consumer recordings before they are distributed, and this is not necessarily the domain of people doing mastering -- but there is apparently one additional step afterwards.

Demos upon request. (This is nothing about selling anything -- purely technical/reverse engineering.)

 

[John Dyson]

In a strict sense, yes. However, it isn't necessarily audibly more dynamic. For example, if ultrasonic bursts are added to a signal, it will measure more dynamic without being audibly different. Maybe clipped peaks can trigger a cartridge resonance producing such an effect.

There is a lot of compression (depending on freq) happening below -10 or -20dB. The compression is so fast that it can sneak in between very short quieter times (2 to 40msec attack time, contoured to the waveform, 40msec to 80msec decay time, contoured to the waveform.) At very low levels, there is actually a doubling up of the attack/release, and above 9kHz (actually around 6kHz), the attack/release also slowly speeds up.

Very early vinyl doesn't have this compression, but I have found (with inputs from skeptics), that this compression started BEFORE CDs (my previous hypothesis about when the scheme started was wrong.)

The ubiuquitious compression DOES NOT affect peaks at normal levels, and has relatively minor effect on DR type meters (and the lower level SoX measurements) when the signal level is very low. Material with normal dynamics only show a moderate change (decrease) in DR when the specialized compression is applied. When comparing with difficult-to-find virgin material, people WILL hear an increase in density, more pronounced/enhanced bass, and much more swirl in the highs/cymbals.

(I do have a related project that undoes the compression, now complete after 8yrs, and works as close to perfectly as the original compressor itself , and have started releasing the related technical information.) I'll be describing things in more detail in the next month or so -- fixing perfection is kind of silly.

Without existence proof about the ubiquitious compression, or with my earlier works (because of incredible compliexty), I would understand skepticism, but NOW there is more than enough evidence that substantiates the details.

If one doesn't like MQA, one should really hate the historical and ongoing compression against consumer recordings. The compression is all about business and protecting IP. Too bad that the IP is being protected against the consumer who paid for assumed clean copies of it. Again, this is BUSINESS -- some people think 'conspiracy'. One persons effective industry standardization is another's conspiracy, I guess.

* People in industry, with dependency on good relations with their employers/distributors will tend to ignore the matter, but with high quality equipment, and perfect media, the time has come to give the consumer the *good stuff*.

 

[dasdoing]

I think nowadays the drums is the culprit for all the compression. that's where engeniers start to compress...then they are forced to sonicly match the rest. listen to this:

the part in the beginning the drums don't even sound real....but this is the sound which is requested. compare to the sweet drum sound later at 1min6s when they only seam to have a slight level match. how can anybody prefer the sound from the beginning?
that's why I love 70ies funk

 

[John Dyson]

I think nowadays the drums is the culprit for all the compression. that's where engeniers start to compress...then they are forced to sonicly match the rest. listen to this:

the part in the beginning the drums don't even sound real....but this is the sound which is requested. compare to the sweet drum sound later at 1min6s when they only seam to have a slight level match. how can anybody prefer the sound from the beginning?
that's why I love 70ies funk

Yea -- I can hear the compression, and it is sad that the specific style of compression is so common. Unfortunately, there is also gating in the recording (for the voices to suppress background noise during quiet times) along with an apparent gain modification after the likelyFA compression.* However, after some guessing, I gave 'decoding' the recording a try, but I didn't name the files 'DECODED' as I usually do, but instead 'proc' which only means 'processed'. It isn't 100% clear to me that this interview is FA encoded, but there is some probability that it is. (You Tube also appears to be a culprit, or at least uses already encoded recordings.) From the above youtube, I have created two before/after segments -- judge for yourself. (I have plenty of Dropbox space nowadays, so will keep these online for a few weeks): (also -- Dropbox has a notoriously poor online .flac player, so for the best comparison and quality, download first!!!!)

* note from above: the apparent gain change can be detected by matching the knee in the compression curve. There isn't a strong knee, except one just strong enough that can sometimes be used to match the needed calibration levels.

https://www.dropbox.com/s/ad0cvo3rnyviu66/fd-raw-0s.flac?dl=0
https://www.dropbox.com/s/7hipdt52s373ife/fd-proc-0s.flac?dl=0

https://www.dropbox.com/s/k782oeap5wgymfj/fd-raw-60s.flac?dl=0
https://www.dropbox.com/s/vewevnh3lqmkd3m/fd-proc-60s.flac?dl=0

I do detect some improvement in dynamics, and sound of the cymbals but that is MY opinion. We each have our own perception, and might actually prefer the compressed sound. In my case, I dislike the compression so strongly that I spent LOTS of time writing a very advanced decoder for the application.

