Even a quick look at the links in this thread and others, shows that R128 numbers HAVE to be read in Conjunction with the Waveform…whereas for several years now, many listeners have become used to hanging their proverbial hat on DR’s above 12 as “good” and below “10” as bad…and that simply is not the case, necessarily
Please, please read more about R128…about how and why it has to be read hand in hand with the Waveform…and not used like Crest DR as some kind of false absolute measurement…I think the more you read about it, the more convinced you will become about why R128 is the way forward…and why Crest DR doesn’t deserve the solo position that it has built up
what does the DR 128 value with the waveform tell us about the Chicago song that the wavefrom would not tell us on its own. What is this read in conjunction that you refer to?
I disagree, and yes I understand 100% about DR range. This is to DISPLAY DR, nothing to do with, or changing the algorithms for R128 volume processing.
It’s because people have become used to using just ONE tool [Crest DR] to describe Dynamic Range…which is wrong because that DR number can and often is misleading
On the other hand, R128 essentially encourages us to use BOTH the Waveform AND the R128 value to be used alongside each other
Take the Chicago track Waveform above…the Waveform shows that Peaks don’t approach Clipping [which is obviously a good thing]…whilst the R128 shows us the Dynamic Range [time weighted over the track duration]…using the Waveform and R128 values TOGETHER will give us a more accurate picture of our music…rather than relying on the Crest DR method on its own, as it can be misleading…for the reasons stated by Danny and the other links posted
Yes and the points I raise have nothing to do with Volume Levelling either
If Roon had decided not to incorporate Volume Levelling, then I still think R128 plus Waveform is the way forward
But equally I could use the argument that most people think that Volume Levelling based on R128 than when using Crest DR…to show that obviously R128 does a better job of measuring how we PERCEIVE Loudness…which in of itself should be persuasive of why R128 is more informative about perceived Loudness and DR
Loudness and DR are two different things…but they are linked…which is why R128 PLUS the Waveform is more informative…and therefore more effective at volume levelling as well [which you seem to accept by the way]
May I suggest, if displaying/analysing standard DR be difficult to implement, then a request for the ability to chose to display R128 Album Gain as well, or instead of R128 DR.
Oh, please don’t do that or make it an option.
I don’t want that numbers in my face all the time. Same for bitrate and other stuff.
After I decide to have an album in my collection, I want to enjoy the music, maybe get some background nformation about the album or the artist. BUT NOT about technical details of the files. I can understand that it can be important for buying the right version of an album. But after that it’s only music, not nerdy numbers.
So, I’m wondering if we are all confused because we haven’t figured out the basic aspects of what the numbers and the waveforms are telling us.
When a track’s “true” dynamic range is expanded (and sadly brickwalled–meaning it is stretched out so much that it clips and the true peaks are now beyond the top limit) the average level is higher (louder) so a low Crest dynamic range correlates to loudness and compression.
The R128 DR standard is less sensitive to the “average” level and is meant to capture only the dynamic range of the track. That’s why a brickwalled track and a well-mastered track can have similar DR’s even though they sound different. This goes to show that R128 DR is a poor metric for actually determining the loudness and compression of the track (which it is not intended to be) and why the Crest DR might be better for that purpose.
However, in theory, DR should not be used as a surrogate for measuring compression and loudness. It seems a measure of DR relative to the average level would be a better measure of the compression. So, in the case of the brickwalled track when you see a low R128 DR number and a high average volume level that should be a warning sign! A low DR but low average volume level is probably ok. I’d like to think that If the R128 DR is large you likely don’t have a compressed track but Adele’s “Hello” easily proves this wrong. It shows an R128 DR of 17 but it is clearly compressed to hell (it seems R128 DR is skewed by the relatively long quiet portion of the track). So, again, DR relative to volume level might be a better choice…a high DR and a high average volume level is likely compressed. A high DR but low volume level is probably ok.
With the R128 standard, volume level seems to track pretty well to the track gain. I’m finding Roon’s volume leveler works pretty well.
If R128 DR and R128 Program Loudness are shown together I think you’ll have informative data points…how many DB dynamic range across the album, and how loud is it. The louder the more compressed and/or limited the proram.
Volume levelling in Roon is a straightforward and high quality 64-bit floating-point (negative) gain operation (with appropriate dither) which reduces the amplitude of the waveform. All good so far.
The potential impact on sound quality is that the DAC is presented with a “quieter” digital signal which is typically less than optimal compared with the unaltered signal. Technically, “a lower resolution digital signal is being presented to the DAC”. There’s no such thing as a free lunch with this type of volume levelling. However, you should let your own ears decide.
However, in the near or distant future, it may be possible for RoonReady DACs to receive the volume levelling offsets via RAAT so they can preserve digital resolution (i.e. by implementing a hybrid digital/analogue volume control).
Another future possibility, if we expose the levelling offsets via the (in Beta) Roon API, might be to implement the offsets using an external volume control; but that’s non-trivial for the time being!
No. It changes the volume that the track is played at, based on the volume change calculated for the whole album. But you can do it on an individual track basis as well (I think), if you really want to.
Now that is interesting. It doesn’t change the volume between tracks within an album, but, I assume, when radio is on or a playlist of tracks from multiple albums is played it knows to adjust each track?
Just under the first screenshot in “enabling volume leveling” there’s a description of what we do here:
Roon supports four Volume Leveling modes:
Off - disable volume leveling for this zone
Track - Performs adjustments on a track-by-track basis. This produces the most consistent volume level during a playback session, but may result in unpleasant jumps in volume during playback of a gapless album.
Album - Performs adjustments on an album-by-album basis. This means that the entirety of an album plays back with the same adjustment, and there are no perceived volume jumps at track boundaries.
Auto - Uses track adjustments when playing adjacent tracks from different albums, and album level adjustments when playing adjacent tracks from the same album.
By default I believe volume leveling is set to “Off”, “Auto” is what I normally use, reasonable people can obviously disagree on what’s best for their own personal collection/gear/habits/etc.
