Other than amps with protection circuitry activating to avoid clipping when played at higher level, the only components I can think of that introduce compression are speakers, the better the speaker, and the lower the sound level, the less it is likely to be. But other than when peaks are approaching the maximum limits of speaker or amp is it likely to be very significant, whereas compression at the mastering stage is a well known and deliberately applied effect, as discussed above, so I think that is far more likely to be the cause of the OP’s frustration.
Well getting the best streaming device you can afford.
A well sorted device should make more sense of a lesser stream than a lesser device making sense of a wide bandwidth high resolution stream ?
Perhaps not more “expensive” but more capable - although things usually follow as such.
For instance replaying a record with a notably wide dynamic range.
On one turntable the increase of information could cause the cartridge,arm,motor and phono stage to produce distortions due to signal handling and mechanical imperfection issues.
On another turntable and phono stage the signal could pass through intact and complete.
Sorry, but I think my analogy goes as far as being a difference between being able to handle extra information - to making sense of extra information.
Apologies for the confusion.
I suspect here that you’re focussing on the resolution of information, as opposed to the range of sound level from quietest playing of an instrument to loudest fortissimo which, for example, might be a 60dB difference in levels in a classical piece. Recorded and distributed without compression that would remain the range. For optimal radio play in a variety of playback (non hifi) environments that might be compressed down to 30dB or less - that is squashed in volume range, but with all detail still there - all the information still present, but simply with less relative loudness between different sounds. 16 bit digital (CD or streaming) can handle the full 60dB so no need for compression, and you can therefore replay retaining the natural relative levels of everything. Vinyl in theory can handle 60dB but that would be losing the quietist sounds into surface noise so may be compressed to, say 50dB, again with all detail still there, but squashed in relative sound level.
N.B these dB levels are just indicative for this example.
I’ll try and find, i am fairly sure it was a recording of Chopin piano waltzes where I heard and it grated last weekend
S
Yes where loudness is simply modulation based, which given our non linear frequency response in our hearing and our current reproduction technology has its limitations. Often in music production adaptive dynamic compression for different frequencies is used to make the recording appear more dynamic but equally loud, energetic and emotional.
Do you use the feature in Roon for volume leveling?
No, I switch off all Roon processing when I use it… do you?
I did awhile ago and it works well. Thought it was worth it for those times when I’m listening to random playlists etc. However getting the best SQ from roon I turn it off. This thread has me thinking of experimenting with it on again.
Coincidence - I received this from Roon today:
It got me thinking that even Roon is doing some DSP!
(I’m a Roon beginner and not using it much…)
1 Like
I’ll check this out. I turned off this feature for SQ but I’ll try it again.
1 Like
JRiver has an analyze audio feature. I believe it runs automatically when I rip a track out (maybe there’s an on/off toggle somewhere). You can also run it manually for other imported files.
I’m curious about something @Innocent_Bystander posted re: the 60dB range between pianissimo and fortissimo in classical music. For something like a Mahler symphony. JRiver might calculate a dB Range in the mid teens. The same appears to be true for Mahler recordings in the Dynamic Range DB. Does that indicate compression during the mastering process?
I think there are multiple questions there: the first being is the version you are playing definitely the same version you’ve seen in the DR database? by version I don’t mean the performance, I mean released recording with its mastering)? If yes, the next would be what is jriver actually doing when it is assessing and giving you a figure for dynamic range?
Sorry, that is answering a slightly different question, the first is in effect your question, but the second is part of it. If jriver is giving a true measure, then it would seem your copy is compressed compared to whatever the DR database figure related to.
JRiver and the DB database use slightly different algorithms (to the best of my knowledge), but in circumstances where I can compare the same masterings, the results are generally the same (sometimes 1 dB difference). I have checked this out for dozens of recordings.
Having said that, I may have had a basic misunderstanding of what DR actually measures. I thought it was the difference between the loudest and softest passage in the track. But apparently not. Here is the JRiver’s explanation:
Crest factor DR measures how many decibels there are between the average loudness of the track and the loudest sound on the track (the peak). The dynamic range database uses crest factor DR measurements. The dynamic range database do not use crest factor in place of actual DR measurements. [This last sentence confuses me - a bit of Googling seems to confirm that the DR Database does use crest factor DR.]
By contrast, R128 Dynamic range measures how many decibels there are between a track’s 95 percentile for loudness and the track’s 10th percentile for loudness. This sounds technical, but in laymen’s terms, it means that the measure screens out the very loudest and very softest sounds from its analysis, to avoid allowing brief sudden loud sounds and momentary silences to affect the measure of dynamic range.
JRiver’s analysis provides both.
https://wiki.jriver.com/index.php/Dynamic_Range
If that is the case, it would explain why the measurements from both the online database and JRiver are much smaller than the 60 db between the loudest and softest passages.
Finally - I don’t use Roon, but Google led me to this thread as well.
1 Like
Yeah I turned it off as it caused levels to jump between continuous tracks on certain albums which was annoying and interrupted the musical appreciation
1 Like
Indeed loudness is aggregated over a period of time, typically 3 seconds with LUFS where as dynamic range can be over a very small period of time. Loudness is quite different to dynamic range and the ubiquitous adaptive frequency based compression used in modern productions.
If a production sounds punchy, dynamic and powerful, it will almost certainly had had adaptive frequency based compression applied to it at maxing or more likely mastering irrespective of the LUFS level.
LUFS compression without adaptive frequency based compression can make a track sound flat, soulless and non engaging.
R128 is about loudness normalisation… here is the EBU guide on how it should be used
And here is the R128 S2 specification for streaming loudness normalisation
Here is a guide for optimising production mastering for popular streaming platforms by a popular compression and loudness processor manufacturer, quite insightful. Written for developing production engineers.
So you can see a given master can sound different on different music streaming platforms… and why.
2 Likes
Just noticed that ProStudioMasters has started to include DR and LUFS info for their releases.
This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.