 

[board]

For anyone interested in corroborate this in a practical way, because all I’m seeing through theory is discrepancy between members:

Listen to RHCP album Stadium Arcadium CD (Redbook Standar) vs the same album digitalized from vinyl. They are practically two different albums. There is harsh clipping in lots of effect on the CD release that are more atenuated or non existing in the Vinyl release. The Dynamic Range difference is night and day, Vinyl release sounds so less harsh, more natural, like if they removed some digital processing from the mixing/mastering stage, Anthony Kiedis voice sounds very different in both mediums.
Since its a dual CD conceptual album I recommend just testing with the first tracks, that will do the trick...Danni California, Charlie, Stadium Arcadium, works for the comparison.

I don't know if you know this, but the CD version of "Stadium Arcadium" is mastered by Vlado Meller, and it's clipping a lot, and EQ'ed in a certain way, while the vinyl version is mastered by Steve Hoffmann, and is EQ'ed a different way.
So, the reason why that particular album sounds so different on the two mediums is because it's two radically different masters and not because of dynamic range compression, nor because of the medium itself.
I certainly also prefer the vinyl edition of that particular album, as Vlado Meller EQ'ed it, and clipped it, in such a horrible way.


I have actually gone about this issue in a practical way:

1: I have compared around 800 albums on vinyl and CD - some from different masters, others from the same masters. When the masters were the same, the real difference seemed to come down to the frequency response of the phono cartridge.

2: Up until the other day I used to say that dynamic range compression completely destroyed music.
So what happened the other day? I took four dynamically uncompressed songs, by Laurence Juber (acoustic guitar, bass and percussion), Pallbearer (doom metal), and Olafur Arnalds (soundtrack music), and then I dynamically compressed them myself: Laurence Juber went from DR14 to DR6, Pallbearer went from DR10 to DR5, and Olafur Arnalds went from DR12 to DR6.
I then level-matched them and did a blind test and could only hear any difference on one of the Laurence Juber songs in a particular part of the song.
I was completely shocked because I had expected to be able to hear a major difference.

So, I'm not trying to be annoying or antagonistic, but I have to ask you:
Have you also done the same, taken an uncompressed song, compressed it dynamically yourself, and then done a blind test?
Admittedly, my assumption is that you haven't. I have yet to meet someone who says that dynamic range compression completely ruins everything, therefore vinyl is much better, etc. and who has actually done the test I did.
Also, I have yet to meet anyone who says that digitization audibly changes the signal dramatically who has actually done a loop-back test (which I have also done). But that's a different topic.

I would be happy to send you these songs, and you can see if you can pass an ABX test .


As for the frequency response of the cartridge, my experience tells me that often the most expensive cartridges are the ones with the least accurate frequency response (although some expensive ones are certainly very accurate), and then when a vinylphile hears this expensive cartridge he simply has a preference for that particular inaccurate sound, which is radically different from that of a CD, and then the states that vinyl is a better technology than digital, simply because he likes the changed frequency response.
So the expensive ones are often being sold for their sonic signature, not for their accuracy. Lyra cartridges are a great example of this, and they're among the most expensive cartridges on the market. Although their distortion level is low, their frequency responses are very inaccurate and have steep rises in the highest frequencies, which adds sheen and sparkle to the music, thus giving the impression of "greater detail retrievement than a CD", when it is in fact simply EQ.

 

[MakeMineVinyl]

The reason to buy vinyl is because you subjectively prefer the sound of it. Vinyl is not a perfect medium by far (neither is analog tape), but the distortions are largely euphonic in nature. Dynamic range really has nothing to do with the medium; many CDs have only a few dB dynamic range while the Telarc 1812 Overture on vinyl has gobs of it and vice versa (although digital has a wider dynamic range potential than any analog media).

 

[paulraphael]

... I have actually gone about this issue in a practical way:

1: I have compared around 800 albums on vinyl and CD - some from different masters, others from the same masters. When the masters were the same, the real difference seemed to come down to the frequency response of the phono cartridge ...

Very interesting post. Confirms a few suspicions of mine, but like most people, I've never done any kind of systematic comparison.

Especially interesting on the topic of level compression. I've heard recordings that sound terrible because of overzealous compression, but I've also heard many (that have earned complaints from reviewers and commenters) that sound fine to me. In my own informal comparison of different CD masters, I typically keep the quieter master, just on general principle. But in many of these cases, even when I casually level-match them, they sound equally good. At least in terms of dynamics.

