Waveform issue on tracks analysed on P4 G2 on 5.1.0

Engine DJ Hardware Standalone Controllers
, , ,
1

Hi. I’ve been a P4 owner since Jun-19 and massively happy to have just updated to the G2. Everything works brilliantly apart from the waveforms.

The issue seems to be isolated to tracks specifically analysed on the G2 (e.g. all new tracks from streaming platforms, new tracks copied to a USB and not pre-analysed on Engine on a laptop).

Basically, the waveforms look absolutely fine but they can’t be relied on at all to execute a mix. I’ve manually adjusted grids 100s of times in the past, so I know this is not a beatgrid issue per se

IMG_4356
IMG_43561179×2556 338 KB

IMG_4357
IMG_43571179×2556 345 KB

See attached photos (sorry they are blurry, I have a video but it wouldn’t upload).

Photo 1

The waveforms and beatgrid is totally misaligned but the tracks play as though they are perfectly in mix.

Photo 2

I then bring both the waveforms together and you can hear the car crash of a mix that results.

Same happens when using Sync. Tracks do not mix at all and you need to nudge the jog wheel a few times to align the sound.

This has something to do with how the waveforms are appearing visually. It’s definitely not a beat grid setting issue

All my old pre analysed tracks seem to work fine. If I get my laptop to analyse the same track from USB on Engine desktop, the analysis is done perfectly

Is anyone else having similar issues?

Thanks :folded_hands:

2

Omg Omg Omg… This is issue only in your brain,. Not in Prime 4 G2. When mixing tracks, rely on your ears, not on the beatgrid displayed on the screen… Thats all. No issue, no problem. Prime 4 G2 works perfectly, without any problems…

4

Can you share the track names and streaming service, so forum members can reproduce something. I use Tidal btw.

5

Thanks for getting back to me. I’m using Beatport Pro with files set to Lossless FLAC.

The two tracks are Cloud Dub (original) by Franksen and Doo Da Doo (Jamie Jones Niagra extended remix) by Robotman.

i gave these 2 someone else with a G2 and they had no issues. I also repeated this on my old P4 and it worked perfectly. Just not on my own G2.

I do now realise these two tracks have 2 very different tempos which isn’t ideal. I can provide other more suitable examples tonight as I get the same problem on most tracks.

Thanks again for your help.

6

Let’s see what others and me see with these tracks.

Also, check if the BPM range setting matches with your OG P4.

7

that’s the exact opposite of what Social media says to do. :sob:

8

Hi Reese. Thanks again for helping

What do you mean by BPM range settings? Default speed range is set to +|-8 on both P4s. BPM range in Library is set to 78-155 on both.

Another 2 tracks I tried were Darksidevinyl “U R” Original and Lexicon Avenue & Matt Black “Parralax” Dave Walker Remix Their key mixes together and both are 122bpm

Mixing both together when the waveforms and beat grids are totally aligned = train wreck of a mix Nudging the second track several times eventually gets the sound to line up. This is totally the opposite behaviour on my old P4 Thanks dean

9

Beatport says the Dave Walker Remix is at 123 BPM, not 122.

So maybe your problem is just that Engine OS calculated the wrong BPM value and set up the beatgrid wrong?

Then again you say you have experience with manually adjusting the beatgrids…

10

Yep. 123 also here in VirtualDJ.

11

Sorry you’re right. It’s 123bpm original but bought down to 122bpm for the mix. Thanks for the correction

12

I have no issues with the last two. The 123 bpm track needs a pitch of about -0.81%, but stays on grid. These are from Tidal.

Tidal doesn’t seem to have Franksen Cloud Dub, but it has the Robotman track. Engine sees its bpm as 131, but the downbeat is off.

13

Thanks for trying. Hmm so wonder what to do then. Do I go back to the retailer and says this one has a bug and ask for a new one?

14

Don’t know. If the retailer finds this a valid reason, why not.

Perhaps create an EngineDJ support ticket first. See if they can resolve this.

Btw. Did you take the foil of the platters? It’s weird the hardware would be the source of the problems.

15

This still looks to me to be just a regular case of Engine OS messing up the fresh track analysis vs Engine DJ that not only works better but has a reanalyze option.

Waveform issue it’s definitely not, it would only be a cese if you hear the sound before the wareform hits the playhead or opposite.

16

It’s the same algo and as it is streaming tracks, only OS is used. The weird thing is here that other P4 G2 devices don’t have this issue.

17

It was always the same algo but we have reports from 2021 onward about streaming tracks analysis being wrong vs desktop.

To confirm no other P4 gen2 has the same we need someone with it to load the Franksen Cloud Dub and see if the beatgrid is in the same position visually vs audibly

18

I meant that streaming tracks aren’t available in Desktop. Only “your” local copy, which most of the time is not the same.

Yep, but I don’t have Beatport anymore.

Plus, as stated:

19

Hi zusammen,

I’ve been playing around with the m.db grid data myself recently, and I’ve already seen some rather curious things there. In my case, however, all of them turned out to be caused by my own modifications, so I wouldn’t draw any conclusions from that.

In my opinion, there is really only one reliable way to investigate this properly:

First, make a backup of the current database (or rename it) so nothing gets lost.

Then do a factory reset and start with a completely clean USB drive. I actually think the factory reset is important here, because it gives us a defined starting point and, just as importantly, allows us to create a database containing only the relevant test tracks.

If necessary, I would even remove the internal SSD to make absolutely sure that no existing database or analysis data can interfere.

Then load only the known affected TIDAL tracks onto the unit and let the P4 G2 analyse them normally. This should give us a clean m.db and the corresponding UUID.rgb data on the USB drive.

This would also make the investigation much easier for Support. Instead of having to dig through a database containing several thousand tracks, they would have a small, clearly defined database containing only the tracks relevant to the problem.

The really useful next step would be to get exactly the same tracks available locally as well, so that Support or someone with the time to investigate could perform a controlled comparison of the resulting data.

There is another important variable here: compression.

The track which is streamed, decrypted and decoded on the device may not necessarily be bit-identical to the corresponding locally stored file. In theory, this should not make any difference to beatgrid data, and probably not to the waveform either — but that is exactly something we shouldn’t have to guess about.

As there seem to have been several reports of similar oddities already, perhaps this would be a good opportunity to actually collect the required data from one affected user and investigate it under controlled conditions.

That would probably get us much further than trying to interpret screenshots and speculate about what might be happening.

BTW, I personally don’t believe this is a hardware defect or anything like that.

My guess is that we are looking at an unfortunate combination of circumstances or settings — and that a clean, controlled test might be enough to find out what is actually causing it.

20

I could be wrong, but the sm.db is/was perhaps for streaming tracks. Your m.db is for local media tracks.

21

Yes, you’re right. I just can’t remember at the moment which database is used for which purpose.

That’s why I was thinking about using a completely empty USB stick. EngineOS would create everything it needs there, so not only would the amount of data be very manageable, but all the relevant database files would be in one place.

The only thing that would then need to be added would be the local media files themselves — essentially creating a clean, self-contained test environment.

Maybe that’s something more useful for the support team internally?

Either way, I think this could be a good opportunity to get to the bottom of what is actually happening.