I was curious as to what formats are more taxing on the hardware/software?
Is flac less easier on the system because it’s uncompressed or mp3 because of its file size? The main reason I ask this is I have seen slow downs in my system. I have tried ssd internally, and a wide variety of usb with my latest being a 3.2 2x2 ssd usb 256 gigs.
It is not uncommon for me to have 3-4 disc’s going at once and I notice my music can start to skip in the ears buds and on the audio graphs. If you wanted to replicate this experience thr. Play a couple of tracks and then start scrolling through your music with a passion and you should at least see the visual hiccups.
Another thing I have noticed is one of my decks while I am loadinga track has to complete restart that deck and will just eventually boot the track with sync turned off.
I have a feeling that whatever my prime 4+ is doing is pushing the system tomits limits. I was just wondering if I could lessen the load on it using less data heavy tracks, or compression heavy tracks, or does it even make a difference? Thank you for your time
So is that easier on the system or is the larger file size harder on the usb joints?
Most of my tunes now have been flac since i find they sound better or at least display more information on the audio graphs. Have you experienced any audio or visual skips?
I don’t know why you experience glitches on your Prime 4+, but decoding FLAC (which is lossless compression) is computationally very lightweight. Any modern processor in our standalone units can decode dozens of FLAC streams in parallel without issue. In fact, MP3 decoding can sometimes be (slightly) more CPU-intensive because of the psychoacoustic modeling, … whereas FLAC decoding is more like unzipping a file …
Of course I don’t know how exactly the Prime 4+ handles loading and decoding internally, but I find it unlikely that the skips you describe are caused by the file format … I use FLAC exclusively, for me it’s definitely the best format. The only problem I sometimes have is that older Pioneer DJ decks like the old XDJ-1000 cannot read FLAC files …
Yes, I would think so. Audio decoding in general, whether FLAC or ALAC, is very easy on the CPU and normally handled without effort. FLAC and ALAC are basically the same kind of format: both are lossless compression, just different containers/codecs. It’s really effects processing like heavy convolution reverbs, time-stretching, or key-shifting that can sometimes stress hardware, not simply loading and playing back a lossless file …
A Prime 4+ chip is literally hundreds to thousands of times more powerful than what ran the first MP3 players. If a 20 MHz DSP from 1998 could handle MP3, a 1.5 GHz quad-core from 2023 can handle FLAC and four decks and effects, and still be chilling.
Are you saying in general there is nothing pushing the system to its limits?
My other concern would be is there a limit to the usb drives? If i was playing 4 flac songs at once is that too much data at once for usb? I know it was for my cheap usb, I even had a error from the system saying the storage was not up to snuff.
One other thing I have never done is a full reset of the board. Is that something I should do after having 6 updates?
Out of curiosity I wrote a small script in Python that loads a FLAC file, converts it to uncompressed WAV in memory (as a NumPy array), and measures the time it takes. Here’s the result:
Time to load and convert FLAC to WAV in memory: 0.214173 seconds
A fraction of a second for a CD-quality track (16-bit, 44.1 kHz, 2:12 minutes … Ludwig van Beethoven’s Adagio from Op. 2 No. 3). Of course this is on an unfairly fast Intel Core Ultra 9 CPU, and the code is not optimized at all, but still this is really fast. You could try this on slower systems or with other libraries … this was just to get a rough idea … So, yes! I would say, FLAC decoding is really trivial compared to what modern DJ hardware can do.
Here is the code:
import soundfile as sf
import numpy as np
import time
# Path to FLAC file
flac_file = "test.flac"
# Start timer
start_time = time.time()
# Load FLAC
data, samplerate = sf.read(flac_file)
# Convert to WAV in memory ... just as a numpy array ...
# Normally sf.write() would save to disk, but here we just hold it in memory
wav_data = np.array(data, dtype=np.int16) # convert to typical WAV format
# Clarification: converting to *int16* is just simulating WAV PCM, so nobody nitpicks 😉
# Stop timer
end_time = time.time()
print(f"Time to load and convert FLAC to WAV in memory: {end_time - start_time:.6f} seconds")
I don’t own a Prime 4+, so I can’t test its exact boundaries, but in general the limits aren’t from decoding audio … they’re more about multitasking: managing several drives, database lookups, drawing the graphics and waveforms, maybe running BPM or key analysis while three or four tracks are playing with effects. That’s where the system could get busy.
As far as I know, once a track is fully loaded and you see the complete waveform, it’s in memory … it’s not constantly streaming from USB. That means four FLAC tracks playing at once shouldn’t overload the USB bus. But loading/unloading, or scrolling fast through the library, could stress the drive … The real bottleneck can be the drive itself. I’d stick with good SSDs/SD cards, formatted in exFAT, and check them now and then so they don’t die in the middle of a set …
And yeah, after a bunch of firmware updates, a reset can sometimes clear out the weird gremlins. Though maybe a Prime 4+ owner can say something about this and give better advise … I’ve never had to reset my Denon players myself …
If anyone has a stand alone system like the prime 4 + or anyone else could you try just scrolling hard through your library while it plays multiple songs at once and see if any slow down or errors occur?
As for your answers, I do feel you are right, its the drawing of the waveforms and the initial analysis it does everytime a track loads are heavy on the system. I tried turning off track analysis in the pc software but i feel like that only applies to the pc side of things.
Maybe I am still cheaping out on usb devices/ssd! I use two usbs. One for playback and one for recording.
This is a link to my sets if anyone has interest. Thank you again for your insights.
Parysatis, if you don’t mind me asking, what is your deck of choice?
Great question. I misspoke a bit when I mentioned analyzing the track. I meant loading the track. The only thing that may happen is when i load a track i may see the different colors on the waveform appear one at a time but still under 2 secs when that happens.
My method lately has been add it to my pc, re analyze the track, then create stems.
AIFF or WAV is uncompressed, so the unit only needs to read it, no CPU processing. Speed of USB or SD card will determine load speed.
Flac / M4A (alac) is compressed but lossless so needs to “unzip”
MP3 is compressed and lossly so most work.
“scrolling hard” is no problem for me, neither on Prime 4 nor on Prime 4 Plus.
But I detected an other issue (imo caused by the limit of the system) as follows:
If you have listed one track or more in the folder “PREPARE” it is very difficult (or even impossible) to load a track to one of the four platters (tapping or via decoder and then tapping a platter).
My workaround (working not every time): Changing the number of the platter (toggling between 1/3 or 2/4) or waiting a little bit or select in the library again (worse).
I think we need more capacity (chip with more performance or something like that) to solve this issue or is here anybody pls. suggesting a better workaround.
It might be worth looking at some of your album art as others have reported having large album art tagged to a file can cause performance issues. See if any of your tracks have say 20mb or higher art embedded.
I have yet to to try the prepare function. So far I have been improvising on my which track to play next but I will give this a try and get back to you if I can replicate it. Thank you
Let us know if that is the reason. It could be a huge help to lots of people with performance issues.
I personally only ever download fairly low res images from google if I’m updating artwork, but I’m now going to go back and check all the ones I get sent by Bandcamp when I purchase music from there. It hasn’t impacted me but I’ve seen a few mentions of it now so there’s definitely something in it.
…
…