Edited to add: have you found any expensive cartridges with rolled-off highs? I've assumed that this was a typical signature of vinyl, which would play to the idea of vinyl warmth, and of CDs being too bright and harsh.

 

[board]

Very interesting post. Confirms a few suspicions of mine, but like most people, I've never done any kind of systematic comparison.

Especially interesting on the topic of level compression. I've heard recordings that sound terrible because of overzealous compression, but I've also heard many (that have earned complaints from reviewers and commenters) that sound fine to me. In my own informal comparison of different CD masters, I typically keep the quieter master, just on general principle. But in many of these cases, even when I casually level-match them, they sound equally good. At least in terms of dynamics.

Edited to add: have you found any expensive cartridges with rolled-off highs? I've assumed that this was a typical signature of vinyl, which would play to the idea of vinyl warmth, and of CDs being too bright and harsh.

As for "I've heard recordings that sound terrible because of overzealous compression", were you able to do any kind of test that showed it was the dynamic range compression that made them sound terrible, or did you simply assume that it must have been the compression?
I'm not attacking you, just asking out of curiosity .

As an example, I find "Villains" by Queens of the Stone Age to sound quite annoying, and I figured that it was due to the massive amount of compression, but after my experiments the other day I've started to think that it most likely is rather due to how it's produced, mixed and EQ'ed.

As for the cartridges, I have seen some with a rolled-off top-end, but I actually think that the claims of vinyl sounding "warm" is not so much about frequencies above 10 kHz, but rather around the harshness region, where many cartridges have a light dip as well.
Here's an example:
https://www.aktivstudio.at/media/pdf/12/49/02/EMT_JSD5-6-Hifirecords-02-2013.pdf

A few other examples are Rega Exact, Denon 103, and Audio Technica AT33PTG/II, the latter being my current cartridge. The Rega Exact has the biggest dip of those three though, and it does seem like this dip is more common in MM cartridges than in MC's. All of them also have a spike around 10-12 kHz, which is something almost all phono cartridges have, although sometimes this spike peaks at a higher frequency, at times higher than at 20 kHz.
What I find much more common than a roll-off in MC cartridges is a rise around 10 kHz or higher.
Here are two examples:

Transfiguration Proteus (you have to press the zoom button to see the frequency response):
http://transfigurationcartridges.co...014/09/Transfiguration-Proteus-Review-HFN.pdf

Sumiko Starling:
https://www.hifinews.com/content/sumiko-starling-cartridge-lab-report

I wrote a long post about frequency response in cartridges on Youtube on the video "The Truth About Vinyl". I can post a link if you're interested.

Here are some examples of roll-offs above 10 kHz, although it should be said that I'm a bit wary of the measurements by Hi-fi News, as I've seen several of their measurements roll off a lot above 10 kHz where the measurements on other websites didn't roll off for the same cartridges, and I also saw someone comparing test records, which seemed to conclude that the test record made by Hi-Fi News (which they would presumably use for their own measurements) rolled off the top. Lastly, to make it even more confusing, it seems that lately Hi-Fi News' cartridge measurements have seemed more accurate, so I'm wondering if they started using another test record.

Koetsu Urushi Sky Blue (from Hi-Fi News):
https://www.absolutesounds.com/pdf/main/press/2013_07_Koetsu_UrushiSkyBlue_HFN.pdf

The same cartridge measured by Hi-Fi Critic (less rolled-off):
http://www.bm.rs/Koetsu/Koetsu Urushi Sky Blue - HiFi Critics March 2009.pdf

Koetsu Onyx Platinum (from Hi-Fi News):
https://www.absolutesounds.com/pdf/main/press/HFN_Koetsu Onyx Platinum_lowres.pdf

Koetsu Vermilion (from Hi-fi Critic):
https://www.absolutesounds.com/pdf/main/press/HiFi Cricti VOL5No2 Koetsu Vermilion.pdf

EAT Jo No 5 ("the talk of 2018's High End Show in Munich") (from Hi-Fi News):
https://www.hifinews.com/content/eat-jo-no5-cartridge-lab-report

TechDas TDC 01 ti (a bargain at £8750) (from Hi-Fi News):
https://www.absolutesounds.com/pdf/main/press/TDC01Ti_HFN_Sept2014.pdf

 

[board]

Which I mostly heard on vinyl. I'll give Manfred Eicher credit, ECM vinyl is of a high standard. That's Keith Jarrett BTW. No, I am not a fan. Partially because of the seasick sound of sustained piano notes and chords as presented in vinyl. But mostly because if I want pomp on the piano, I'll stick to Liszt.

Manfred Eicher crossed over into "New Age" music. The defining characteristic of new age music is long sustained notes, the sorts of sounds better presented by digital sources than LPs or cassettes. That is the 'droid you are seeking.

If you like I can send you a recording of two songs from the LP "Passengers" by The Gary Burton Quartet on ECM. They sound terrific.
When I started replacing my records with CDs I also bought this album on CD, but the LP sounds quite a lot better, as it has a bit more bass, but also quite a lot more high frequency content, so the CD sounds quite dull.

 

[Robin L]

If you like I can send you a recording of two songs from the LP "Passengers" by The Gary Burton Quartet on ECM. They sound terrific.
When I started replacing my records with CDs I also bought this album on CD, but the LP sounds quite a lot better, as it has a bit more bass, but also quite a lot more high frequency content, so the CD sounds quite dull.

It's alright, I'm cured.

 

[Frank Dernie]

Here are some examples of roll-offs above 10 kHz, although it should be said that I'm a bit wary of the measurements by Hi-fi News, as I've seen several of their measurements roll off a lot above 10 kHz where the measurements on other websites didn't roll off for the same cartridges, and I also saw someone comparing test records, which seemed to conclude that the test record made by Hi-Fi News (which they would presumably use for their own measurements) rolled off the top. Lastly, to make it even more confusing, it seems that lately Hi-Fi News' cartridge measurements have seemed more accurate, so I'm wondering if they started using another test record.

Paul Millar wrote somewhere that he has a large collection of test records collected over the years.
I have no idea what the test records published by magazines are like.
For measurements we had test records cut by Bruel and Kjaer which had a "guaranteed" deviation from flat but, iirc, it was only guaranteed for 5 uses (I could be have been more but it wasn't many).
I wasn't measuring the cartridge though but the influence of turntable and arm performance on its output.
Given that the scales used for cartridge FR are quite course the main thing I take from measurements is that none of them are that good but even that with a pinch of salt.
The peak in the hf response is a resonance. If it is the natural frequency of the stylus effective mass on the vinyl compliance - which is the high frequency bandwidth limit - it seems very low, we were doing better than that 45 years ago!

 

[board]

It's alright, I'm cured.

No worries .

 

[board]

The peak in the hf response is a resonance. If it is the natural frequency of the stylus effective mass on the vinyl compliance - which is the high frequency bandwidth limit - it seems very low, we were doing better than that 45 years ago!

I believe that many cartridge manufacturers nowadays are deliberately producing non-neutral cartridges because that's exactly what the buyers want - vinyl lovers want their vinyl to sound different from digital, i.e. not neutral. Then they can claim "vinyl is a superior technology, because I like it more".

Was it your job to measure tonearms? If so, where did you work doing this?

 

[Frank Dernie]

Was it your job to measure tonearms? If so, where did you work doing this?

I was part of a research group set up at Garrard to investigate how to get the best performance from a record player efficiently.
It was set up in the mid 1970s with the first laser vibration measuring kit I had ever seen and state of the art measuring kit.
I learned a lot but, sadly, very little came of it commercially.
The management were more interested by the styling than the engineering in the end, and the company went under following poor management decisions (actually I think the MD was making decisions to make the balance sheet look good short term for his next job move) so few of the findings found their way into production.

As far as I can see nothing new has been learned about the physics of what how a TT system works since then and quite a lot forgotten. Some of what we were doing was ground breaking and difficult and would be quite easy to prototype now using distributed mass FEA but it seems nobody is actually doing it.

There are plenty of technical non-sequiturs in TT "common knowledge" simply because static explanations have been made, and accepted by the business and hence the market, for a dynamic system.
It never ceases to amaze me.

Anyway, what we looked at was the whole dynamic system and my belief is, and was, all about minimising disturbance either side of the transducer, which largely boiled down to making sure there was no anti-node at the headshell at audio frequencies.
I measured that simply by looking at the cartridge output on the basis that any shortcoming (and there are always some) mustn't influence the cartridge output.

The whole LP business is fashion industry today IMO not technically based at all.

I believe that many cartridge manufacturers nowadays are deliberately producing non-neutral cartridges

Some have always produced non-neutral cartidges tuned to taste. Some have always tried for neutral, I think now it is such a long time since proper engineering techniques were used to measure stuff the equipment needed, specifically accurate test records, are no longer made by B&K.

 

[magicscreen]

Vinyl ha bigger dynamic range if you listen to it.
There is some confusion about the 1 bit = 6 dB rule, which is incredibly wrong.
So CD has no 96 dB because of the 16 bit.